Good, Affordable PC Diagnostic Software?

RebornData asks: "I do freelance computer support for small businesses, and after running into a string of hairy hardware problems, decided to buy some generic PC diagnostic software. My searches turned up numerous vendors but very few independent or reputable reviews and comparisons, so I embarked on an evaluation of my own. What I found was an industry filled with con artists, bitter feuds, and outdated products. I'm now out $400 and am wondering whether my hope of finding useful diagnostic software was a naive dream. Has anyone found something that works for you?"

"The premise of PC diagnostics software is simple: provide an easy way to test for PC hardware problems, independent of software configuration. Some hardware vendors (like Dell) provide diagnostics with their systems, but they are usually model-specific and not even all major vendors provide them. Of course there are free utilities like the well-known memtest86, but I was wanted something more comprehensive.

So I started my research, and found a variety of packages, including PC Doctor, PC Check, Microscope, PC Certify, Tufftest Pro, among many others, ranging in price from $500 to $35. Some come with associated hardware, such as loopback connectors for parallel, serial, network or USB ports, or ISA / PCI cards that will show low-level POST codes for machines that appear completely dead.

Some of the vendors provided demos, but most were severely crippled. The cheaper software tended to be outdated and incomplete, lacking support for newer hardware features. Almost all practiced high-pressure sales tactics over the phone, and I discovered that one company was actually a spinoff of another by a disgruntled former employee, resulting in a bitter, lawsuit-ridden feud.

Microscope, by Micro 2000, seemed to have the most online feedback, mostly positive, but they didn't provide a demo. After contacting their sales, they suggested that if I bought a full copy for my evaluation, I could return it in 30 days if it didn't meet my needs. Well, it turned out to be buggy and missing important features found in other, cheaper products. When I called to return the product, the salesman disclaimed all knowledge of the promise they made, and they've refused to take it back. Some further digging found that I'm not the first person to be taken in by these tactics.

I still would like to find worthwhile PC diagnostics software, but the (a) lack of independent reviews, (b) shady industry sales tactics and (c) poor performance of a 'well regarded' package leave me wondering... am I a sucker for buying into the whole concept in the first place? Can anyone point me towards a reputable vendor, or an alternative set of independent tools that will do the same job?"

Sandra

SiSoft's Sandra [sisoftware.net] is good for some basic hardware info on the machine.
It was nice finding out that the RAM I bought from Coast-to-Coast memory that I got a "deal" on was actually a step down in terms of speed (which they were selling for the "sale" price...so it all worked out).

They have diagnosit tests, but I've only used the free version. But its a nice first-line strategy for sizing up machines.

Try AIDA

Hi, I recommend AIDA as a Sysinfo tool. Free and powerful, can even run in batch mode.
http://www.aida32.hu/aida32.php [aida32.hu]

Ziad

Sounds like AIDA is on its last legs

The future of AIDA does not sound all that bright. From their website:

Due to the constantly decreasing interest about our 16-bit DOS system information tool AIDA16, as of February 16, 2004 we have decided to discontinue the product and to remove all AIDA16 related contents from our website aida32.hu. ... Please do NOT ask for source code of AIDA16, it is confidential and will not go public.

and

So as of February 16, 2004 we do NOT provide any official support for AIDA32. ... We preserve the right to ignore all messages about installation, configuration and application problems, however we will still be active on both the international and the Hungarian discussion forums.

Re:Sounds like AIDA is on its last legs

Your interpretation is very different from mine. I read from this that they are no longer going to update AIDA16, but that they will continue to develop AIDA32, they just won't be providing free support for their free product -- other than on the discussion forums.

That doesn't sound anything like being on it's last legs. It sounds like a sensible response for dealing with support for a free product.

Re:Sandra

I've ordered a good amount of memory from Coast to Coast [18004memory.com]. I've saved a lot of money for my friends/family/clients, especially when it comes to SODIMMs (Laptop memory) and router memory. I don't buy their memory because it's the best. I buy from them because I've only had one compatibility issue (Which was promptly resolved) and the price/performance ratio is tremendously in their favor. The relative running Windows XP on 128MB of RAM is in heaven when they get an additional stick thrown in, regardless of how fast it is. Before the upgrade they were running 1000 times slower on the pagefile!

Going back on topic, let me share what I've learned about troubleshooting hardware issues. First, let me stress that it's usually not a hardware issue. Drivers, resource conflicts, and buggy BIOS/Firmware code is usually the issue. When it comes to "diagnostic" software for real hardware problems, there's little you can do. For memory, there's MemTest-86 [memtest86.com]. This program hasn't failed to spot bad memory for me yet. For hard drives, go to the manufacturer's web site to get a bootable diagnostic floppy. Usually the quick tests are all that's needed, but it could pay off to do the extended test if the quick one says everything's OK. Even if the extended test comes back fine, don't rule out the drive. The problem could be intermittent. For the rest of the hardware, simply swap components around until you narrow it down. Take out that video card and toss in a spare one that you've had for several years and know works. For the CPU, toss in any other CPU that matches the original's architecture (Just a different speed rating is OK). I think you get the idea on the (Swapping out) method. Also, never underestimate the crap a poor power supply unit (PSU) can throw at you! Stability issues could be coming from an unreliable current. Of course, temperature problems could also cause these same stability issues, but that's much easier to monitor. When you get to a point where you're truly stumped, some manufacturers have forums that you can go to for help. This is especially true in the homebuilt market with Abit, VIA, AMD, etc.

Good luck and happy hunting!

