This week, I’d like to discuss the last few cost-effective computer upgrades. Some of these upgrades, such as CPU and system board replacements, are more likely to require the services of an experienced local technician to actually implement them.
CD/DVD drives. Modern CD and DVD drives are inexpensive and faster than earlier drives. I’ve bought good-quality Sony CD/DVD drives for as little as $20 to $40. This is an easy and worthwhile upgrade. Remove both side panels and unplug the power and ribbon cables, noting the correct orientation for later reassembly. Unscrew the four screws attaching the drive to the case, slide out the old drive through the front panel and slide in the new drive from the front panel. Reattach screws, power plug and data cable.
Upgrading CPU processors and system boards. I usually don’t recommend that an inexperienced home user upgrade either the CPU processor or the system board even though they can be a worthwhile upgrade. Unless you have prior experience replacing a CPU, there is a very good chance that you will permanently damage this expensive but fragile component. A professional technician can usually do the upgrade in a fairly short time.
The most cost-effective CPU upgrade would be one in which you substitute a significantly faster version of the same processor type. Usually, the most economical and beneficial approach involves replacing an entry-level CPU with a later, faster version of the same product line, installing additional cooling as needed. I believe the best cost-performance ratio is usually found in CPUs that are one speed grade lower than the current top-end model. Because system board and memory hardware may differ among product lines, you probably should consult an experienced local technician to evaluate whether it’s worth upgrading your system, determining the most appropriate replacement CPU, and have that technician do the upgrade.
Upgrading system boards is even more difficult. You’ll need to make many different sorts of connections and set numerous options during system board installation. Unless you already have significant experience installing new system boards, this is definitely a job for an experienced technician.
My current preference for cost-effective CPUs is the 2.8 Ghz AMD quad-core Phenom II line. Although most operating systems do not efficiently use all four CPUs on any processing chip, this new processor line does have significant overall performance improvements compared to earlier models. You’ll most likely need to upgrade the system board and memory. The Phenom II requires DDR2 1066 memory and a system board with an AM2+ socket. I like the Gigabyte boards built around the AMD 790 series chipsets.
Power supplies. The fast new hardware that we’ve discussed over the past several weeks does require a lot of power and cooling, and you may need to upgrade your power supply. At this time, I am using 600-watt power supplies to ensure that there’s ample power, especially if you have installed more than one hard disk. The brand is not important but be sure it has enough of the right kinds of power plugs — at least two, preferably four, SATA hard disk power plugs, at least one six-pin PCI-E video card plug, plus the usual 20-pin and four-pin system board main power plugs. You’ll want a floppy disk power connector and some spare, traditional four-pin Molex power plugs for your CD/DVD, fans and front panel power lights. The power supply should have one or two large, quiet fans. I prefer 120-millimeter fans to move a lot of air quietly.
Replacing the power supply is not difficult if you take care to avoid disturbing other cables, connections and settings inside your computer. Open up both side panels of the computer case. Disconnect the main AC power card from the wall outlet. Then disconnect all internal connections to system boards, disk drives of all sorts, fans and video cards. Remove the four rear screws that attach the power supply to the case, carefully sliding it out to avoid hitting and damaging other components. Slide in the new power supply, attach the screws and reconnect the various power plugs.
Ideally, rather than leaving a tangled mess of power wires inside the computer case where they can get caught in fans or tug on other connections, take some heavy wire ties, loop the extra connectors and overly long cables into one or more neat bundles and wire tie them together.
Video card upgrades. Upgrading to a really fast video card is worthwhile only if you are a serious gamer, a designer using AutoCAD or using the latest version of Adobe PhotoShop CS4 and other higher-end Abode products. Most business users will not see any obvious benefit.
Having said that, upgrading a video card can be easy if your intended upgrade card contains the same brand video chipsets as the video card you’re replacing. Although there are many video card brands, almost all include basic video chipsets made by either ATI or nVidia.
Both of these vendors use what’s often termed “unified drivers,” which means the recent ATI or nVidia device driver software will work with nearly all brands of recent video cards using a wide range of low-end to high-end ATI or nVidia chipsets. That’s a positive development because mismatched video driver software will usually result in a complete system crash at boot-up.
For this reason, a home upgrader is well-advised to purchase an upgrade card that incorporates the basic video chips from the same vendor as their existing display hardware and to first ensure they have already installed the latest stable device driver software from either ATI or nVidia, as appropriate. Under the best circumstances, physically swapping the video cards will be all that’s required.
To determine whether your computer currently uses either ATI or nVidia video chipsets, check the printed materials that came with your computer. If that is not clear, go to “Start,” “Settings,” “Control Panel,” “System,” “Hardware” tab, “Device Manager” button, “Display Adapters” and then click on the “Display Adapters” + box. Windows will list the video chipset manufacturer and series. Double click on the chipset listing and you will bring up the “Properties” box. Click on the “Driver” table and Windows will list complete information about which video card device driver software is currently installed. If it’s within a year or so, there should be no problem.
You can manually check for a standard update by clicking on “Start,” “Windows Update” and select the “custom” button. After Windows Update checks your computer for currently installed software, click on the “Hardware, optional” button to see whether newer video software is available and certified by Microsoft. If so, choose and install it.
If you decide to upgrade with a video card using a different vendor’s chipsets, then you simply need to temporarily set your display to VGA 640 x 480 or 800 x 600 before starting your upgrade. To do this, right click on a blank area of your Windows desktop, choose “Properties,” then the “Settings” tab and adjust the “screen resolution” slider to these lower resolutions.
Because I use several high-end Adobe products in my legal practice, I looked carefully for an advanced video chipset that supported all of the various Adobe Master Suite CS4 products, not just PhotoShop. Weirdly, some of the Adobe products packaged together as Abode Master Collection CS4 are certified with some video chipsets and other Adobe products are certified with an entirely different group of video chipsets.
The only economical video chipset that I found to be broadly compatible was the ATI Radeon 3870 set of products. Because I have been using lower-end Radeon video cards for some time, installing a 3870-based video card was easy, at least for Windows XP systems. I found the single ATI 3870 GPU cards made by Sapphire were fast, well made and affordable, costing between $85 and $100.
Summary Upgrade Specifications. Here’s a modern computer system that I would consider a very good balance of performance and cost-effectiveness:
1. Mid-tower case with 600-watt power supply, brand unimportant.
2. AMD Phenom II 2.8 gigahertz CPU processor.
3. Gigabyte Socket AM2+ system board, model depending upon your particular peripheral device needs. I prefer the newer AMD 790 series of chipsets.
4. Four gigabytes DDR2 1066 DRAM memory.
5. Video card based upon ATI single-processor 3870 GPU chipset. Sapphire brand has been reliable and inexpensive.
6. Sony or comparable 16x DVD and CD burner.
7. Standard 3.5-inch, 1.44 MB floppy disk for operating system installation.
8. Western Digital Black Caviar 640GB WD6401AALS hard disk. If you are interested in maximum performance, then use a 10,000-rpm Western Digital. 300GB VeliceRaptor hard disk as your C: boot drive and program storage drive and the 640 GB WD6401AALS drive as your D: drive where your data would be stored.
9. Logitech wireless keyboard and mouse. Surprisingly, the less expensive models seem more reliable.
10. Digital monitor to taste. At the moment, the 23- and 24-inch 1080p flat screen monitors from Acer, ASUS, HP and Samsung seem like the best deals.
Video Card Setup: WARNING: If any technical setup information below seems at all confusing to you, then I would urge you to stop at this time and either abandon any video upgrade or get the job done by a competent local technician.
A few days before you update your video device driver software, and again when you later upgrade your video hardware go to Windows “Start,” “Settings,” “Control Panel,” “System,” “System Restore” and be absolutely sure that “System Restore” has been operating and tracking all drives. Use your system regularly for a few days to ensure that you have at least one good system restore point..
When you are ready to install the new video card, shut down your computer, ground yourself to avoid static discharge, and unplug the power cord. Open the side panel that allows access to the system board. Find the first PCI-E card slot. If your computer is too old to include a PCI-E slot, then it is too old to be worth upgrading even if you could still find the older hardware.
The physical side of the upgrade process can be straightforward unless other computer components physically get in the way. Many modern computers, particularly less expensive systems, often include a lower-end video display on the system board rather than as a separate video card. In that case, you will not have a separate display card to first remove but instead you’ll just plug the new video card directly into the first PCI-E slot, secure it to the case with the appropriate screw, and, if required, plug in a 6-pin auxiliary power cable.
However, when you next reboot, you will first need to go into the BIOS setup and configure your hardware to set the new video card as the default boot-up display. If you don’t have a free 6-pin power plug, then you will need to get a simple adapter from older 4-pin Molex power plugs to a 6-pin PCI-E video card power plug.
Again, don’t try this at home if you are at all uncertain about what to do. Instead, have a local technician do the work for you. Making a mistake in the BIOS setup can make your system unusable until a technician restores the BIOS to factory defaults.
Physically installing a powerful new video card can be sublimely easy or really tricky, depending upon whether your system has a modern power supply with a free 6-pin power plug for the PCI-E video card and depending upon whether your existing computer has system board components, hard disks, cables or other hardware that physically get in the way.
Even if your system does not have a free six-pin PCI-E power connector, a computer technician often can provide a simple adapter that converts traditional four pin Molex power connectors to 6-pin PCI-E, so this is not a serious problem. Likewise, if you have cables or hard disks that physically interfere, these can be moved if you know what you’re doing. However, both of these factors should be taken into account before you decide whether to install a faster video card. Upgrading simply may not be worth the trouble.
I upgraded three Windows XP and Windows XP x64 computers with these Sapphire video cards and found that the upgrades to Windows XP systems were easy and worthwhile. Upgrading the Windows XP x64 system was troublesome because ATI’s newest 2009 x64 drivers were not stable, forcing me rebuild the system and revert to ATI’s older but stable July 2008 x64 software. I would expect the same stability problems with the 64-bit version of Windows Vista, which is based upon the XP x64 software.
If you experience any crashing after upgrading device driver hardware, then boot up in Windows Safe Mode by passing the F8 key several times during your next boot-up. Windows Safe Mode will start and then ask you whether you wish to boot normally. Answer No and then choose a system restore point that you know will work reliably. Your system will revert to the earlier software and settings, thus giving you a stable system and another chance to cause yourself grief.
After you have physically installed any new video card, then run the installation disk that comes with your new hardware and again go to “properties”, “settings” to configure your new display to your taste.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
Showing posts with label technology. Show all posts
Showing posts with label technology. Show all posts
Tuesday, April 7, 2009
Tuesday, March 31, 2009
Plugged In: Don’t let Redoubt’s ash be downfall of electronics
Before concluding our series about cost-effective computer upgrades, here’s a brief word about dealing with the byproducts of our namesake across Cook Inlet, whose unpredictable geological mutterings are likewise kicking up some dust on the central Kenai Peninsula.
Volcanic emissions are potentially damaging to almost any mechanical or electrical device. Damage can occur in several ways. Although light traces of ash are not a problem, it’s wise to remember that local volcanic ash is an industrial-grade abrasive, which, in sufficient quantities, can abrade and mechanically damage engines, bearings, fans or any other rotating or mechanically actuated devices.
This can quickly cause major computer failures. For example, if the bearings on your CPU’s cooling fan fail, either your computer will shut down almost immediately or your CPU will basically suffer catastrophic heat stroke within a few seconds. Modern processors really do put out enough heat to cook themselves unless constantly and efficiently cooled. In addition to general abrasion of moving parts, other electronic devices like printers, scanners and photocopiers are highly vulnerable to mechanical scratching of interior parts and can become quickly and permanently unusable unless covered and cleaned with compressed air. Don’t try to wipe the dust off such parts and surfaces — blow it off.
Volcanic ash can also damage electronics because it’s electrically conductive, which can cause short circuits. It’s also chemically corrosive due to sulfur oxides which, when combined with atmospheric water, hydrolyze into sulfuric acid, among other noxious chemicals. That’s the same process that results in long-term toxic acid drainage from open-pit mines and mine tailings exposed to oxygen.
Even if abrasion, corrosion or short circuits don’t get you, and even if the ash doesn’t clog your computer’s cooling fans, accumulated ash and other dust is a good heat insulator, causing electronic devices to retain heat, sometimes to the point that the device fails.
I have lost high-end Western Digital Raptor hard disks because I forgot to take a computer outside every few months and thoroughly blow out any and all dust, including dust hiding on the bottom of exposed hard disk circuit boards. The older, 36-gigabyte Raptor hard disk ran pretty warm even when everything was being cooled properly. That disk didn’t have a chance when its circuit board had a nice warm insulating layer of dust. Redoubt’s ash falls will greatly increase the need for regular internal cleaning.
What to do? Appropriate precautions are quite straightforward. Monitor the Alaska Volcano Observatory’s Web site, www.avo.alaska.edu, throughout the day and evening so you’re informed about any explosive ash events and the probable trajectory for any ash that might be ejected over the next 24 hours. You can’t predict whether there will be ash at any given time but you can predict where any emissions will travel on the winds over the next 18 to 24 hours.
When there’s any chance of an ash fall coming down over your area, shut down all of your electronic equipment immediately. Also shut down any uninterruptible power supplies or other battery-operated devices to avoid possible short circuits due to conductive ash. Unplug them from the wall if you don’t have high-grade surge protectors installed. If there’s a power loss, then there’s also the possibility of damaging voltage surges when electrical power is restored. Homer Electric Association’s state-approved tariffs provide that you, the end consumer, are responsible for preventing surge damage and that HEA is not liable.
If possible, keep dust from infiltrating into your premises. Seal leaky windows with wide masking tape. Prevent ash from being tracked into your house and workplace. Put extra filters on any air intakes. For example, I put some regular furnace filters in front of the normal dust filters in my office’s air-handling system. This should provide an easily replaced second line of defense against ash being sucked into my building and being uniformly distributed by the heating system. Also consider shutting down your air-handling systems if you’re present during the day, or setting them so low at night that they’ll not activate very much, if at all, throughout the night.
Covering all electronic and sensitive mechanical equipment is obvious. For more expensive or sensitive devices, I have been fully covering the top and sides of each individual device with a 33-gallon trash bag and then covering the entire area with large, heavy-duty plastic drop cloths bought at Trustworthy Hardware. That gives me a second layer of protection. Ash slides easily, so I have been turning up the edges of the drop clothes to prevent ash from sliding off them on to whatever I am trying to protect.
You’ll need to internally clean any ash or dust regularly. Even if you’re not in the path of an obvious ash event, there’s enough circulating ash that your computer will build up a damaging layer of heat-insulating internal dust more quickly than usual. The best way to clean this dust is to buy several cans of compressed nitrogen or dried air. These are available locally at Three Bears, Save-U-More and other stores.
Unplug your computer or printer, first noting and marking where each cable attaches so that you’ll be able to reassemble the computer without any configuration problems. Then, take the entire system outside, remove each of the sliding side panels and thoroughly blow out any and all dust. In addition to obvious general dust, take particular care to clean the system board, the underside of any hard disks, all fan blades and fan housings, all power supply openings, and the air spaces between the fins on the coolers on the CPU, the video card and any smaller fans on the system board. Repeat regularly while Redoubt’s dust is floating about.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
Volcanic emissions are potentially damaging to almost any mechanical or electrical device. Damage can occur in several ways. Although light traces of ash are not a problem, it’s wise to remember that local volcanic ash is an industrial-grade abrasive, which, in sufficient quantities, can abrade and mechanically damage engines, bearings, fans or any other rotating or mechanically actuated devices.
This can quickly cause major computer failures. For example, if the bearings on your CPU’s cooling fan fail, either your computer will shut down almost immediately or your CPU will basically suffer catastrophic heat stroke within a few seconds. Modern processors really do put out enough heat to cook themselves unless constantly and efficiently cooled. In addition to general abrasion of moving parts, other electronic devices like printers, scanners and photocopiers are highly vulnerable to mechanical scratching of interior parts and can become quickly and permanently unusable unless covered and cleaned with compressed air. Don’t try to wipe the dust off such parts and surfaces — blow it off.
Volcanic ash can also damage electronics because it’s electrically conductive, which can cause short circuits. It’s also chemically corrosive due to sulfur oxides which, when combined with atmospheric water, hydrolyze into sulfuric acid, among other noxious chemicals. That’s the same process that results in long-term toxic acid drainage from open-pit mines and mine tailings exposed to oxygen.
Even if abrasion, corrosion or short circuits don’t get you, and even if the ash doesn’t clog your computer’s cooling fans, accumulated ash and other dust is a good heat insulator, causing electronic devices to retain heat, sometimes to the point that the device fails.
I have lost high-end Western Digital Raptor hard disks because I forgot to take a computer outside every few months and thoroughly blow out any and all dust, including dust hiding on the bottom of exposed hard disk circuit boards. The older, 36-gigabyte Raptor hard disk ran pretty warm even when everything was being cooled properly. That disk didn’t have a chance when its circuit board had a nice warm insulating layer of dust. Redoubt’s ash falls will greatly increase the need for regular internal cleaning.
What to do? Appropriate precautions are quite straightforward. Monitor the Alaska Volcano Observatory’s Web site, www.avo.alaska.edu, throughout the day and evening so you’re informed about any explosive ash events and the probable trajectory for any ash that might be ejected over the next 24 hours. You can’t predict whether there will be ash at any given time but you can predict where any emissions will travel on the winds over the next 18 to 24 hours.
When there’s any chance of an ash fall coming down over your area, shut down all of your electronic equipment immediately. Also shut down any uninterruptible power supplies or other battery-operated devices to avoid possible short circuits due to conductive ash. Unplug them from the wall if you don’t have high-grade surge protectors installed. If there’s a power loss, then there’s also the possibility of damaging voltage surges when electrical power is restored. Homer Electric Association’s state-approved tariffs provide that you, the end consumer, are responsible for preventing surge damage and that HEA is not liable.
If possible, keep dust from infiltrating into your premises. Seal leaky windows with wide masking tape. Prevent ash from being tracked into your house and workplace. Put extra filters on any air intakes. For example, I put some regular furnace filters in front of the normal dust filters in my office’s air-handling system. This should provide an easily replaced second line of defense against ash being sucked into my building and being uniformly distributed by the heating system. Also consider shutting down your air-handling systems if you’re present during the day, or setting them so low at night that they’ll not activate very much, if at all, throughout the night.
Covering all electronic and sensitive mechanical equipment is obvious. For more expensive or sensitive devices, I have been fully covering the top and sides of each individual device with a 33-gallon trash bag and then covering the entire area with large, heavy-duty plastic drop cloths bought at Trustworthy Hardware. That gives me a second layer of protection. Ash slides easily, so I have been turning up the edges of the drop clothes to prevent ash from sliding off them on to whatever I am trying to protect.
You’ll need to internally clean any ash or dust regularly. Even if you’re not in the path of an obvious ash event, there’s enough circulating ash that your computer will build up a damaging layer of heat-insulating internal dust more quickly than usual. The best way to clean this dust is to buy several cans of compressed nitrogen or dried air. These are available locally at Three Bears, Save-U-More and other stores.
Unplug your computer or printer, first noting and marking where each cable attaches so that you’ll be able to reassemble the computer without any configuration problems. Then, take the entire system outside, remove each of the sliding side panels and thoroughly blow out any and all dust. In addition to obvious general dust, take particular care to clean the system board, the underside of any hard disks, all fan blades and fan housings, all power supply openings, and the air spaces between the fins on the coolers on the CPU, the video card and any smaller fans on the system board. Repeat regularly while Redoubt’s dust is floating about.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
Tuesday, March 24, 2009
Plugged In: Hardware upgrades can be relatively easy fix to sluggish PC performance
After this weekend’s bout of upgrading my own computers, recalibrating printers and then shutting down for Mount Redoubt’s tantrums, I’m not inclined to write a sassy lead, so let’s dive straight into geek speak mode and examine how to implement some useful and cost-effective PC upgrades.
Luckily, upgrading DRAM is fairly easy, and DRAM has become quite inexpensive. I’ve seen a four-gigabyte matched pair of brand-name DDR2 1066 DRAM for as little as $63 plus shipping prior to a substantial rebate.
Modern replacement DRAM must always be installed in matched pairs and be of the correct type and speed for your system. There are several current types of DRAM, termed DDR, DDR2 and DDR3. These are absolutely NOT interchangeable.
Plan upon removing any DRAM that’s already installed in your system — it’s not going to work reliably with your newer-upgrade DRAM. All installed DRAM in all channels should be identical.
A 32-bit Windows XP or 32-bit Windows Vista operating systems are physically unable to use more than four gigabytes DRAM. Although two gigabytes should be sufficient for most consumers, DRAM has become so inexpensive that there’s little economic benefit to installing less than a full four gigabytes.
A-64 bit operating system like Windows XP x64 or 64-bit Vista can physically access much more DRAM, so an upgrade to at least four gigabytes DRAM would be in order.
Before you start your upgrade, discharge any static electricity by touching the metal exterior of the computer case, then unplug your computer’s power cord from the back of the system.
Memory modules are notched off-center so they will snap and lock into place only when correctly inserted. Don’t force the module but be sure that the lock on each end of the memory slot snaps securely into the notch at each end of the DRAM module.
Modern computers typically include four slots for two separate memory channels; each channel will have two separate memory slots. Be sure that your upgrade modules are properly inserted into the two matching slots for the first memory channel. Although these are usually color-coded, check your computer and system board manual if you’re not sure which slots to use. These manuals are often found online at the system board manufacturer’s Web site. If you decide to upgrade to four gigabytes, then installing two, two-gigabyte memory modules will usually be preferable and less expensive.
If you haven’t installed the new DRAM properly, your system will not boot, a situation that’s usually the result of bad memory or improperly installed memory. Be sure that your memory is fully and properly seated in the correct slots. If you’re not sure, remove all memory modules, check your system board manual and start over.
Be sure to buy DRAM that is the correct type and has a rated performance at least as fast as the minimum speed rating required by your CPU and system board. As a relatively inexpensive precaution, I generally prefer to buy upgrade DRAM that is one grade faster than the minimum required by my system. It will work at the lower speed and provide a little extra margin of reliability. For example, if your old DRAM is DDR2 800, then I personally would prefer DDR2 1066.
Do the easy part first. Not all new model hard disks provide equally good performance. Generally, hard disks with higher rotational speeds and higher storage capacity tend to perform better for the simple reason that high-rpm hard disks, and hard disks with a great amount of data stored in a very small area, will mechanically rotate that much more data under the magnetic read-write heads inside the hard disk.
I have not experienced any problems associated with compressing the disks of my own computers, but I strongly recommend against trying compression on a multidisk RAID array because RAID arrays store data in a manner very different from an individual hard disk. I believe that the risks associated with compressing a RAID disk array are unpredictable but potentially severe.
Disk compression is very slow, so it’s something that should be done overnight. To access disk compression, click on My Computer, then right click on the C: drive icon. Left click on the Properties menu item, and then choose Disk Cleanup to delete any junk files. Empty the Recycle Bin. Then, in the same disk Properties window, check the Compress Drive check box, choose to apply to all subdirectories and files, and then wait several hours.
Do not use your computer and do not turn it off until the process is competed. Repeat this sequence for any D: or other internal hard disks. Ideally, any computer running disk compression should be connected to an uninterruptible power supply to avoid an unplanned shutdown in the event that there’s a power outage.
The fastest consumer hard disks are all SATA type drives, and of these the 10,000-rpm Western Digital Raptor drives are remain the fastest consumer-level hard disks. The 300-GB VelociRaptor currently costs about $229, while the slightly slower 150-GB version costs about $179. The older but still very quick 74-GB Raptor is now $99. You’re paying a premium, though, per gigabyte of storage for the fastest possible hard disk performance.
The overall performance of Western Digital’s Black Caviar and Caviar SE16 hard disks is only slightly slower than the Raptors and the cost per gigabyte is quite a bit lower. The 7,200-rpm, 640-GB Black Caviar drive (WD6401AALS) is an unusually good performer among 7,200-rpm drives, and only costs about $80 plus shipping, a very reasonable price that works out to about 12 cents per gigabyte.
My recommendation for the performance-oriented user would be to use a Raptor drive as the C: boot drive, from which Windows and all application programs are loaded, and a WD, 640-GB Black Caviar drive as a D: drive, where all data and documents are stored. Not only does this arrangement optimally balance performance, cost and storage capacity, but it’s easier to reliably back up your data if the data is located on a physically separate hard disk.
If your system board is fairly new and has the necessary SATA RAID hardware attachments, and if your computer’s case and power supply are sufficiently robust, using four WD6401AALS drives in a RAID 10 disk array would result in a very low-cost. but fast and reliable, RAID 10 disk array. I recently installed such an array in my own law office file server and have been very happy with it. Although the RAID 10 disk array would not have as much storage capacity as the four individual drives, the cost per gigabyte is still excellent and your shard disk performance will be quite a bit faster and more reliable.
It’s not physically difficult to replace a hard disk, mostly consisting of loosening four screws, an SATA power plug, and an SATA data plug, switching the hard disks, and then replacing screws and the data and power cables. The real trick is to move your operating system, data and paid-up program activations as easily and smoothly as possible.
To do that, you’ll need fresh hard disks, some external SATA disk enclosures that connect to your computer system through USB 2.0 ports, and a special program from Acronic called Easy Migrate 7. You can buy a copy of the Easy Migrate program as an Internet download from www.acronis.com. Easy Migrate basically copies the internal hard disks of your computer to the new target hard disks, which are usually connected through USB external hard disk enclosures.
Most external hard disk enclosures will work OK if you make sure the enclosures are specifically intended for use with SATA drives. I’ve gone through many external hard disk enclosures over the years and they’re mostly a yawn. As long as they work reliably, that’s usually good enough.
However, I recently found a series from Bytecc, the ME-300SU-BK 3.5 Black USB 2.0 Easy-Open Enclosure, that cost only $27 each including base and power supply. This external disk enclosure seems faster than average and is quite compact because it fits the SATA drive so snugly that the case itself becomes the cooling unit for the hard disk, neatly solving several problems at once. These are also recommended as a home and office data backup solution when used with the WD 640 GB Black Caviar drives and Microsoft’s basic backup accessory.
Acronis Easy Migrate 7 generally does a very good job of transferring the Windows operating system, application programs and data to new hard disks and then setting these disks up properly for physical transfer into your computer.
I found the program properly transferred most program activation data but failed to transfer activation and license data for some lesser-known utility programs and some device drivers. As a result, I had to delete, then reinstall, a few scanners and printers. I also had to find my purchase invoices and transfer the activation data and serial numbers manually into several programs. Aside from that fairly minor glitch, using Easy Migrate was fairly reliable and certainly much easier than re-installing operating systems and application programs to new hard disks.
Next week, we’ll continue the hardware upgrade discussion with information on changing out CPUs, video cards and power supplies.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
Adding RAM and why
Microsoft and most third-party software vendors constantly add more and more “little” bits of software that automatically load when your system first boots. Every “little” bit demands more of your system’s DRAM memory, sometimes as much as an additional 25 megabytes. As a result, 2- or 3-year-old Windows XP systems, which typically sold with 512 MB installed DRAM memory, frequently no longer have enough free DRAM to work quickly and efficiently, even with undemanding application programs like basic word processing, Web access or e-mail. Demanding applications like photographic programs need a great deal of free DRAM or they’ll slow down to the point of unusability.Luckily, upgrading DRAM is fairly easy, and DRAM has become quite inexpensive. I’ve seen a four-gigabyte matched pair of brand-name DDR2 1066 DRAM for as little as $63 plus shipping prior to a substantial rebate.
Modern replacement DRAM must always be installed in matched pairs and be of the correct type and speed for your system. There are several current types of DRAM, termed DDR, DDR2 and DDR3. These are absolutely NOT interchangeable.
Plan upon removing any DRAM that’s already installed in your system — it’s not going to work reliably with your newer-upgrade DRAM. All installed DRAM in all channels should be identical.
A 32-bit Windows XP or 32-bit Windows Vista operating systems are physically unable to use more than four gigabytes DRAM. Although two gigabytes should be sufficient for most consumers, DRAM has become so inexpensive that there’s little economic benefit to installing less than a full four gigabytes.
A-64 bit operating system like Windows XP x64 or 64-bit Vista can physically access much more DRAM, so an upgrade to at least four gigabytes DRAM would be in order.
Before you start your upgrade, discharge any static electricity by touching the metal exterior of the computer case, then unplug your computer’s power cord from the back of the system.
Memory modules are notched off-center so they will snap and lock into place only when correctly inserted. Don’t force the module but be sure that the lock on each end of the memory slot snaps securely into the notch at each end of the DRAM module.
Modern computers typically include four slots for two separate memory channels; each channel will have two separate memory slots. Be sure that your upgrade modules are properly inserted into the two matching slots for the first memory channel. Although these are usually color-coded, check your computer and system board manual if you’re not sure which slots to use. These manuals are often found online at the system board manufacturer’s Web site. If you decide to upgrade to four gigabytes, then installing two, two-gigabyte memory modules will usually be preferable and less expensive.
If you haven’t installed the new DRAM properly, your system will not boot, a situation that’s usually the result of bad memory or improperly installed memory. Be sure that your memory is fully and properly seated in the correct slots. If you’re not sure, remove all memory modules, check your system board manual and start over.
Be sure to buy DRAM that is the correct type and has a rated performance at least as fast as the minimum speed rating required by your CPU and system board. As a relatively inexpensive precaution, I generally prefer to buy upgrade DRAM that is one grade faster than the minimum required by my system. It will work at the lower speed and provide a little extra margin of reliability. For example, if your old DRAM is DDR2 800, then I personally would prefer DDR2 1066.
Replacing hard disks
Hard disk performance is another area where a fairly inexpensive replacement can yield substantial performance benefits. However, changing to a newer hard disk can be daunting.Do the easy part first. Not all new model hard disks provide equally good performance. Generally, hard disks with higher rotational speeds and higher storage capacity tend to perform better for the simple reason that high-rpm hard disks, and hard disks with a great amount of data stored in a very small area, will mechanically rotate that much more data under the magnetic read-write heads inside the hard disk.
Before you upgrade
Before deciding to upgrade your hard disk, however, I suggest you try a free optional feature in Windows XP — disk compression, which improves the efficiency of how Windows stores hard disks data. My tests indicate that compressing a disk can improve some hard disk performance metrics as well as freeing up some storage capacity.I have not experienced any problems associated with compressing the disks of my own computers, but I strongly recommend against trying compression on a multidisk RAID array because RAID arrays store data in a manner very different from an individual hard disk. I believe that the risks associated with compressing a RAID disk array are unpredictable but potentially severe.
Disk compression is very slow, so it’s something that should be done overnight. To access disk compression, click on My Computer, then right click on the C: drive icon. Left click on the Properties menu item, and then choose Disk Cleanup to delete any junk files. Empty the Recycle Bin. Then, in the same disk Properties window, check the Compress Drive check box, choose to apply to all subdirectories and files, and then wait several hours.
Do not use your computer and do not turn it off until the process is competed. Repeat this sequence for any D: or other internal hard disks. Ideally, any computer running disk compression should be connected to an uninterruptible power supply to avoid an unplanned shutdown in the event that there’s a power outage.
Recommended hard disks
Consumers have relatively limited choices in high performance hard disks because the really fast 15,000-rpm SCSI drives are too expensive and too complex for consumer-grade computer systems. Most consumer-grade hard disks are SATA drives that rotate at 7,200 rpm. Some of these can be surprisingly good performers.The fastest consumer hard disks are all SATA type drives, and of these the 10,000-rpm Western Digital Raptor drives are remain the fastest consumer-level hard disks. The 300-GB VelociRaptor currently costs about $229, while the slightly slower 150-GB version costs about $179. The older but still very quick 74-GB Raptor is now $99. You’re paying a premium, though, per gigabyte of storage for the fastest possible hard disk performance.
The overall performance of Western Digital’s Black Caviar and Caviar SE16 hard disks is only slightly slower than the Raptors and the cost per gigabyte is quite a bit lower. The 7,200-rpm, 640-GB Black Caviar drive (WD6401AALS) is an unusually good performer among 7,200-rpm drives, and only costs about $80 plus shipping, a very reasonable price that works out to about 12 cents per gigabyte.
My recommendation for the performance-oriented user would be to use a Raptor drive as the C: boot drive, from which Windows and all application programs are loaded, and a WD, 640-GB Black Caviar drive as a D: drive, where all data and documents are stored. Not only does this arrangement optimally balance performance, cost and storage capacity, but it’s easier to reliably back up your data if the data is located on a physically separate hard disk.
If your system board is fairly new and has the necessary SATA RAID hardware attachments, and if your computer’s case and power supply are sufficiently robust, using four WD6401AALS drives in a RAID 10 disk array would result in a very low-cost. but fast and reliable, RAID 10 disk array. I recently installed such an array in my own law office file server and have been very happy with it. Although the RAID 10 disk array would not have as much storage capacity as the four individual drives, the cost per gigabyte is still excellent and your shard disk performance will be quite a bit faster and more reliable.
How to change a hard disk
Changing hard disks is not necessarily for the computer novice, so if you’re uncertain, have the job done by a local technician. Before you do anything, be sure you have completely backed up every hard disk in your system and that the backups are reliable.It’s not physically difficult to replace a hard disk, mostly consisting of loosening four screws, an SATA power plug, and an SATA data plug, switching the hard disks, and then replacing screws and the data and power cables. The real trick is to move your operating system, data and paid-up program activations as easily and smoothly as possible.
To do that, you’ll need fresh hard disks, some external SATA disk enclosures that connect to your computer system through USB 2.0 ports, and a special program from Acronic called Easy Migrate 7. You can buy a copy of the Easy Migrate program as an Internet download from www.acronis.com. Easy Migrate basically copies the internal hard disks of your computer to the new target hard disks, which are usually connected through USB external hard disk enclosures.
Most external hard disk enclosures will work OK if you make sure the enclosures are specifically intended for use with SATA drives. I’ve gone through many external hard disk enclosures over the years and they’re mostly a yawn. As long as they work reliably, that’s usually good enough.
However, I recently found a series from Bytecc, the ME-300SU-BK 3.5 Black USB 2.0 Easy-Open Enclosure, that cost only $27 each including base and power supply. This external disk enclosure seems faster than average and is quite compact because it fits the SATA drive so snugly that the case itself becomes the cooling unit for the hard disk, neatly solving several problems at once. These are also recommended as a home and office data backup solution when used with the WD 640 GB Black Caviar drives and Microsoft’s basic backup accessory.
Acronis Easy Migrate 7 generally does a very good job of transferring the Windows operating system, application programs and data to new hard disks and then setting these disks up properly for physical transfer into your computer.
I found the program properly transferred most program activation data but failed to transfer activation and license data for some lesser-known utility programs and some device drivers. As a result, I had to delete, then reinstall, a few scanners and printers. I also had to find my purchase invoices and transfer the activation data and serial numbers manually into several programs. Aside from that fairly minor glitch, using Easy Migrate was fairly reliable and certainly much easier than re-installing operating systems and application programs to new hard disks.
Next week, we’ll continue the hardware upgrade discussion with information on changing out CPUs, video cards and power supplies.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
Tuesday, March 17, 2009
Plugged in: Troubleshoot PC problems before throwing money them
A few months ago, I became concerned that the central file server at my law office was performing much slower than expected, given the very fast components from which I built that device. Because I didn’t want to just throw money at a balky computer, my first reaction was to systematically troubleshoot the system.
The best way to troubleshoot a slow computer is to individually test each subset for overall performance and see whether you have a bottleneck that can be easily and inexpensively upgraded. For this purpose, I like PassMark’s Performance Test 6.1, which comes in both a 32-bit version for regular Windows XP and 32-bit Vista and a 64-bit version for more advanced systems like Windows XP x64, Windows Server and 64-bit Vista. You can get a fully functional 30-day trial version from www.passmark.com, but the purchase price is so reasonable that I suggest buying a long-term license so you can later test your system as needed.
Before testing your system, be sure to completely defragment the boot drive using a disk defragmenter like Diskeeper (www.diskeeper.com). A fragmented hard disk will result in unreliably low performance metrics and will not give you an accurate understanding of where your computer’s weaknesses lie. Similarly, prior to testing, be sure that you have run Windows Update so that your Windows operating system is fully up to date.
Performance Test includes many individual tests, but the most important ones are the combined scores near the end, particularly “CPU Mark,” “Memory Mark,” and “Disk Mark” for business users. Persons using the newest CS4 version of Adobe Photoshop and computer gamers will also be very interested in “2D Graphics Mark” and in “3D Graphics Mark.” I believe that these individual subsystem tests are more meaningful than an overall “PassMark Rating” for the tested computer. These tell you what subsystem most affects your overall performance. PassMark provides a number of comparative test results for standard models, allowing you to directly compare your system with other computers.
What you’re looking for is well-balanced computer performance that’s tuned to your individual needs. A user who mostly does word processing and e-mail does not need high performance in any particular area. For low-demand users, almost anything made within the past three years or so will be adequate. Business users who use Excel or other spreadsheets to crunch a lot of numbers will want very high CPU performance, particularly “CPU Integer Math,” “CPU Floating Point Math” and “CPU Find Prime Numbers.” Business users who store a lot of very large documents or Photoshop users will especially need good hard disk performance, shown by the “Disk Mark” aggregate score.
Now, put a CD in your CD reader, cancel any auto-start programs and run all Performance Test 6.1 tests as a single operation. To do this, choose the “Tests, Run All Tests” menu item. Print out the results, label them appropriately and move to the next phase of testing.
Using the “Windows Update, Custom, Hardware — Optional” selection, check to see whether there are any approved device driver updates for the software that meshes individual hardware to the overall Windows operating system. Install the updated device driver software. Reboot your computer and run Performance Test again, as above. Print out and label these results and see whether you notice any significant changes. Often, updating to a manufacturer’s latest device driver software from Microsoft’s necessarily generic drivers will provide a decent and free performance boost.
Not all driver software makes it to Microsoft’s Windows Update service. Sometimes, you’ll want to find something newer or perhaps a little more exotic, in which case you’ll need to go to the device manufacturer’s own Web site for the latest driver software. Be a little more careful here — sometimes device driver software is posted by a vendor before all of the bugs are worked out and buggy driver software will often crash a system upon the next boot-up. Personally, I would avoid device drivers that are not dated at least a month earlier.
It always makes sense to ensure that you can recover from a crashing system that’s become inherently unstable due to bad device drivers. The best way to do this is to ensure that your system always has Windows System Restore enabled. To check, go to “Control Panel, System, System Restore” and ensure that System Restore is turned on for all drives.
If your system should start crashing after any software or hardware upgrades, don’t just keep trying to reboot the system continuously. You’ll only corrupt the underlying Windows installation sooner or later. Instead, use the F8 key on the next reboot to go into “safe mode” and, rather than trying to boot normally, instead enter the “system restore” function and revert back to the prior-known good system restore point. It will take awhile for Windows to restore your system back to an earlier status, but after you reboot your system should be stable again.
After updating your operating system and device driver software, there’s another potential bottleneck to examine before throwing money at new hardware. Many programs install “helpful” bits that automatically start up as “processes” or “services,” and their individual or cumulative effect is to slow down your system, often remarkably so, or even make it unstable. In years past, among the biggest offenders, in my opinion, were earlier versions of Norton Utilities, which not only caused system instability but actually reduced overall performance by as much as 30 percent, the equivalent of going backwards one full CPU processor generation.
You’ll be able to check and see which processes are hogging CPU and disk performance by using Windows Task Manager, which is accessed by simultaneously hitting the three-key CNTRL-ALT-DELETE combination. When Task Master pops up, first check the Applications tab to see which programs are running. Close them to get a more accurate reading. Then, look at the Performance tab to see whether there are any unusual spikes in CPU, memory or hard disk usage. Finally, check the Processes tab to see what processes are running and how much CPU and memory resources they are using.
During this check, you should not have any application programs open and no routine system maintenance operations like disk defragmenting or virus checking under way. If necessary, wait until these housekeeping chores are completed. CPU and hard disk activity should be near zero during this rest state check. If they are not near zero, something unusual is happening that may be a wasteful use of computer resources.
While you’re at it, use the Performance and Process tabs in Windows Task Manager check to see which optional processes are using a lot of memory. This does not necessarily slow down your system unless the optional processes grab and hold on to so much RAM memory that your overall system does not have enough memory to run applications quickly without having to resort to much-slower hard disk swapping. Photo editing programs like Adobe Photoshop and Lightroom demand a lot of RAM memory, by the way, so check memory status while running these programs, as well. If you are potentially short on available RAM, then the least-expensive and best approach is to simply replace your existing RAM with a matched, two- or four-gigabyte pair for 32-bit systems and a matched, four-gigabyte pair for 64-bit operating systems.
The best way to test whether any optional boot-up software is dragging down your system is to use the Ace Utilities (www.acelogix.com) Optimize, Auto-Start Manager to unload all your optional start-up programs, reboot and then use Performance Test 6.1 as described above to measure the raw performance of your system without any secondary software. If there’s a major improvement, decide what you really need to run routinely and what is superfluous. Again use the Ace Auto-Start Manager to add any necessary programs, such as anti-virus software, back into the boot-up sequence.
Make one final performance test, compare it with your prior test results and decide whether there are any hardware bottlenecks that can be readily upgraded without a great deal of cost and hassle. Generally, the most cost-effective upgrades are faster hard disks, faster video cards and increasing the amount of RAM memory in your system. On the other hand, if your CPU Mark and Memory Mark measurements seem low, you’ll probably be better off simply replacing the entire system if you really need the extra performance.
Next week, we’ll talk about how to do some hardware upgrades and the potential pitfalls along the way.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
The best way to troubleshoot a slow computer is to individually test each subset for overall performance and see whether you have a bottleneck that can be easily and inexpensively upgraded. For this purpose, I like PassMark’s Performance Test 6.1, which comes in both a 32-bit version for regular Windows XP and 32-bit Vista and a 64-bit version for more advanced systems like Windows XP x64, Windows Server and 64-bit Vista. You can get a fully functional 30-day trial version from www.passmark.com, but the purchase price is so reasonable that I suggest buying a long-term license so you can later test your system as needed.
Before testing your system, be sure to completely defragment the boot drive using a disk defragmenter like Diskeeper (www.diskeeper.com). A fragmented hard disk will result in unreliably low performance metrics and will not give you an accurate understanding of where your computer’s weaknesses lie. Similarly, prior to testing, be sure that you have run Windows Update so that your Windows operating system is fully up to date.
Performance Test includes many individual tests, but the most important ones are the combined scores near the end, particularly “CPU Mark,” “Memory Mark,” and “Disk Mark” for business users. Persons using the newest CS4 version of Adobe Photoshop and computer gamers will also be very interested in “2D Graphics Mark” and in “3D Graphics Mark.” I believe that these individual subsystem tests are more meaningful than an overall “PassMark Rating” for the tested computer. These tell you what subsystem most affects your overall performance. PassMark provides a number of comparative test results for standard models, allowing you to directly compare your system with other computers.
What you’re looking for is well-balanced computer performance that’s tuned to your individual needs. A user who mostly does word processing and e-mail does not need high performance in any particular area. For low-demand users, almost anything made within the past three years or so will be adequate. Business users who use Excel or other spreadsheets to crunch a lot of numbers will want very high CPU performance, particularly “CPU Integer Math,” “CPU Floating Point Math” and “CPU Find Prime Numbers.” Business users who store a lot of very large documents or Photoshop users will especially need good hard disk performance, shown by the “Disk Mark” aggregate score.
Now, put a CD in your CD reader, cancel any auto-start programs and run all Performance Test 6.1 tests as a single operation. To do this, choose the “Tests, Run All Tests” menu item. Print out the results, label them appropriately and move to the next phase of testing.
Using the “Windows Update, Custom, Hardware — Optional” selection, check to see whether there are any approved device driver updates for the software that meshes individual hardware to the overall Windows operating system. Install the updated device driver software. Reboot your computer and run Performance Test again, as above. Print out and label these results and see whether you notice any significant changes. Often, updating to a manufacturer’s latest device driver software from Microsoft’s necessarily generic drivers will provide a decent and free performance boost.
Not all driver software makes it to Microsoft’s Windows Update service. Sometimes, you’ll want to find something newer or perhaps a little more exotic, in which case you’ll need to go to the device manufacturer’s own Web site for the latest driver software. Be a little more careful here — sometimes device driver software is posted by a vendor before all of the bugs are worked out and buggy driver software will often crash a system upon the next boot-up. Personally, I would avoid device drivers that are not dated at least a month earlier.
It always makes sense to ensure that you can recover from a crashing system that’s become inherently unstable due to bad device drivers. The best way to do this is to ensure that your system always has Windows System Restore enabled. To check, go to “Control Panel, System, System Restore” and ensure that System Restore is turned on for all drives.
If your system should start crashing after any software or hardware upgrades, don’t just keep trying to reboot the system continuously. You’ll only corrupt the underlying Windows installation sooner or later. Instead, use the F8 key on the next reboot to go into “safe mode” and, rather than trying to boot normally, instead enter the “system restore” function and revert back to the prior-known good system restore point. It will take awhile for Windows to restore your system back to an earlier status, but after you reboot your system should be stable again.
After updating your operating system and device driver software, there’s another potential bottleneck to examine before throwing money at new hardware. Many programs install “helpful” bits that automatically start up as “processes” or “services,” and their individual or cumulative effect is to slow down your system, often remarkably so, or even make it unstable. In years past, among the biggest offenders, in my opinion, were earlier versions of Norton Utilities, which not only caused system instability but actually reduced overall performance by as much as 30 percent, the equivalent of going backwards one full CPU processor generation.
You’ll be able to check and see which processes are hogging CPU and disk performance by using Windows Task Manager, which is accessed by simultaneously hitting the three-key CNTRL-ALT-DELETE combination. When Task Master pops up, first check the Applications tab to see which programs are running. Close them to get a more accurate reading. Then, look at the Performance tab to see whether there are any unusual spikes in CPU, memory or hard disk usage. Finally, check the Processes tab to see what processes are running and how much CPU and memory resources they are using.
During this check, you should not have any application programs open and no routine system maintenance operations like disk defragmenting or virus checking under way. If necessary, wait until these housekeeping chores are completed. CPU and hard disk activity should be near zero during this rest state check. If they are not near zero, something unusual is happening that may be a wasteful use of computer resources.
While you’re at it, use the Performance and Process tabs in Windows Task Manager check to see which optional processes are using a lot of memory. This does not necessarily slow down your system unless the optional processes grab and hold on to so much RAM memory that your overall system does not have enough memory to run applications quickly without having to resort to much-slower hard disk swapping. Photo editing programs like Adobe Photoshop and Lightroom demand a lot of RAM memory, by the way, so check memory status while running these programs, as well. If you are potentially short on available RAM, then the least-expensive and best approach is to simply replace your existing RAM with a matched, two- or four-gigabyte pair for 32-bit systems and a matched, four-gigabyte pair for 64-bit operating systems.
The best way to test whether any optional boot-up software is dragging down your system is to use the Ace Utilities (www.acelogix.com) Optimize, Auto-Start Manager to unload all your optional start-up programs, reboot and then use Performance Test 6.1 as described above to measure the raw performance of your system without any secondary software. If there’s a major improvement, decide what you really need to run routinely and what is superfluous. Again use the Ace Auto-Start Manager to add any necessary programs, such as anti-virus software, back into the boot-up sequence.
Make one final performance test, compare it with your prior test results and decide whether there are any hardware bottlenecks that can be readily upgraded without a great deal of cost and hassle. Generally, the most cost-effective upgrades are faster hard disks, faster video cards and increasing the amount of RAM memory in your system. On the other hand, if your CPU Mark and Memory Mark measurements seem low, you’ll probably be better off simply replacing the entire system if you really need the extra performance.
Next week, we’ll talk about how to do some hardware upgrades and the potential pitfalls along the way.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
Tuesday, March 10, 2009
Plugged In: Printers — A big decision for small businesses
Large-format printers are a breed apart. They’re physically larger, with a higher initial purchase price but lower ink costs, and can produce very high-quality prints at least 13-by-19 or larger.
All affordable large-format printers, even those designed for professionals like engineers and graphic designers, are based upon inkjet technology.
Only a few vendors make consumer-grade, large-format printers. Should you consider a large-format printer? A surprisingly large number of businesses might find one useful, even if you don’t need photographic-quality printing.
For example, most real property documents, such as aerial photos, plats and surveys, and most construction plans are intended to be printed at 11-by-17 or larger, and smaller copies are hard to read and use.
Most home and small-business printers can print letter-sized documents and photographs with varying degrees of quality. Such printers are usually relatively inexpensive to purchase but have very high ink costs per print.
For the routine printing needs of a small business, an inexpensive laser printer makes the most sense. Laser printers are faster and less expensive per page compared to low-end inkjet printers, but laser printers have two glaring weaknesses.
Even fairly expensive laser printers are limited to prints not wider than standard 8.5-inch-width letter and legal paper, although an office store like Soldotna’s UPS store can often print the occasional larger image, for a price. Secondly, laser printers do not do a very good job printing color images, especially photographs.
If you want to make high-quality, large, color images at home or in your business, then you’ll need a large-format printer.
There are several good models currently offered in the $200 to $800 range. Unfortunately, since the bankruptcy of CompUSA last year, you can’t find any of these products on Southcentral store shelves for immediate purchase. You’ll need to buy them from a reliable Internet vendor like www.amazon.com or www.newegg.com.
Consumer-grade large format printers usually make prints up to 13-by-19. Somewhat more expensive printers can use 17- or 18-inch-wide paper while professional models can handle 24-inch or wider rolls of high-end paper.
Only three vendors, HP, Canon and Epson, sell readily available, consumer-grade, large-format printers. All of them produce quality systems but the right choice depends upon your needs.
HP pioneered several of the most important printing technologies, including laser printing and photo-grade inkjets. For many years, HP basically owned the market for business printers.
Epson focused upon very high-quality photographic printing. Canon put out excellent models in both of these areas but tended to be underappreciated.
There are a few basic printing concepts that are useful in determining which printer is best for you. Generally, printers with a larger number of separate ink colors produce more natural-looking photographs and detailed color images.
At least six different ink colors, including light magenta and light cyan, are needed for high-quality results. Printers that use larger (27-milliliter or more) separate cartridges for each color are much more economical to use. If you want to make high-grade, black-and-white prints for exhibition, then you’ll need a printer with at least two, preferably three or four, shades of black, gray, light gray, and light, light gray. Almost all printers now use the USB 2.0 computer interface.
Paper handling is important, too, because paper jams and scratched prints are a true annoyance. Printers with a straight-through rear paper feed, like that used by Canon and Epson, are usually more reliable. HP’s tray and folded paper path approach is usually more troublesome when making larger prints and photographs.
Using rolls of paper with an automatic roll feeder attached to the printer is much more reliable and far more economical in the long run, sometimes saving as much as 75 percent to 80 percent in paper costs.
There are two general types of high-grade printing inks: dye-based and pigment-based. Pigment inkjets are generally usable on many different brands of photo and fine art paper and are considered to be less prone to fading.
Dye inkjets are a slightly older technology that often can provide more brilliant colors but generally must be used with certain types of paper specified by the printer’s manufacturer. Dye-ink prints are more vulnerable to moisture damage and thus must be handled with greater care. Except for HP’s archivally rated Vivera inks, dye-ink images are not very resistant to fading over the years.
However, in comparison, traditional color prints made from film usually fade more quickly than dye inks.
For consistent and economical results, the entire printing system from monitor through printer should be regularly checked, calibrated and “color managed” so that what you see on the monitor is what you get on the print. Printers should be used, or at least cleaned and calibrated, regularly to avoid problems.
Be sure that you know what you’re getting into with a larger-format printer. They’re big, heavy and need a lot of tabletop space, with sufficient room to allow top loading of 19-inch or larger cut paper. Getting consistently high-quality prints and good economy from large-format printers will require some study and effort on your part.
Unlike cameras, whose models usually change from year to year, specific printer models are usually marketed without significant changes for two to four years. High-grade inkjet printing is a fairly mature technology and improvements from model to model are usually incremental, rather than revolutionary.
Be careful when purchasing: The real cost for these printers is in the ink and many of them ship with low-capacity, quickly exhausted “starter” cartridges. Before you know it, you’ll be spending another few hundred dollars for real ink cartridges. Such practices are something of a scam that vendors occasionally do with some large-format printers.
The following models are among the more common large-format printers, warts and all. All prices were current on Amazon.com as of March 2. There were no new printers of consequence introduced at the annual photo marketing show, PMA, last week, so printers listed here should be current through at least early summer. There is no one perfect choice, so shop carefully and know your own needs and technical limitations. Replacement ink sets for most printers listed here can be ordered through Stub’s Office Supply in Soldotna.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
All affordable large-format printers, even those designed for professionals like engineers and graphic designers, are based upon inkjet technology.
Only a few vendors make consumer-grade, large-format printers. Should you consider a large-format printer? A surprisingly large number of businesses might find one useful, even if you don’t need photographic-quality printing.
For example, most real property documents, such as aerial photos, plats and surveys, and most construction plans are intended to be printed at 11-by-17 or larger, and smaller copies are hard to read and use.
Most home and small-business printers can print letter-sized documents and photographs with varying degrees of quality. Such printers are usually relatively inexpensive to purchase but have very high ink costs per print.
For the routine printing needs of a small business, an inexpensive laser printer makes the most sense. Laser printers are faster and less expensive per page compared to low-end inkjet printers, but laser printers have two glaring weaknesses.
Even fairly expensive laser printers are limited to prints not wider than standard 8.5-inch-width letter and legal paper, although an office store like Soldotna’s UPS store can often print the occasional larger image, for a price. Secondly, laser printers do not do a very good job printing color images, especially photographs.
If you want to make high-quality, large, color images at home or in your business, then you’ll need a large-format printer.
There are several good models currently offered in the $200 to $800 range. Unfortunately, since the bankruptcy of CompUSA last year, you can’t find any of these products on Southcentral store shelves for immediate purchase. You’ll need to buy them from a reliable Internet vendor like www.amazon.com or www.newegg.com.
Consumer-grade large format printers usually make prints up to 13-by-19. Somewhat more expensive printers can use 17- or 18-inch-wide paper while professional models can handle 24-inch or wider rolls of high-end paper.
Only three vendors, HP, Canon and Epson, sell readily available, consumer-grade, large-format printers. All of them produce quality systems but the right choice depends upon your needs.
HP pioneered several of the most important printing technologies, including laser printing and photo-grade inkjets. For many years, HP basically owned the market for business printers.
Epson focused upon very high-quality photographic printing. Canon put out excellent models in both of these areas but tended to be underappreciated.
There are a few basic printing concepts that are useful in determining which printer is best for you. Generally, printers with a larger number of separate ink colors produce more natural-looking photographs and detailed color images.
At least six different ink colors, including light magenta and light cyan, are needed for high-quality results. Printers that use larger (27-milliliter or more) separate cartridges for each color are much more economical to use. If you want to make high-grade, black-and-white prints for exhibition, then you’ll need a printer with at least two, preferably three or four, shades of black, gray, light gray, and light, light gray. Almost all printers now use the USB 2.0 computer interface.
Paper handling is important, too, because paper jams and scratched prints are a true annoyance. Printers with a straight-through rear paper feed, like that used by Canon and Epson, are usually more reliable. HP’s tray and folded paper path approach is usually more troublesome when making larger prints and photographs.
Using rolls of paper with an automatic roll feeder attached to the printer is much more reliable and far more economical in the long run, sometimes saving as much as 75 percent to 80 percent in paper costs.
There are two general types of high-grade printing inks: dye-based and pigment-based. Pigment inkjets are generally usable on many different brands of photo and fine art paper and are considered to be less prone to fading.
Dye inkjets are a slightly older technology that often can provide more brilliant colors but generally must be used with certain types of paper specified by the printer’s manufacturer. Dye-ink prints are more vulnerable to moisture damage and thus must be handled with greater care. Except for HP’s archivally rated Vivera inks, dye-ink images are not very resistant to fading over the years.
However, in comparison, traditional color prints made from film usually fade more quickly than dye inks.
For consistent and economical results, the entire printing system from monitor through printer should be regularly checked, calibrated and “color managed” so that what you see on the monitor is what you get on the print. Printers should be used, or at least cleaned and calibrated, regularly to avoid problems.
Be sure that you know what you’re getting into with a larger-format printer. They’re big, heavy and need a lot of tabletop space, with sufficient room to allow top loading of 19-inch or larger cut paper. Getting consistently high-quality prints and good economy from large-format printers will require some study and effort on your part.
Unlike cameras, whose models usually change from year to year, specific printer models are usually marketed without significant changes for two to four years. High-grade inkjet printing is a fairly mature technology and improvements from model to model are usually incremental, rather than revolutionary.
Be careful when purchasing: The real cost for these printers is in the ink and many of them ship with low-capacity, quickly exhausted “starter” cartridges. Before you know it, you’ll be spending another few hundred dollars for real ink cartridges. Such practices are something of a scam that vendors occasionally do with some large-format printers.
The following models are among the more common large-format printers, warts and all. All prices were current on Amazon.com as of March 2. There were no new printers of consequence introduced at the annual photo marketing show, PMA, last week, so printers listed here should be current through at least early summer. There is no one perfect choice, so shop carefully and know your own needs and technical limitations. Replacement ink sets for most printers listed here can be ordered through Stub’s Office Supply in Soldotna.
Canon:
- Canon Pixma 9000 ($410). This is an eight-dye ink printer that is excellent for color photos up to 13-by-19. Because it does not have a second gray cartridge, it’s less useful for exhibition-grade, black-and-white prints. I own the predecessor to this model and it has been fast, reliable and easy to use, so long as you use Canon’s own printer software to control the look of final output. This would be an excellent choice for most users. My only complaint is that Canon’s small ink cartridges increase the price per print.
- Canon Pixma Pro9500 ($760). The Pixma Pro 9500 is a 10-pigment ink, 13-by-19 printer that includes several levels of black. This printer has been criticized for very slow output speed but the output is very nice. Given the choice in this price range, though, I would get the Epson 2880, especially if doing a lot of black and white.
Epson:
- Epson Stylus Photo 1400 ($214). This is the least-expensive, large-format, 13-by-19 printer from a major vendor. It uses six dye inks and is a good choice for a low-cost, general-use business and photo large-format printer for color images. It is not intended to print exhibit-grade, black-and-white images. Low-capacity ink cartridges significantly increase the cost per page.
- Epson Stylus R1900 ($549). The R1900 is a 13-by-19 printer using seven pigment inks plus a gloss optimizer. It’s basically designed to produce high-quality, glossy photos and includes a manual roll feed that allows you to print longer panoramic photos while reducing your paper costs. Similar to the Epson 1400, the R1900 does not have an intermediate gray level ink, reducing its suitability for fine art and black-and-white photos. Low-capacity ink cartridges significantly increase the cost per page.
- Epson Stylus Pro 2880 ($735) includes the full Epson K3 eight-pigment photo ink set, including three levels of black and gray. Its output is considered to be the best of any printer under $1,000. If you need to do occasional 13-by-19, exhibit-grade photos, including black-and-white images, then this is the printer for you. However, it has been criticized because its expensive, low-capacity ink cartridges increase the cost per page.
- Epson Stylus Pro 3800 ($1,095) is a 17-by- 22 printer that, although a 3-year-old design, is still considered to be the best all-around fine art printer in its size and price range. The 3800 is a large printer using much more economical, high-capacity ink sets that include three levels of black/gray for optimum black-and-white printing. Unfortunately, it does not have a roll feed option but paper handling is considered to be quite good. Output quality is excellent, but this is really intended as a high-quality, fine-art printer rather, than a production business machine.
HP:
- HP 8350 ($290) is HP’s least expensive, 13-by-19 large-format printer that includes a photo gray option for better black-and-white images but it has mixed customer reviews.
- HP B8550 ($304) This printers uses five-dye inks, prints up to 13-by-19, and has decent customer reviews. Ink is a little less expensive than equivalent Canon and Epson printers.
- HP B8850 ($472) is a 13-by-19 printer that uses eight-pigment inks and includes a straight rear paper path, although HP does not include the very useful fold-down rear paper loading trays included with the Canon printers. This is a somewhat newer, slightly stripped-down version of the highly regarded HP B9180 professional printer. It has received excellent reviews and would be a very good choice for an amateur photographer intending to print occasional exhibit-quality photos. It uses higher-capacity ink cartridges for better economy.
- HP B9180 ($594) is another 13-by-19 printer quite similar to the B8850 but includes an LCD panel and network connection. This is also a highly regarded choice for a semiprofessional photographer, with excellent quality color and black-and-white output, but the B9180’s paper handling has been criticized on occasion.
- HP DesignJet 90 (Q6656A, under $1,000, Q6656B with roll feed, list $1,150) is an interesting printer that includes six long-life, Vivera dye inks. This is the least expensive printer for 18-inch wide paper. I have the 24-inch DesignJet 130 version with roll feed. If you get the roll feed version and use HP’s Professional Plus papers, you will find this to be a very economical 18-inch-wide printer that’s able to produce exhibition-grade color images with a rated 82-year archival life. Because it does not include a second black/gray ink, it is not as useful for black-and-white images.
- DJ 110Plus, ($995). This printer uses four Vivera dye inks and prints up to 24 inches wide. It’s useful for making big posters and signs in a business setting, but not for high-quality work. It uses economical, high-capacity ink cartridges.
- DJ 130 ($1,150 but also get $450 roll feed for best reliability and economy). I use this 24-inch printer regularly to make very high-color images for exhibit. In fact, I used it so regularly that I finally wore one out. It’s the DesignJet 90’s big brother and the most economical all-around large-format choice for the color photographer or business requiring high-quality color images up to 24 inches wide. Because it does not include a second gray/black ink, black-and-white image quality is decent but not spectacular. It uses economical, high-capacity ink cartridges.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
Tuesday, February 24, 2009
Plugged In: To print, or not to print
Editor’s Note: The following column contains puns. Be forewarned.
Whether ’tis nobler to suffer the slings and arrows of outrageous fortune-gouging, or to take arms against a sea of paper, and end it?
All mock seriousness aside, whether to print a document, and the best way to do so, is a question to which there are a variety of right, or nearly right, answers. Actually, I was tempted to say “write answers,” but I’ll forego that homonymic bad pun. Really. One needs to be restrained in these printed pages and, as a modest word-wright, show the Write Stuff.
Paper documents remain highly useful, even though more than 95 percent of all business documents are electronic, and e-mails with grandparents are now more common than scrawled thank-you notes. Paper handouts are usually a more effective way to conduct meetings. A few sheets of paper can be folded and carried easily in a pant or jacket pocket. Printed documents can be read by any literate person when there’s no computer around and are often a more effective way to present data and reasoning, particularly in complex matters — for example, note that the homonymic puns in the first two paragraphs work only in written form. Courts require paper exhibits, even if they are photocopies. A high-quality, fine-art photo enlargement is almost always preferable to photos displayed on a computer screen.
In my opinion, though, paper is no longer the data recording, archival and filing medium of choice. Paper is easily misfiled, cannot be easily searched and is expensive to reproduce and store off-premises in case of a disaster. Losing your business records, especially accounting records, to any sort of casualty like fire, flood or storm, is tantamount to ultimately losing your business. That’s happened to several of my business clients over the years.
I believe that the best overall approach is to preserve documents electronically in a standard, easily searchable format, and to print paper copies only when needed. This approach has several benefits. It maximizes your business efficiency and effectiveness. It minimizes storage and filing costs, is much more economical, and is more ecologically sound. It’s also more convenient, especially when you can e-mail signed copies in seconds and enable electronic marginal notes and comments by readers, who can then electronically return their comments to you with a few mouse clicks.
Let’s first look at the more modern approach to “printing” electronically. Adobe Acrobat PDF files are already the de facto standard for the federal government, most state and local governments and businesses generally. Thus, printing a document to Adobe’s PDF makes the most sense as a long-term medium for storing data electronically in a way that’s easy to search, back up and protect.
It’s always been important to store data in an open data file format like Acrobat’s PDF, and to avoid long-term storage of data in proprietary data file formats. It’s unwise to trust that most niche vendors, or their file formats, will be around next year, or that the data will be usable with another program. That’s especially true in tough economic times.
Acrobat documents can be “printed” to a standard format electronic file directly from digital data stored on your computer, such as e-mail, Web browsers, spreadsheets, word processing programs or photographic programs. Basically, printing data to a PDF file costs you nothing in supplies. Unlike printing paper documents, “printing” to an electronic PDF file is essentially without any cost — it’s only some electrons being moved around the computer and hard disk.
If you already have paper documents that you wish to preserve electronically, then you can scan them directly into Acrobat using a wide variety of flat image scanners and document scanners that can feed and scan many pages a minute. I’ll discuss scanning in the near future.
Adobe has recently provided a new, archival version of Acrobat, PDF/X, that should be suitable for long-term data storage so long as you take care to ensure that the physical storage (hard disks, portable USB flash drives, CD disks, etc.) are in good working order and that you regularly transfer the electronic files to newer types of data storage hardware.
In order to make Acrobat PDF documents easy to search across a large hard disk, you’ll need to run the Optical Character Recognition (OCR) process so that the internal contents become readable to search programs and to Acrobat itself. I like Copernic’s Desktop Search because commercial versions can search an entire network, rather than your computer alone. Google’s desktop search program also has its enthusiasts. Downloads of each are free and worth trying.
Even if you are committed to storing all of your data electronically, there will be many times when you will want, or need, paper copies. That obviously requires a printer. There are several types of printers in common use: color laser printers, monochrome laser printers, ordinary inkjet printers and photo-grade inkjet printers. Each has unique uses, advantages and disadvantages.
Inexpensive, compact home printers are usually inkjets, often part of “multifunction” devices that include a slow scanner, light copying functions and perhaps some fax capability. These are typically inexpensive to purchase, are suitable for light home and home office use but are too slow, and insufficiently robust, for heavier business use.
Because smaller inkjet printers use very low-capacity but expensive ink cartridges, your ink cost is the killer. Light-duty inkjet printers are like the reputed Gillette razors of years past — the razor itself is basically given away to induce you to buy only Gillette razor blades and at a pretty high unit cost. As with razor blades, the real profit for inkjet vendors is in the supplies.
Some midrange consumer inkjets, such as HP’s Photosmart series, the Kodak ESP-7 and ESP-9, or some Canon or Epson multifunction devices, can do a very creditable job printing lab-quality photographs up to 8.5-by-11. If you don’t care about photo-quality printing, then almost any inexpensive multifunction device will be suitable for home use.
Businesses should generally consider getting a laser printer. Laser output is typically faster, looks better and is more water-resistant than inkjets. If you do not believe that you will ever need color output, then a monochrome (black and white) laser printer will be sufficient. Low-end monochrome laser printers tend to be fairly inexpensive, while upper-end ones tend to be quite fast. I have tried several brands, but HP LaserJets have always proven to be the most reliable, and the HP dealer in Anchorage, Lewis and Lewis, has been very good to work with.
Realistically, though, color laser output is becoming the norm and it is quite useful. Color laser printers are often more convenient. Most color laser printers run slower than advertised, so don’t buy an inexpensive one that claims high output speed and then expect quick results. I have had several Lexmark color laser printers costing less than $1,000 and I have been disappointed.
I have, as a result, reverted to HP’s slow Color LaserJet 2,600 series. These printers do a quality job but are suitable only for fairly low volume, despite being advertised as a printer sufficiently fast to service an entire small business. I found that was not the case and purchased an excellent Konica-Minolta 5,670 printer. This series is very fast and the quality is adequate, although not quite as good, in my opinion, as the photo output from my HP LaserJet 2,605. Konica-Minolta printers are sold locally at Frontier Business Systems and Hi-Speed Gear. Konica-Minolta also makes some less-expensive, slightly slower color laser printers that I found to be good values.
Color laser printers are generally limited to 8.5-by-11 output, and their photographic print quality is not very good compared to good inkjet printers. If you want any sort of printout larger than 8.5-by-11, particularly photographic work, then you’ll need to buy a large-format inkjet printer. Large-format photo printers deserve an entire article, and we’ll do just that next week.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
Whether ’tis nobler to suffer the slings and arrows of outrageous fortune-gouging, or to take arms against a sea of paper, and end it?
All mock seriousness aside, whether to print a document, and the best way to do so, is a question to which there are a variety of right, or nearly right, answers. Actually, I was tempted to say “write answers,” but I’ll forego that homonymic bad pun. Really. One needs to be restrained in these printed pages and, as a modest word-wright, show the Write Stuff.
Paper documents remain highly useful, even though more than 95 percent of all business documents are electronic, and e-mails with grandparents are now more common than scrawled thank-you notes. Paper handouts are usually a more effective way to conduct meetings. A few sheets of paper can be folded and carried easily in a pant or jacket pocket. Printed documents can be read by any literate person when there’s no computer around and are often a more effective way to present data and reasoning, particularly in complex matters — for example, note that the homonymic puns in the first two paragraphs work only in written form. Courts require paper exhibits, even if they are photocopies. A high-quality, fine-art photo enlargement is almost always preferable to photos displayed on a computer screen.
In my opinion, though, paper is no longer the data recording, archival and filing medium of choice. Paper is easily misfiled, cannot be easily searched and is expensive to reproduce and store off-premises in case of a disaster. Losing your business records, especially accounting records, to any sort of casualty like fire, flood or storm, is tantamount to ultimately losing your business. That’s happened to several of my business clients over the years.
I believe that the best overall approach is to preserve documents electronically in a standard, easily searchable format, and to print paper copies only when needed. This approach has several benefits. It maximizes your business efficiency and effectiveness. It minimizes storage and filing costs, is much more economical, and is more ecologically sound. It’s also more convenient, especially when you can e-mail signed copies in seconds and enable electronic marginal notes and comments by readers, who can then electronically return their comments to you with a few mouse clicks.
Let’s first look at the more modern approach to “printing” electronically. Adobe Acrobat PDF files are already the de facto standard for the federal government, most state and local governments and businesses generally. Thus, printing a document to Adobe’s PDF makes the most sense as a long-term medium for storing data electronically in a way that’s easy to search, back up and protect.
It’s always been important to store data in an open data file format like Acrobat’s PDF, and to avoid long-term storage of data in proprietary data file formats. It’s unwise to trust that most niche vendors, or their file formats, will be around next year, or that the data will be usable with another program. That’s especially true in tough economic times.
Acrobat documents can be “printed” to a standard format electronic file directly from digital data stored on your computer, such as e-mail, Web browsers, spreadsheets, word processing programs or photographic programs. Basically, printing data to a PDF file costs you nothing in supplies. Unlike printing paper documents, “printing” to an electronic PDF file is essentially without any cost — it’s only some electrons being moved around the computer and hard disk.
If you already have paper documents that you wish to preserve electronically, then you can scan them directly into Acrobat using a wide variety of flat image scanners and document scanners that can feed and scan many pages a minute. I’ll discuss scanning in the near future.
Adobe has recently provided a new, archival version of Acrobat, PDF/X, that should be suitable for long-term data storage so long as you take care to ensure that the physical storage (hard disks, portable USB flash drives, CD disks, etc.) are in good working order and that you regularly transfer the electronic files to newer types of data storage hardware.
In order to make Acrobat PDF documents easy to search across a large hard disk, you’ll need to run the Optical Character Recognition (OCR) process so that the internal contents become readable to search programs and to Acrobat itself. I like Copernic’s Desktop Search because commercial versions can search an entire network, rather than your computer alone. Google’s desktop search program also has its enthusiasts. Downloads of each are free and worth trying.
Even if you are committed to storing all of your data electronically, there will be many times when you will want, or need, paper copies. That obviously requires a printer. There are several types of printers in common use: color laser printers, monochrome laser printers, ordinary inkjet printers and photo-grade inkjet printers. Each has unique uses, advantages and disadvantages.
Inexpensive, compact home printers are usually inkjets, often part of “multifunction” devices that include a slow scanner, light copying functions and perhaps some fax capability. These are typically inexpensive to purchase, are suitable for light home and home office use but are too slow, and insufficiently robust, for heavier business use.
Because smaller inkjet printers use very low-capacity but expensive ink cartridges, your ink cost is the killer. Light-duty inkjet printers are like the reputed Gillette razors of years past — the razor itself is basically given away to induce you to buy only Gillette razor blades and at a pretty high unit cost. As with razor blades, the real profit for inkjet vendors is in the supplies.
Some midrange consumer inkjets, such as HP’s Photosmart series, the Kodak ESP-7 and ESP-9, or some Canon or Epson multifunction devices, can do a very creditable job printing lab-quality photographs up to 8.5-by-11. If you don’t care about photo-quality printing, then almost any inexpensive multifunction device will be suitable for home use.
Businesses should generally consider getting a laser printer. Laser output is typically faster, looks better and is more water-resistant than inkjets. If you do not believe that you will ever need color output, then a monochrome (black and white) laser printer will be sufficient. Low-end monochrome laser printers tend to be fairly inexpensive, while upper-end ones tend to be quite fast. I have tried several brands, but HP LaserJets have always proven to be the most reliable, and the HP dealer in Anchorage, Lewis and Lewis, has been very good to work with.
Realistically, though, color laser output is becoming the norm and it is quite useful. Color laser printers are often more convenient. Most color laser printers run slower than advertised, so don’t buy an inexpensive one that claims high output speed and then expect quick results. I have had several Lexmark color laser printers costing less than $1,000 and I have been disappointed.
I have, as a result, reverted to HP’s slow Color LaserJet 2,600 series. These printers do a quality job but are suitable only for fairly low volume, despite being advertised as a printer sufficiently fast to service an entire small business. I found that was not the case and purchased an excellent Konica-Minolta 5,670 printer. This series is very fast and the quality is adequate, although not quite as good, in my opinion, as the photo output from my HP LaserJet 2,605. Konica-Minolta printers are sold locally at Frontier Business Systems and Hi-Speed Gear. Konica-Minolta also makes some less-expensive, slightly slower color laser printers that I found to be good values.
Color laser printers are generally limited to 8.5-by-11 output, and their photographic print quality is not very good compared to good inkjet printers. If you want any sort of printout larger than 8.5-by-11, particularly photographic work, then you’ll need to buy a large-format inkjet printer. Large-format photo printers deserve an entire article, and we’ll do just that next week.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com.
Tuesday, February 17, 2009
Plugged In: Get video carded for better digital photo use
Digital photography and desktop computing are fraternal twins. They’re not exactly alike but have a lot in common.
Decent digital photography requires some pretty heavy-duty computing and desktop printing power. At the same time, digital photography and videography add a great deal of clarity, power and persuasiveness to presentations and written documents.
Readers of previous columns will recall my view that getting a really fast, expensive video card is an unnecessary expense for most consumers. However, if you plan to use your computer for digital photography, then you should consider getting a Photoshop-compatible video card with at least 512 megabytes of memory installed directly on the video card. Digital photographers should avoid video cards that share memory with the overall computer and operating system.
Here’s why a video card is becoming so important to serious digital photographers and technical users, like engineers. Processing digital photographs, engineering drawings and other technical files demands a great deal of graphics manipulation and computing power.
A general-purpose CPU is not optimized for the sort of intense graphical processing required to process these high megapixel digital images. Even though the several most recent versions of Adobe’s industry standard digital photo programs — Photoshop, Photoshop Elements and Photoshop Lightroom — can recognize and use dual or quad core CPUs, they still tend to bog down when processing big images and large technical files. Upgrading system memory to at least 2 gigabytes DRAM and installing a very fast hard disk and a quad core processor helps reduce long processing times, but it’s still not quite enough.
There is another solution that happens to be fairly easy and inexpensive to implement. Starting with the very recent Adobe Photoshop CS4, Adobe now uses the Graphics Processing Unit (GPU) found on high-end video cards to reduce Photoshop’s processing times. GPUs are optimized for precisely the sorts of demands made by Photoshop. Recent high-end GPUs are actually very fast computer cores that specialize in graphics manipulation, where general-purpose CPU processors fall behind.
Replacing an older video card with a new, higher-end video processor is not very difficult or expensive. It’s probably one of the best ways to upgrade a computer used for digital imaging.
Remember, though, that GPU processing only works with the most recent versions of Photoshop CS4, Photoshop Elements 7 and Lightroom 2.2. Older versions of Photoshop programs will not see any improvement. GPU processing of Photoshop works with Mac OS 10.5.4 and 10.4.11, 64-bit versions of Windows Vista and 32-bit versions of Windows XP. The 64-bit version of Windows XP x64 is not supported. In addition, you’ll need to obtain and install the most recent Open GL 2.0 driver software for your new video card in Photoshop.
Not all video cards are suitable, though. To be supported, a video card can be any brand that is built around one of the supported GPU chipsets listed below, and the card must support Open GL 2.0. A current listing of supported GPU chipsets can be found by searching www.adobe.com. We recommend that you download the latest GPU chipset driver software from ATI and nVidia at these links: ATI video cards: www.ati.amd.com/support/driver.html ; nVidia video cards: www.nvidia.com/Download/index.aspx?lang=en-us.
Among the more common, high-end video GPU chipsets supported by Photoshop CS4 are:
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com, along with links to legal and community resources.
Decent digital photography requires some pretty heavy-duty computing and desktop printing power. At the same time, digital photography and videography add a great deal of clarity, power and persuasiveness to presentations and written documents.
Readers of previous columns will recall my view that getting a really fast, expensive video card is an unnecessary expense for most consumers. However, if you plan to use your computer for digital photography, then you should consider getting a Photoshop-compatible video card with at least 512 megabytes of memory installed directly on the video card. Digital photographers should avoid video cards that share memory with the overall computer and operating system.
Here’s why a video card is becoming so important to serious digital photographers and technical users, like engineers. Processing digital photographs, engineering drawings and other technical files demands a great deal of graphics manipulation and computing power.
A general-purpose CPU is not optimized for the sort of intense graphical processing required to process these high megapixel digital images. Even though the several most recent versions of Adobe’s industry standard digital photo programs — Photoshop, Photoshop Elements and Photoshop Lightroom — can recognize and use dual or quad core CPUs, they still tend to bog down when processing big images and large technical files. Upgrading system memory to at least 2 gigabytes DRAM and installing a very fast hard disk and a quad core processor helps reduce long processing times, but it’s still not quite enough.
There is another solution that happens to be fairly easy and inexpensive to implement. Starting with the very recent Adobe Photoshop CS4, Adobe now uses the Graphics Processing Unit (GPU) found on high-end video cards to reduce Photoshop’s processing times. GPUs are optimized for precisely the sorts of demands made by Photoshop. Recent high-end GPUs are actually very fast computer cores that specialize in graphics manipulation, where general-purpose CPU processors fall behind.
Replacing an older video card with a new, higher-end video processor is not very difficult or expensive. It’s probably one of the best ways to upgrade a computer used for digital imaging.
Remember, though, that GPU processing only works with the most recent versions of Photoshop CS4, Photoshop Elements 7 and Lightroom 2.2. Older versions of Photoshop programs will not see any improvement. GPU processing of Photoshop works with Mac OS 10.5.4 and 10.4.11, 64-bit versions of Windows Vista and 32-bit versions of Windows XP. The 64-bit version of Windows XP x64 is not supported. In addition, you’ll need to obtain and install the most recent Open GL 2.0 driver software for your new video card in Photoshop.
Not all video cards are suitable, though. To be supported, a video card can be any brand that is built around one of the supported GPU chipsets listed below, and the card must support Open GL 2.0. A current listing of supported GPU chipsets can be found by searching www.adobe.com. We recommend that you download the latest GPU chipset driver software from ATI and nVidia at these links: ATI video cards: www.ati.amd.com/support/driver.html ; nVidia video cards: www.nvidia.com/Download/index.aspx?lang=en-us.
Among the more common, high-end video GPU chipsets supported by Photoshop CS4 are:
- nVidia geForce: nVidia 260 GTX 896MB, nVidia 9800 GTX 512MB, nVidia 9600 GT 512MB, nVidia 8800 GTX 768MB, nVidia 8800 GT 512MB, nVidia 8600M 256MB, nVidia 7900 GTX 512MB, nVidia 7900 GS 256MB, nVidia 7800 GTX 256MB, nVidia 7600 256MB, nVidia 6800 256MB.
- nVidia Quadro: Quadro FX 4600 768 MB, Quadro FX 4500 512MB, Quadro FX 4400 512MB, Quadro FX 3700 512MB, Quadro FX 3500 256MB, Quadro FX 1700 512MB, Quadro FX 1500 256MB, Quadro FX 1400 -128MB, Quadro FX 570 -256MB, Quadro FX 370 256MB.
- ATI Radeon: ATI Radeon 4850 512 MB, Radeon x3870 x2 512MB, Radeon x2900HD 512MB, ATI Radeon HD 2400 256MB, Radeon x1900XT 512MB, ATI Radeon x1800 - 512MB, Radeon x1800 256MB.
- ATI Fire GL: ATI Fire GL 7700 512 MB, ATI Fire GL 7600 512 MB, Fire GL V7200 256MB, ATI Fire GL 5600 512MB, Fire GL V3600 256MB, Fire GL 3350 256 MB.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com, along with links to legal and community resources.
Tuesday, February 10, 2009
Plugged In: Upgrading computers doesn’t have to be costly
You don’t need to spend yourself into bankruptcy to acquire effective computer hardware. The key to smart business automation is careful research and long-term planning.
In tough times, you need solutions that increase your efficiency and your effectiveness, at a smart price. One of the best ways to improve your profitability and viability is to make smart decisions about investing in new technology and to then crisply implement your decisions.
This week, as a wrap-up of our recent computing hardware discussions, I’ll address some more general business and planning considerations that business users should take into account when purchasing new PC computer automation.
Compared to other steps that you can take in your business, prudent investment in technology can produce comparatively large benefits. Small- to medium-sized businesses are a particularly good fit for Michael Porter’s business strategy, developed at Harvard Business School, that focuses upon optimum profitability and survivability in tough times through developing low-volume, high-margin work. Smart technology implementation often is key to developing that sort of resilient business because it allows you to both improve your responsiveness and effectiveness while simultaneously reducing your base overhead costs.
Even though simple approaches and programs often provide the best return for your automation investment and are usually the easiest to implement and use, businesses commonly throw a lot of immature “bleeding-edge” technology, expensive hardware and even more expensive consultant fees and staff time at what is really a business management problem.
Gradually implementing a well-thought-out strategic business automation plan is usually the most effective and efficient way to transform your business, buying only the hardware and software that you will be able to install and begin using within the next two months or so and then implementing your next phase.
Purchasing binges are inefficient. By the time that you get around to installing some of your purchases, you likely can buy a newer, better version of the same product for less money.
Choose your technology for long-term, low-cost usefulness. Buy mature mainstream technology wherever possible and avoid both dead-end and bleeding-edge hardware and software. You’ll not only save money by purchasing mature, proven software and hardware, but likely also spare yourself an expensive, frustrating experience installing and using immature technology.
The hidden costs in terms of your time, staff time and general disruption when you install new computer technology usually costs a business much more than the actual purchase price of the hardware and software. Keep these less-obvious costs in mind when deciding what technology to purchase. It may be less costly in the end for a business to buy something that’s initially more expensive but ultimately more easily installed, maintained and used.
Use the least complex technology that efficiently does the job for you, provided that it’s supported by a financially stable vendor and has reasonable long-term growth potential. Choose your business software with great care after careful investigation. Application software is the heart of any business automation. Without good business software, a computer network is just a collection of electrical and mechanical parts. Choosing the right application software is more difficult and of great consequence than buying your computer hardware. If the computer becomes unreliable or too slow, then it’s easy enough to upgrade your hardware and migrate your application programs and data to the new computer, but changing over to an entirely new software program is much more expensive and disruptive.
Further, your accumulated business data is the heart of your business — it is the sole reason that you bought and implemented all of that expensive hardware and software. It’s more valuable than all your hardware and business application software combined. Be sure that your data will be usable for many years in the future and can be converted as needed should you decide upon different application software or if your current vendor goes bust. Try to avoid becoming locked into a vendor’s proprietary data format that cannot be later converted or upgraded to other application programs in the future.
Basic engineering and overall system reliability and performance don’t vary a great deal from brand to brand anymore, even though brand-name manufacturers often use system boards that are proprietary in how they mechanically attach to the computer case, an approach that unfortunately precludes less-costly, third-party upgrades later.
Even though brand-name systems from first-tier vendors can often be excellent buys, you should favor locally purchasing more readily upgradable generic systems that use high-quality components mounted inside an industry-standard, ATX-style system case.
Most brand-name vendors assemble components made by the same vendors used by better “white box” generic computer systems. Again, the system case ideally would be a vertical, floor-mounted system that’s capacious and well-cooled.
Consider making periodic upgrades to your systems, which is easier to do with generic “white box” hardware. Periodic upgrades are often a highly economical, less disruptive way to regularly improve computing performance.
At a minimum, you should buy medium-high performance computer hardware rather than the fastest bleeding-edge system or the slowest, cheapest system, which is destined to be soon obsolete and unrepairable. Try to avoid systems that use wholly proprietary components.
Don’t buy more hardware than you need and avoid advertising-driven computer consumerism. As a business user, you are buying a cost-effective business tool, not a hobby whose main purpose is emotional satisfaction, even though most of us also enjoy the latter. Typically, online and retail dealers push their most expensive, highest-margin units. The extra cost of the highest-priced computers derives primarily from highly touted consumer features, such as superfast video cards, features that have little real benefit in most offices.
Historically, system performance increases while prices simultaneously plummet. There’s no immediate end in sight to either trend, although the rate of useful improvement has been diminishing lately as the technology matures. Make purchasing decisions based on current needs, rather than perceptions of what you might need in a year or two. Cutting-edge technology is typically overpriced, immature and unreliable while it’s still “hot.” Many manufacturers try to sell you their higher-margin, top-of-the-line systems and fastest components by promising that purchasing marginally more computing power ostensibly avoids the need to upgrade hardware as often. That’s false economy at best, and it’s probably not even true.
I believe it’s most sensible to buy quality, mature technology that’s about a half generation behind the current top of the line. Buying about a half a generation behind the leading edge saves you a lot of money while providing more reliable technology with enough performance to work satisfactorily for at least two to three years. These cost savings alone should allow you to regularly and economically upgrade the critical computer system components, the CPU, DRAM and hard disk, or to replace the system more often — a policy that saves money and keeps your hardware more generally current for the overall life of the system.
I usually recommend a three-year cycle for complete replacement of a computer, printer or scanner, although you might want to make partial computer system upgrades more frequently, given today’s very low component prices. It’s false economy to retain or not upgrade a too-slow system until it’s been fully depreciated based upon an artificially long depreciation schedule. Remember, modern computer and communications technology are now the basic tools and lifeblood of any business.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com, along with links to legal and community resources.
In tough times, you need solutions that increase your efficiency and your effectiveness, at a smart price. One of the best ways to improve your profitability and viability is to make smart decisions about investing in new technology and to then crisply implement your decisions.
This week, as a wrap-up of our recent computing hardware discussions, I’ll address some more general business and planning considerations that business users should take into account when purchasing new PC computer automation.
Compared to other steps that you can take in your business, prudent investment in technology can produce comparatively large benefits. Small- to medium-sized businesses are a particularly good fit for Michael Porter’s business strategy, developed at Harvard Business School, that focuses upon optimum profitability and survivability in tough times through developing low-volume, high-margin work. Smart technology implementation often is key to developing that sort of resilient business because it allows you to both improve your responsiveness and effectiveness while simultaneously reducing your base overhead costs.
Even though simple approaches and programs often provide the best return for your automation investment and are usually the easiest to implement and use, businesses commonly throw a lot of immature “bleeding-edge” technology, expensive hardware and even more expensive consultant fees and staff time at what is really a business management problem.
Gradually implementing a well-thought-out strategic business automation plan is usually the most effective and efficient way to transform your business, buying only the hardware and software that you will be able to install and begin using within the next two months or so and then implementing your next phase.
Purchasing binges are inefficient. By the time that you get around to installing some of your purchases, you likely can buy a newer, better version of the same product for less money.
Choose your technology for long-term, low-cost usefulness. Buy mature mainstream technology wherever possible and avoid both dead-end and bleeding-edge hardware and software. You’ll not only save money by purchasing mature, proven software and hardware, but likely also spare yourself an expensive, frustrating experience installing and using immature technology.
The hidden costs in terms of your time, staff time and general disruption when you install new computer technology usually costs a business much more than the actual purchase price of the hardware and software. Keep these less-obvious costs in mind when deciding what technology to purchase. It may be less costly in the end for a business to buy something that’s initially more expensive but ultimately more easily installed, maintained and used.
Use the least complex technology that efficiently does the job for you, provided that it’s supported by a financially stable vendor and has reasonable long-term growth potential. Choose your business software with great care after careful investigation. Application software is the heart of any business automation. Without good business software, a computer network is just a collection of electrical and mechanical parts. Choosing the right application software is more difficult and of great consequence than buying your computer hardware. If the computer becomes unreliable or too slow, then it’s easy enough to upgrade your hardware and migrate your application programs and data to the new computer, but changing over to an entirely new software program is much more expensive and disruptive.
Further, your accumulated business data is the heart of your business — it is the sole reason that you bought and implemented all of that expensive hardware and software. It’s more valuable than all your hardware and business application software combined. Be sure that your data will be usable for many years in the future and can be converted as needed should you decide upon different application software or if your current vendor goes bust. Try to avoid becoming locked into a vendor’s proprietary data format that cannot be later converted or upgraded to other application programs in the future.
Basic engineering and overall system reliability and performance don’t vary a great deal from brand to brand anymore, even though brand-name manufacturers often use system boards that are proprietary in how they mechanically attach to the computer case, an approach that unfortunately precludes less-costly, third-party upgrades later.
Even though brand-name systems from first-tier vendors can often be excellent buys, you should favor locally purchasing more readily upgradable generic systems that use high-quality components mounted inside an industry-standard, ATX-style system case.
Most brand-name vendors assemble components made by the same vendors used by better “white box” generic computer systems. Again, the system case ideally would be a vertical, floor-mounted system that’s capacious and well-cooled.
Consider making periodic upgrades to your systems, which is easier to do with generic “white box” hardware. Periodic upgrades are often a highly economical, less disruptive way to regularly improve computing performance.
At a minimum, you should buy medium-high performance computer hardware rather than the fastest bleeding-edge system or the slowest, cheapest system, which is destined to be soon obsolete and unrepairable. Try to avoid systems that use wholly proprietary components.
Don’t buy more hardware than you need and avoid advertising-driven computer consumerism. As a business user, you are buying a cost-effective business tool, not a hobby whose main purpose is emotional satisfaction, even though most of us also enjoy the latter. Typically, online and retail dealers push their most expensive, highest-margin units. The extra cost of the highest-priced computers derives primarily from highly touted consumer features, such as superfast video cards, features that have little real benefit in most offices.
Historically, system performance increases while prices simultaneously plummet. There’s no immediate end in sight to either trend, although the rate of useful improvement has been diminishing lately as the technology matures. Make purchasing decisions based on current needs, rather than perceptions of what you might need in a year or two. Cutting-edge technology is typically overpriced, immature and unreliable while it’s still “hot.” Many manufacturers try to sell you their higher-margin, top-of-the-line systems and fastest components by promising that purchasing marginally more computing power ostensibly avoids the need to upgrade hardware as often. That’s false economy at best, and it’s probably not even true.
I believe it’s most sensible to buy quality, mature technology that’s about a half generation behind the current top of the line. Buying about a half a generation behind the leading edge saves you a lot of money while providing more reliable technology with enough performance to work satisfactorily for at least two to three years. These cost savings alone should allow you to regularly and economically upgrade the critical computer system components, the CPU, DRAM and hard disk, or to replace the system more often — a policy that saves money and keeps your hardware more generally current for the overall life of the system.
I usually recommend a three-year cycle for complete replacement of a computer, printer or scanner, although you might want to make partial computer system upgrades more frequently, given today’s very low component prices. It’s false economy to retain or not upgrade a too-slow system until it’s been fully depreciated based upon an artificially long depreciation schedule. Remember, modern computer and communications technology are now the basic tools and lifeblood of any business.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com, along with links to legal and community resources.
Tuesday, February 3, 2009
Plugged In: Get with the system (board) in building computers
Over the past several weeks, we’ve examined some of the most important factors that affect getting the most for your money when setting up and maintaining a properly balanced Windows computer.
This week, we’ll consider the remaining major components of your computer, including the system board, hard disks, video and CD/DVD drives.
If you haven’t had more than enough technical jargon by the end of this article, then you’re either a serious case or you’ve missed our prior articles, which you can find online at http://redoubtreporter.blogspot.com.
It also means that you’ll need to be very careful when choosing your system board because it’s the heart of your computer, even more so than the CPU processor. At the most basic level, your system board must mechanically fit your case and the case’s attachment points. Generally, so long as you get an ATX or Micro-ATX style board, mechanical compatibility should not be a problem. I prefer the larger ATX boards because they usually provide a little more elbowroom on the board, but Micro-ATX boards are often quite acceptable and are required for small cases.
Your consideration should start with deciding which CPU processor you’ll use because the system board must support the specific CPU processor. At the most basic level of plain mechanical compatibility, the system board must use precisely the same socket type that fits the intended CPU. Assuming that the CPU fits, then the system board should have a chipset that is compatible with the CPU and that includes all of the functions that you’ll want.
At the moment, NVidia and Intel make the most powerful chipsets for current Intel CPUs, most of which fit the LGA 775 pin socket. The best-rated system boards seem to use the following chipsets for Intel CPU processors: Intel P35, P45, G31, Q35 and Q45 chipsets, and NVidia GeForce 9400, 790i and GE 43 chipsets.
Most current AMD CPU processors use either the AM2 or AM2-plus socket and the best-rated system boards for AMD CPUs seem to use the following chipsets: AMD 770, 780 and 790 chipsets or NVidia 570 or 790 chipsets.
Some budget system boards will include onboard video, which eliminates the purchase of a separate video card. Onboard video is usually quite acceptable for average business and home use but is too slow for games or Photoshop. Well-regarded onboard video chipsets include AMD’s Radeon HD 3200/3300 series and NVidia’s GeForce 8200/8300 series.
Because it’s so easy to build a basic budget system board using off-the-shelf chipsets, there are many, many brands. I’ve found the most generally reliable system boards to be made by Gigabyte, MSI, Intel and AMD. ECS, DFI, ASUS and ASrock system boards used to have quite a following, but Gigabyte and MSI seem to be leading the pack lately in general user satisfaction. I’ve had good results with both Gigabyte and MSI boards, which are sold locally and are priced midrange. There are, however, many different models, even within the same brand, and models change very rapidly as features are added or subtracted.
Spending a little more for a high-quality system board is never wasted. The higher reliability and additional hardware functions are usually quite worthwhile add-ons, and you’ll likely be able to get by with a less-expensive processor in the short term and then upgrade when prices on the top-of-the-line CPU drop dramatically in six months. This is actually a very practical strategy that provides an inexpensive path for regular upgrades because CPU prices drop regularly. A good system board should have all, or at least almost all, of the following:
Generally, a separate plug-in 16-channel PCI-X video card provides the best performance. Get a video card that includes at least 256 megabytes, preferably 512 megabytes, of video memory installed on the video card itself. Video cards that “share” the DRAM memory installed on the system board are definitely budget items. They work OK but are not the preferred option. Be sure that the video card’s resolution and other characteristics are matched to your video monitor.
If you are an Adobe Photoshop CS4 or Lightroom 2.x user, check Adobe’s Web site for a list of video card chipsets that are certified to accelerate photo processing in Adobe Photoshop CS4 and Lightroom 2.x. Video card photo processing is a new feature in the most recent Adobe products. It’s a very worthwhile performance boost and is recommended. Gamers definitely need the fastest video cards available.
Decent video monitors are now quite inexpensive compared to even a few years ago. I have always had excellent luck with monitors from Viewsonic and their budget brand, Optiquest. Get a good monitor, though, because recent studies strongly suggest that higher-quality monitors improve productivity and reduce fatigue and eye strain. Wide-screen monitors are OK, but if you do get a wide-screen monitor, you’ll probably want to get at least a 22-inch, preferably a 24-inch, monitor. Otherwise, you’ll be squinting.
Look for these features in a good hard disk:
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com, along with links to legal and community resources.
This week, we’ll consider the remaining major components of your computer, including the system board, hard disks, video and CD/DVD drives.
If you haven’t had more than enough technical jargon by the end of this article, then you’re either a serious case or you’ve missed our prior articles, which you can find online at http://redoubtreporter.blogspot.com.
System Boards
The system board is the single most critical and complex part of your computer system. Most modern computers contain few, if any, traditional expansion cards anymore. The hardware and functions that used to reside on six or eight plug-in expansion cards are now integrated directly into the “chipset” on the system board, which is sometimes called a “motherboard.” Usually, that means that you’ll get more functionality and reliability for less money. It also means that you’ll need to be very careful when choosing your system board because it’s the heart of your computer, even more so than the CPU processor. At the most basic level, your system board must mechanically fit your case and the case’s attachment points. Generally, so long as you get an ATX or Micro-ATX style board, mechanical compatibility should not be a problem. I prefer the larger ATX boards because they usually provide a little more elbowroom on the board, but Micro-ATX boards are often quite acceptable and are required for small cases.
Your consideration should start with deciding which CPU processor you’ll use because the system board must support the specific CPU processor. At the most basic level of plain mechanical compatibility, the system board must use precisely the same socket type that fits the intended CPU. Assuming that the CPU fits, then the system board should have a chipset that is compatible with the CPU and that includes all of the functions that you’ll want.
At the moment, NVidia and Intel make the most powerful chipsets for current Intel CPUs, most of which fit the LGA 775 pin socket. The best-rated system boards seem to use the following chipsets for Intel CPU processors: Intel P35, P45, G31, Q35 and Q45 chipsets, and NVidia GeForce 9400, 790i and GE 43 chipsets.
Most current AMD CPU processors use either the AM2 or AM2-plus socket and the best-rated system boards for AMD CPUs seem to use the following chipsets: AMD 770, 780 and 790 chipsets or NVidia 570 or 790 chipsets.
Some budget system boards will include onboard video, which eliminates the purchase of a separate video card. Onboard video is usually quite acceptable for average business and home use but is too slow for games or Photoshop. Well-regarded onboard video chipsets include AMD’s Radeon HD 3200/3300 series and NVidia’s GeForce 8200/8300 series.
Because it’s so easy to build a basic budget system board using off-the-shelf chipsets, there are many, many brands. I’ve found the most generally reliable system boards to be made by Gigabyte, MSI, Intel and AMD. ECS, DFI, ASUS and ASrock system boards used to have quite a following, but Gigabyte and MSI seem to be leading the pack lately in general user satisfaction. I’ve had good results with both Gigabyte and MSI boards, which are sold locally and are priced midrange. There are, however, many different models, even within the same brand, and models change very rapidly as features are added or subtracted.
Spending a little more for a high-quality system board is never wasted. The higher reliability and additional hardware functions are usually quite worthwhile add-ons, and you’ll likely be able to get by with a less-expensive processor in the short term and then upgrade when prices on the top-of-the-line CPU drop dramatically in six months. This is actually a very practical strategy that provides an inexpensive path for regular upgrades because CPU prices drop regularly. A good system board should have all, or at least almost all, of the following:
- A chipset that supports the fastest DRAM speeds and “Hypertransport” speeds of which your intended CPU is capable. Otherwise, you’ll be wasting a lot of that expensive processing power.
- At least eight high-speed USB ports with at least four, preferably six, physically present back-panel USB connectors and internal wiring for two to four more USB 2.0 connectors for front panel and internal use. Almost every external device connects to your computer through a fast USB port anymore, especially scanners, printers, microphones, Web cameras and mice.
- One or two FireWire connectors would be nice, especially for video camcorders and fast external hard disks.
- At least one gigabit-speed Ethernet network adapter. I prefer the Marvell gigabit Ethernet chip because it includes a unique test program that helps diagnose networking cable problems.
- At least four fast 3.0-gigabit SATA connectors for internal hard disks and newer DVD drives and also at least one IDE connector for older CD/DVD drives. If you plan to use this computer to act as a central storage “file server” for a small office network, then be sure that it includes RAID 0+1 or RAID 5 disk array functions on the system board.
- A decent audio chipset to power an external speaker set. Although the Realtek audio chipset is fine, I prefer the NVidia audio chipset. The audio control software seems more straightforward and effective and that’s most of it.
- A floppy disk drive connector ¬— everyone except Microsoft knows that the floppy disk is obsolete. Unfortunately, installing many versions of Windows still requires you to have the specialized driver software for your hard disk controller on a floppy disk and you’ll need to connect that floppy disk.
- Several small internal fan connectors.
- The usual keyboard, mouse, and printer port connectors.
- Several slots for expansion cards.
- Small fans on the main system board chipsets, particularly fans that don’t interfere with seating a video card. Modern chipsets run hot and I do not believe that simple metal cooling fans provide enough cooling for long-term reliability. Small fans are better, assuming that they are of reliable quality.
Video
Unless you are a gamer or a professional user of CAD, Adobe Photoshop or Adobe Lightroom, nearly any stable video card with driver software for your specific operating system will be adequate. Video functions built directly into the system board tend to be the slowest but are usually OK for business users. As with system boards, there are numerous video card brands and models, all of which are built using a few standard video chipsets from NVidia or AMD. I have used video boards built around the AMD/ATI Radeon for some years with good reliability and quite adequate performance for the price. Generally, a separate plug-in 16-channel PCI-X video card provides the best performance. Get a video card that includes at least 256 megabytes, preferably 512 megabytes, of video memory installed on the video card itself. Video cards that “share” the DRAM memory installed on the system board are definitely budget items. They work OK but are not the preferred option. Be sure that the video card’s resolution and other characteristics are matched to your video monitor.
If you are an Adobe Photoshop CS4 or Lightroom 2.x user, check Adobe’s Web site for a list of video card chipsets that are certified to accelerate photo processing in Adobe Photoshop CS4 and Lightroom 2.x. Video card photo processing is a new feature in the most recent Adobe products. It’s a very worthwhile performance boost and is recommended. Gamers definitely need the fastest video cards available.
Decent video monitors are now quite inexpensive compared to even a few years ago. I have always had excellent luck with monitors from Viewsonic and their budget brand, Optiquest. Get a good monitor, though, because recent studies strongly suggest that higher-quality monitors improve productivity and reduce fatigue and eye strain. Wide-screen monitors are OK, but if you do get a wide-screen monitor, you’ll probably want to get at least a 22-inch, preferably a 24-inch, monitor. Otherwise, you’ll be squinting.
Hard disks
Twenty-five years after IBM introduced the first desktop hard disk, your hard disk’s performance remains one of the single most critical determinants of overall computer performance. (What I actually wanted to write was “Redoubt Reporter readers will readily recall my recent review of really radical performance with rapidly revolving disk drives,” but I’m not sure that I can slip it by our eagle-eyed, style-manualed editor, and the thesaurus ran out of synonyms starting with “re.”)Look for these features in a good hard disk:
- SATA II interface — it’s faster and easy to set up.
- Fast rotation — 7,200 RPM drives are standard. Western Digital’s 10,000 RPM Raptor drives are really fast when matched to the right system board SATA disk controller, but they’re a lot more expensive. If you want the ultimate in desktop PC performance, be ready to spend at least two to three times as much per gigabyte capacity if you buy a 150 GB or 300 GB Raptor drive. On the other hand, these drives are rated as highly reliable.
- High capacity — not only do you want at least 500 gigabytes storage, but modern high capacity hard disks pack the data more tightly, resulting in more data mechanically passing under the read-write head on each revolution, resulting in better overall performance.
- A large “buffer,” at least 8 MB built in, preferably 16 MB or 32 MB.
- High shock and drop resistance. None of us has ever dropped a hard disk or computer. None of us. Of course.
- Two hard disks in your computer. Put your operating system and application programs on the boot drive (C) and your data on a second, physically different drive (D). That way, if Windows goes south, which it will in time, you won’t lose data when you need to reformat the boot hard disk. In addition, using a second hard disk to hold your data makes it easier to upgrade to a larger hard disk when needed and to back up your data.
Odds and ends
- DVD drive and floppy disk — just about any brand-name DVD read/write driver and any brand-name floppy disk will do the trick. These have become commodity items.
- Keyboard and mouse — wired keyboards and optical mice are generic commodity items. I prefer the convenience of a wireless keyboard. I’ve had my best luck with the less-expensive Logitech models, but even these fail more often than hard-wired devices.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com, along with links to legal and community resources.
Tuesday, January 20, 2009
Plugged in: With computers, more money doesn’t always mean more speed
Editor’s note: This is the second in a series of stories about making smart PC purchases.
Warning: Spending more money does not necessarily buy you greater useful computing power and reliability. Any computer is only as fast as its slowest bottleneck, rather like a chain’s weakest link, and each computer should be carefully balanced for cost-effective performance.
DRAM memory is the very high-speed “volatile” electronic chip memory that holds programs and data only so long as a computer is powered up. When the computer is turned off, whatever unsaved data remains in volatile DRAM memory disappears.
DRAM memory retrieves and manipulates data far faster than your spinning mechanical hard disk but the hard disk retains its data as a magnetic recording when powered down and holds hundreds of times more data. Not having enough DRAM forces your computer to frequently use the slow mechanical hard disk for primary computing, which greatly slows down any computer. DRAM memory is now very inexpensive.
One of the first and most inexpensive ways to often improve computer performance is to add additional memory, particularly if you are using the computer for memory-intensive tasks such as processing photos or videos. Ordinary 32-bit Windows XP computers should be upgraded by installing at least 2 gigabytes (GB) DRAM if the computer exhibits any slowdowns that coincide with heavy hard disk activity. (Check to see if your hard disk indicator light is on most of the time and under what circumstances.) Sixty-four-bit Vista and Windows XP x 64 computers should have at least 4 GB of installed DRAM. Adding more than 4 GB DRAM to a computer using a 32-bit operating system like Windows XP is a waste — the computer can’t even recognize that it’s installed. This is not a limitation with 64-bit operating systems.
There are several types of DRAM, usually designated DDR, DDR2 and DDR3, along with the DRAM’s speed ratings. The correct type and speed for each CPU processor and system board must be installed. Tests show that adding DRAM whose rated external speed is significantly faster than the minimum specified for your system board and CPU is a lot more expensive but only provides a very slight performance advantage. That’s because current DRAM technologies can only go so fast internally before becoming unreliable. When the external speed (the bus speed) is increased, something that makes marketing departments happy, the internal chips can’t really keep up and provide data all that much faster. In order to avoid crashing, high speed DDR memory inserts internal “wait states” that slow the data transfer rate. The net result is that there’s only a very modest improvement in overall system performance. One real benefit of AMD CPU processors, by the way, is that they include the DRAM memory controller directly on the CPU and the memory controller runs at full CPU speed, which does result in better DRAM performance under most circumstances.
When you upgrade the amount of DRAM in your computer, replace all of the DRAM at the same time, adding two matched sticks at a time in each DRAM channel. All installed DRAM needs to be matched in order to avoid instability and to maximize performance.
When you’re adding DRAM, be sure that its rated speed is at least as fast as the standard for that system board and installed CPU; I prefer to install DRAM that is rated for one speed increment faster than the minimum required for a particular CPU and system board. Slower DRAM is slightly cheaper but more prone to instability or slower performance. Using DRAM that’s rated as faster than the minimum required usually results in a more reliable system, provided you operate the computer at its default memory speed. That gives you a little in reserve.
There’s a really useful program, CPUID, that will give you a wealth of information about your installed processor, your system board and your DRAM memory. It’s a free download from www.cpuid.com. This free diagnostic is recommended.
Also recommended when you’re trying to balance the performance of your computer system is a benchmark program called “Performance Test”, now in version 6.1. Although the software license activation for Performance Test costs $24, a very reasonable price, you can get a 30-day, full-function download from the vendor’s Web site, www.passmark.com. Passmark also sells a “burn-in” program to stress test new hardware and provides a 30-day, full-function trial of that program, as well.
I’ve used Performance Test for many years as a means for identifying the weak points in my computer systems. For example, I recently found that my law office file server had quite adequate CPU and network performance but that the hard disk performance was very much sub-par, even though I had installed a RAID disk array of 10,000 rpm Western Digital Raptor SATA hard disks, some of the faster hard disks available. Raptor hard disks should be capable of reading and writing 100 to 200 megabytes of data per second. I was getting between 4 MB and 7 MB per second. Spending hundreds of dollars to replace these fast hard disks was obviously not the answer. That brings us to our next low-cost means of improving computer performance.
Driver Software: Software “drivers” are the bits of programming code that allow any attached hardware to work properly with the Windows operating system. Microsoft provides a lot of the basic driver software needed to get Windows installed and running but the Microsoft drivers, although generally pretty reliable and stable, are necessarily generic and do not always provide the best performance with specific computing hardware. This is particularly true for high-end video cards, fast hard disk controller chips, scanners and network connections.
Updating the “driver” software for too-slow computer components is one of the best ways to improve the system performance and stability of an unbalanced computer system, and it’s usually free. My own file server’s hard disk performance tripled after I updated the driver software for the hard disk controllers. It’s still not really good enough, but three times faster than before. That’s a good start for a 10-minute process that cost me nothing.
When you buy new computer hardware that requires some sort of driver software, the hardware vendor usually provides their optimized driver software on a disk or CD. However, the optimized vendor software is not always installed when a new computer is set up and vendors constantly update their software drivers to reduce bugs and improve performance, posting new driver software to their Web sites. So, your next step after identifying specific performance weaknesses is to update the drivers for the related hardware. For example, if hard disk performance seems weak, update the IDE and SATA disk controller driver software.
There are several ways to update the driver software for balky hardware. You can do so within the Windows Control Panel system tab or using Windows Update over the Web. You can use the vendor’s CD that came with your computer or go directly to the hardware vendor’s Web site. Some Web sites include Internet-based scans and automatic updating while others require you to know what’s installed and manually choose and download updated driver software.
You must be very careful, though, when updating driver software. If the new software is not stable, then your system will crash. Before you update driver software, be sure that you completely back up your computer system. Also, use the Windows System Restore function to set a restore point with known good driver software, so that you can roll your system back in the event of instability.
Finally, don’t install a driver that the vendor just posted for the first time last week. Unless you’re in a critical situation, don’t be a test pilot — let some other eager fellow take the first risks. Wait a few weeks before updating to new driver software so that the vendor has a chance to correct any problems. The slowest computer of all is one that won’t even work.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com, along with links to legal and community resources.
Warning: Spending more money does not necessarily buy you greater useful computing power and reliability. Any computer is only as fast as its slowest bottleneck, rather like a chain’s weakest link, and each computer should be carefully balanced for cost-effective performance.
DRAM memory is the very high-speed “volatile” electronic chip memory that holds programs and data only so long as a computer is powered up. When the computer is turned off, whatever unsaved data remains in volatile DRAM memory disappears.
DRAM memory retrieves and manipulates data far faster than your spinning mechanical hard disk but the hard disk retains its data as a magnetic recording when powered down and holds hundreds of times more data. Not having enough DRAM forces your computer to frequently use the slow mechanical hard disk for primary computing, which greatly slows down any computer. DRAM memory is now very inexpensive.
One of the first and most inexpensive ways to often improve computer performance is to add additional memory, particularly if you are using the computer for memory-intensive tasks such as processing photos or videos. Ordinary 32-bit Windows XP computers should be upgraded by installing at least 2 gigabytes (GB) DRAM if the computer exhibits any slowdowns that coincide with heavy hard disk activity. (Check to see if your hard disk indicator light is on most of the time and under what circumstances.) Sixty-four-bit Vista and Windows XP x 64 computers should have at least 4 GB of installed DRAM. Adding more than 4 GB DRAM to a computer using a 32-bit operating system like Windows XP is a waste — the computer can’t even recognize that it’s installed. This is not a limitation with 64-bit operating systems.
There are several types of DRAM, usually designated DDR, DDR2 and DDR3, along with the DRAM’s speed ratings. The correct type and speed for each CPU processor and system board must be installed. Tests show that adding DRAM whose rated external speed is significantly faster than the minimum specified for your system board and CPU is a lot more expensive but only provides a very slight performance advantage. That’s because current DRAM technologies can only go so fast internally before becoming unreliable. When the external speed (the bus speed) is increased, something that makes marketing departments happy, the internal chips can’t really keep up and provide data all that much faster. In order to avoid crashing, high speed DDR memory inserts internal “wait states” that slow the data transfer rate. The net result is that there’s only a very modest improvement in overall system performance. One real benefit of AMD CPU processors, by the way, is that they include the DRAM memory controller directly on the CPU and the memory controller runs at full CPU speed, which does result in better DRAM performance under most circumstances.
When you upgrade the amount of DRAM in your computer, replace all of the DRAM at the same time, adding two matched sticks at a time in each DRAM channel. All installed DRAM needs to be matched in order to avoid instability and to maximize performance.
When you’re adding DRAM, be sure that its rated speed is at least as fast as the standard for that system board and installed CPU; I prefer to install DRAM that is rated for one speed increment faster than the minimum required for a particular CPU and system board. Slower DRAM is slightly cheaper but more prone to instability or slower performance. Using DRAM that’s rated as faster than the minimum required usually results in a more reliable system, provided you operate the computer at its default memory speed. That gives you a little in reserve.
There’s a really useful program, CPUID, that will give you a wealth of information about your installed processor, your system board and your DRAM memory. It’s a free download from www.cpuid.com. This free diagnostic is recommended.
Also recommended when you’re trying to balance the performance of your computer system is a benchmark program called “Performance Test”, now in version 6.1. Although the software license activation for Performance Test costs $24, a very reasonable price, you can get a 30-day, full-function download from the vendor’s Web site, www.passmark.com. Passmark also sells a “burn-in” program to stress test new hardware and provides a 30-day, full-function trial of that program, as well.
I’ve used Performance Test for many years as a means for identifying the weak points in my computer systems. For example, I recently found that my law office file server had quite adequate CPU and network performance but that the hard disk performance was very much sub-par, even though I had installed a RAID disk array of 10,000 rpm Western Digital Raptor SATA hard disks, some of the faster hard disks available. Raptor hard disks should be capable of reading and writing 100 to 200 megabytes of data per second. I was getting between 4 MB and 7 MB per second. Spending hundreds of dollars to replace these fast hard disks was obviously not the answer. That brings us to our next low-cost means of improving computer performance.
Driver Software: Software “drivers” are the bits of programming code that allow any attached hardware to work properly with the Windows operating system. Microsoft provides a lot of the basic driver software needed to get Windows installed and running but the Microsoft drivers, although generally pretty reliable and stable, are necessarily generic and do not always provide the best performance with specific computing hardware. This is particularly true for high-end video cards, fast hard disk controller chips, scanners and network connections.
Updating the “driver” software for too-slow computer components is one of the best ways to improve the system performance and stability of an unbalanced computer system, and it’s usually free. My own file server’s hard disk performance tripled after I updated the driver software for the hard disk controllers. It’s still not really good enough, but three times faster than before. That’s a good start for a 10-minute process that cost me nothing.
When you buy new computer hardware that requires some sort of driver software, the hardware vendor usually provides their optimized driver software on a disk or CD. However, the optimized vendor software is not always installed when a new computer is set up and vendors constantly update their software drivers to reduce bugs and improve performance, posting new driver software to their Web sites. So, your next step after identifying specific performance weaknesses is to update the drivers for the related hardware. For example, if hard disk performance seems weak, update the IDE and SATA disk controller driver software.
There are several ways to update the driver software for balky hardware. You can do so within the Windows Control Panel system tab or using Windows Update over the Web. You can use the vendor’s CD that came with your computer or go directly to the hardware vendor’s Web site. Some Web sites include Internet-based scans and automatic updating while others require you to know what’s installed and manually choose and download updated driver software.
You must be very careful, though, when updating driver software. If the new software is not stable, then your system will crash. Before you update driver software, be sure that you completely back up your computer system. Also, use the Windows System Restore function to set a restore point with known good driver software, so that you can roll your system back in the event of instability.
Finally, don’t install a driver that the vendor just posted for the first time last week. Unless you’re in a critical situation, don’t be a test pilot — let some other eager fellow take the first risks. Wait a few weeks before updating to new driver software so that the vendor has a chance to correct any problems. The slowest computer of all is one that won’t even work.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com, along with links to legal and community resources.
Tuesday, January 13, 2009
Plugged In: Self configuring a good way to save on computers
This week and next, I’d like to discuss what’s cost-effective when purchasing a new Windows-based PC computer. A little care in choosing your new computer’s specifications can result in major savings without seriously compromising performance. Over the past 20 years, I have personally built a few hundred Windows-based computer systems and these articles are a distillation of that experience.
Name-brand systems are often advertised in print or Internet media with very low starting prices, but their ultimate purchase price can be just as high as locally purchased systems after fully “configuring” the system and paying freight. Aside from Costco and the now-defunct CompUSA, the retail electronics chains that I’ve shopped in Anchorage try to sell high-margin consumer electronics for full list price, which I consider ludicrous. If you do want to buy a name-brand system, then you’ll usually get the best with HP, sold by both Costco and Fred Meyer.
However, I believe that locally purchasing a “white box” computer custom-configured to your needs often makes the most sense, particularly in a less-populous area like Kenai-Soldotna where same-day manufacturer’s service is usually not available. Any prices that I mention in these articles were current in December 2008 at Soldotna’s Peninsula Technologies, where I usually buy my routine computer components, like hard disks and cases, although there are several reputable vendors in the Kenai-Soldotna area.
Don’t waste money “buying for the future.” In six months, today’s super expensive, top-end hardware will be ho-hum, and in a year it will be obsolete.
Any modern PC computer will contain the following basic components:
System case: Modern computers, especially powerful ones, generate a lot of heat, and excessive heat is the deadly enemy of computer reliability. Small cases in which components fit tightly may seem cute or easy on desktop space, but they’re not very expandable and are difficult to cool. You’ll do much better with a midsized tower case that you can put under a desk. You’ll be able to expand it much more readily and are less likely to fry internal components. Chrome, flashy designs and multicolored lights may look cool, but add nothing to performance.
On the other hand, the flash built into a lot of cases usually doesn’t cost any more, so why not? Just be sure that you get a case that is rugged, with lots of room for extra drives and that’s easy to access from both sides. A decent case will cost in the range of $80 to $90 locally, which is pretty reasonable, considering shipping costs.
Cooling fans: You will need to add some extra case ventilation fans, preferably at least one large, quiet, 120-mm fan on the back side that exhausts hot air. These are cheap insurance against hardware failure and data loss. Exhaust fans work better than fans that try to force air into the case. A second fan on the top or side of the case is a good idea. I prefer fans that plug directly into the system board and are controlled by it. Avoiding heat buildup is critical to computing reliability.
Remember to periodically remove dust and pet fur from all cooling fans, air intakes and internal components. Doing so is crucial to avoiding heat-related failures — as I know from recent personal experience. A 120-mm fan will probably cost about $12 to $15 dollars and take about three minutes to install, if you know where to plug it into the system board.
Most central processing unit processors sold in the manufacturer’s original retail packaging include a cooling fan that’s matched to the CPU when used at its rated speed. AMD seems to be particularly good about including very robust CPU cooling fans with their retail-boxed processors. However, if you plan to overclock your computer, as computer gamers are wont to do, you may need to buy a third-party cooling system. These tend to be expensive and are unnecessary for routine business use.
Power supply: Although a power supply may seem pretty boring compared to the latest and greatest Intel or AMD processor, it’s one of the single most critical components in a computer, and one of the most likely to fail. Although many computer cases include basic power supplies, these are often barely adequate, on the order of 350 to 400 watts. To ensure better internal cooling and increased reliability, spend the extra money for a decent quality, 500- to 600-watt power supply for a basic business computer and something even bigger if you are a gamer using a powerful video card. A good-quality, 500- to 600-watt power supply will cost somewhere between $55 and $80, but is money well spent if you are putting together a powerful computer.
Central processing unit: The CPU is the heart of your computer and will be built either by Intel or by AMD. Both companies make processors that are comparably reliable and fast. AMD processors tend to be exceptionally reliable and somewhat less expensive for the same performance. Cutting to the chase, I recommend a dual-core AMD Athlon X2 6000 CPU costing about $139 locally. Get the version that uses the current AM2 socket. This is neither the latest nor the greatest, but performs nearly as well in the real world for a lot less money. Here’s why:
Unless experimentally “overclocked” by a knowledgeable technician, a CPU will operate at its rated speed, usually in the range of 2 to 3 gigahertz (2 billion to 3 billion cycles per second). Overclocking, by itself, only marginally increases performance because other components like DDR memory and hard disks become bottlenecks that seriously limit any performance increases. Overclocking often introduces serious system instability problems and usually requires faster memory and expensive third-party CPU cooling fans. Business users should avoid overclocking (for one thing, it voids your warranty) but overclocking can be an interesting challenge to the technically adept who have nothing better to do at the moment.
However, just as more megapixels don’t necessarily translate into a better digital photograph, a higher CPU speed doesn’t always translate into a faster computer. In fact, even though quad core CPUs running around 3 GHz are definitely more expensive, all other things being equal, the overall performance improvement is usually only a few percent, basically unnoticeable unless you spend your work day doing nothing but playing demanding games or running synthetic performance benchmark programs. (I’ve actually been there and done that in years past when writing my Lawyers Lab technology columns for Law Office Computing Magazine.)
Much more important is a processor’s efficient internal processing. AMD Athlon X2 and Phenom quad core CPUs and the newer Intel Core 2 Duo and Core 2 Quad processors are both generally more reliable and more efficient than earlier Intel Pentiums and AMD single core Athlons of any speed. Essentially all current desktop CPUs have at least two, and often three or four, processing units located inside that single CPU chips, and even humble 32-bit Windows XP now recognizes quad core CPUs as four separate processing units.
All current desktop CPUs are inherently “64-bit” processors, which means they can be compared to a superhighway with many high-speed lanes and reduced traffic congestion in each direction. In theory, all of this should mean that a dual-core CPU arguably should be at least twice as fast, and a quad core CPU four times as fast, than earlier single-core processors. Sadly, that’s not true. You get less than you pay for and that has nothing to do with AMD or Intel and everything to do with Microsoft and application software vendors.
Windows and most current application software do not efficiently allocate processing tasks (threads) to several different processor cores, and thus do not take full advantage of the extra CPU cores. In fact, except for the newest versions of Photoshop and a few less-common application programs, most business programs, such as spreadsheets or word processors, can only use a single processing core. The remaining one to three CPU cores are basically wasted, using more electricity and producing a lot of heat but no illumination, although they do help a bit with background system chores like antivirus scans.
In addition, unless you’re using a 64-bit version of Windows XP or Windows Vista, you’ll only be processing at 32 bits, basically trying to navigate a freeway that’s become congested because half its lanes are blocked off. The 64 bit version of Windows XP is hard to find, although very reliable and quick, while the 64 bit version of Windows Vista is slowed due to fancy interface geegaws that demand a lot of processing power to look pretty, although they contribute little or nothing to efficient computing.
Next week, I’ll discuss what to look for when specifying the remaining major computer system components, including the system board, which now generally includes most of the controllers needed by a modern computer, such as audio, networking, USB, CD/DVD and hard disk controllers, video cards, DDR memory, hard disks, DVD/CD reader/writer, floppy disk drive, video monitor, keyboard, mouse and operating system.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com, along with links to legal and community resources.
Name-brand systems are often advertised in print or Internet media with very low starting prices, but their ultimate purchase price can be just as high as locally purchased systems after fully “configuring” the system and paying freight. Aside from Costco and the now-defunct CompUSA, the retail electronics chains that I’ve shopped in Anchorage try to sell high-margin consumer electronics for full list price, which I consider ludicrous. If you do want to buy a name-brand system, then you’ll usually get the best with HP, sold by both Costco and Fred Meyer.
However, I believe that locally purchasing a “white box” computer custom-configured to your needs often makes the most sense, particularly in a less-populous area like Kenai-Soldotna where same-day manufacturer’s service is usually not available. Any prices that I mention in these articles were current in December 2008 at Soldotna’s Peninsula Technologies, where I usually buy my routine computer components, like hard disks and cases, although there are several reputable vendors in the Kenai-Soldotna area.
Don’t waste money “buying for the future.” In six months, today’s super expensive, top-end hardware will be ho-hum, and in a year it will be obsolete.
Any modern PC computer will contain the following basic components:
System case: Modern computers, especially powerful ones, generate a lot of heat, and excessive heat is the deadly enemy of computer reliability. Small cases in which components fit tightly may seem cute or easy on desktop space, but they’re not very expandable and are difficult to cool. You’ll do much better with a midsized tower case that you can put under a desk. You’ll be able to expand it much more readily and are less likely to fry internal components. Chrome, flashy designs and multicolored lights may look cool, but add nothing to performance.
On the other hand, the flash built into a lot of cases usually doesn’t cost any more, so why not? Just be sure that you get a case that is rugged, with lots of room for extra drives and that’s easy to access from both sides. A decent case will cost in the range of $80 to $90 locally, which is pretty reasonable, considering shipping costs.
Cooling fans: You will need to add some extra case ventilation fans, preferably at least one large, quiet, 120-mm fan on the back side that exhausts hot air. These are cheap insurance against hardware failure and data loss. Exhaust fans work better than fans that try to force air into the case. A second fan on the top or side of the case is a good idea. I prefer fans that plug directly into the system board and are controlled by it. Avoiding heat buildup is critical to computing reliability.
Remember to periodically remove dust and pet fur from all cooling fans, air intakes and internal components. Doing so is crucial to avoiding heat-related failures — as I know from recent personal experience. A 120-mm fan will probably cost about $12 to $15 dollars and take about three minutes to install, if you know where to plug it into the system board.
Most central processing unit processors sold in the manufacturer’s original retail packaging include a cooling fan that’s matched to the CPU when used at its rated speed. AMD seems to be particularly good about including very robust CPU cooling fans with their retail-boxed processors. However, if you plan to overclock your computer, as computer gamers are wont to do, you may need to buy a third-party cooling system. These tend to be expensive and are unnecessary for routine business use.
Power supply: Although a power supply may seem pretty boring compared to the latest and greatest Intel or AMD processor, it’s one of the single most critical components in a computer, and one of the most likely to fail. Although many computer cases include basic power supplies, these are often barely adequate, on the order of 350 to 400 watts. To ensure better internal cooling and increased reliability, spend the extra money for a decent quality, 500- to 600-watt power supply for a basic business computer and something even bigger if you are a gamer using a powerful video card. A good-quality, 500- to 600-watt power supply will cost somewhere between $55 and $80, but is money well spent if you are putting together a powerful computer.
Central processing unit: The CPU is the heart of your computer and will be built either by Intel or by AMD. Both companies make processors that are comparably reliable and fast. AMD processors tend to be exceptionally reliable and somewhat less expensive for the same performance. Cutting to the chase, I recommend a dual-core AMD Athlon X2 6000 CPU costing about $139 locally. Get the version that uses the current AM2 socket. This is neither the latest nor the greatest, but performs nearly as well in the real world for a lot less money. Here’s why:
Unless experimentally “overclocked” by a knowledgeable technician, a CPU will operate at its rated speed, usually in the range of 2 to 3 gigahertz (2 billion to 3 billion cycles per second). Overclocking, by itself, only marginally increases performance because other components like DDR memory and hard disks become bottlenecks that seriously limit any performance increases. Overclocking often introduces serious system instability problems and usually requires faster memory and expensive third-party CPU cooling fans. Business users should avoid overclocking (for one thing, it voids your warranty) but overclocking can be an interesting challenge to the technically adept who have nothing better to do at the moment.
However, just as more megapixels don’t necessarily translate into a better digital photograph, a higher CPU speed doesn’t always translate into a faster computer. In fact, even though quad core CPUs running around 3 GHz are definitely more expensive, all other things being equal, the overall performance improvement is usually only a few percent, basically unnoticeable unless you spend your work day doing nothing but playing demanding games or running synthetic performance benchmark programs. (I’ve actually been there and done that in years past when writing my Lawyers Lab technology columns for Law Office Computing Magazine.)
Much more important is a processor’s efficient internal processing. AMD Athlon X2 and Phenom quad core CPUs and the newer Intel Core 2 Duo and Core 2 Quad processors are both generally more reliable and more efficient than earlier Intel Pentiums and AMD single core Athlons of any speed. Essentially all current desktop CPUs have at least two, and often three or four, processing units located inside that single CPU chips, and even humble 32-bit Windows XP now recognizes quad core CPUs as four separate processing units.
All current desktop CPUs are inherently “64-bit” processors, which means they can be compared to a superhighway with many high-speed lanes and reduced traffic congestion in each direction. In theory, all of this should mean that a dual-core CPU arguably should be at least twice as fast, and a quad core CPU four times as fast, than earlier single-core processors. Sadly, that’s not true. You get less than you pay for and that has nothing to do with AMD or Intel and everything to do with Microsoft and application software vendors.
Windows and most current application software do not efficiently allocate processing tasks (threads) to several different processor cores, and thus do not take full advantage of the extra CPU cores. In fact, except for the newest versions of Photoshop and a few less-common application programs, most business programs, such as spreadsheets or word processors, can only use a single processing core. The remaining one to three CPU cores are basically wasted, using more electricity and producing a lot of heat but no illumination, although they do help a bit with background system chores like antivirus scans.
In addition, unless you’re using a 64-bit version of Windows XP or Windows Vista, you’ll only be processing at 32 bits, basically trying to navigate a freeway that’s become congested because half its lanes are blocked off. The 64 bit version of Windows XP is hard to find, although very reliable and quick, while the 64 bit version of Windows Vista is slowed due to fancy interface geegaws that demand a lot of processing power to look pretty, although they contribute little or nothing to efficient computing.
Next week, I’ll discuss what to look for when specifying the remaining major computer system components, including the system board, which now generally includes most of the controllers needed by a modern computer, such as audio, networking, USB, CD/DVD and hard disk controllers, video cards, DDR memory, hard disks, DVD/CD reader/writer, floppy disk drive, video monitor, keyboard, mouse and operating system.
Local attorney Joe Kashi received his bachelor’s and master’s degrees from MIT and his law degree from Georgetown University. He has published many articles about computer technology, law practice and digital photography in national media since 1990. Many of his technology and photography articles can be accessed through his Web site, www.kashilaw.com, along with links to legal and community resources.
Subscribe to:
Posts (Atom)