-Lucas

Re:Sandra

Good point! MemTest-86 also comes on bootable CD. I should also mention that it's licensed under the GPL (FREE!!!!). :)

Also, you can always make a bootable CD from a bootable floppy using most CD burning programs assuming you have one floppy drive somewhere that you could use for the transformation. Here's a good article [pcsupportadvisor.com] I found on doing this.

-Lucas

Re:Sandra

At least as it's packaged in Debian, memtest86 has a binary image you can just add to your lilo.conf so you can pick it when you boot. Neither a floppy nor a CD is needed.

Re:Sandra

As far as boot floppies, memtestx86, hard drive diagnostics, etc, I would send you to Bart's Boot Disk Creation [nu2.nu] site (the one mentioned the other day here for their Windows boot disk) and look into their "corporate boot disk" routine, especially the boot CD part of it. I heavily modified this (to allow the network and CD-rom to co-exist) and added things like hard drive diag tools from all of the major vendors (some were easier than others, some insisted on creating ISO images, so I need to mount that ISO, then extract their utils, etc). It already includes support for memtestx86, and I also added the off-line NT password editor (you can select multiple boot images, so just needed to add the linux one mentioned on other part of the site.

It also uses some cygwin tools (dd, etc) to make it so that you can create a bootable CD by emulating the creation of a boot floppy via DD, so you can create one w/o the need for an actual floppy (I would recommend hacking it to create 2.88MB floppies rather than 1.44 MB floppies (need to search its config files to handle this)

Re:Sandra

[memtest86] Let me send you my IProc PC-100 SDRAM DIMM... the idiots put the wrong timing values in it's SPD. I've only found one machine, ever, to work properly with that damn thing. Tyan MB's tend to lock as soon as the POST is complete. Memtest86 ran for 7 days and could not find a problem with the DIMM.

[hard drive diag floppy] I just returned two Maxtor drives that passed multiple "extended" tests with their diag utils. BOTH have entire tracks that aren't readable -- sector mark not found... they aren't there anymore.

Just because it passed a limited set of tests doesn't mean it's not broken.

Re:Sandra

I too have seen memtest fail to find bad memory, but only once so far. (I switched out memory and since then, no problems.) Also keep in mind, that due to the way it works, memtest errors could be power supply, motherboard, or CPU related in addition to memory related.

I've also seen the hd tests fail to spot a hard drive that was bad too. I spent over a week on one of them trying to figure out why it kept having problems. Since it was in my possession, I didn't have to worry out the cleaning crew causing the problem (see this story [computerworld.com] under neat idea..)

I've come to the conclusion that there is NO test that can say if the hardware is good, only tests that can say its bad. Between memtest and knoppix (and the hd tests), I can USUALLY spot faulty hardware quickly, but every so often..GRRRR

BTW, I haven't seen it mentioned yet, but the ultimate boot CD [ultimatebootcd.com] (free!) contains lots of useful tools all on one CD. These include memtest, and the various hd manufacturer's tests.

Brute Force Troubleshooting

I have been in the PC service and repair field for over 15 years. I've worked with PC technicians for several companies in several states, in a wide variety of industries (some of the companies provide PC technical support to customers, but others were in-house).

It is a shame that most PC technicians don't actually diagnose problems. Instead, they guess and swap, until the system happens to work, again. They don't really know what was wrong (though they probably will claim otherwise), and they certainly don't know if they fixed the problem.

I've met several technicians who claim that modern microchips are less-sensitive to electrostatic discharge than obsolete microchips were, but the microchip industry says exactly the opposite. Most PC technicians take very little, if any, precaution against electrostatic discharge. They assume that if the component works, it isn't damaged, and they lack the skills and tools to find any real damage. Instead, they simply swap out parts if something stops working.

I can't entirely fault the PC troubleshooting industry, though. Electronics are too cheap, most of the time, for technicians to spend very much time troubleshooting them. Speed is the most important asset in the PC industry. It is better to be fast than correct, whether troubleshooting systems or writing software code or technical manuals.

That might be reasonable for PC technicians, but one could find the same attitude in other troubleshooting industries. I just took my car in for repairs, because I often had to push-start it. This after a week of repairs for various problems. In that week, the mechanics never found anything wrong with my car starting, and this last trip dedicated to that problem was no different... until the mechanics got ready to return my vehicle to me. When they tried to drive back to the parking lot, my vehicle would not start. A new starter appears to have taken care of that problem.

Doctors are the same way. It costs far too much to find the real problem, I suppose, so doctors rely on rules-of-thumb and shotgun approaches. Many diagnosis are through the process of elimination; one treatment didn't work, so they try another. Doctors probably never know exactly what is wrong with the patient, but they often get close enough for the body to heal itself, to some degree.

Re:Sandra - NO GOOD

Sandra has some uses, but it, like a lot of other "hardware diagnostics" software, has a HUGE disadvantage.... It expects your operating system to be in good working order. It also expects that you have current (and appropriate) drivers installed for all your hardware. Any good technician already knows the tenet "Divide and conquer". When troubleshooting, you want to know if the problem is caused by the hardware or the software. If your software is screwed up, how can you trust the diagnostics to properly report that the HARDWARE is good?
Likewise, anyone can write a driver bad enough to make the video tests fail, but that doesn't mean the video card is actually defective.

Years ago, I worked at a PC manufacturer and we used QAPLUS FE. It was small enough to fit on a floppy disk, and had modules for all the independant subsystems: CPU, RAM, VIDEO, IO Ports, Timer channels, interrupts, Hard disk... You could select all the tests and let it run all night. If it failed on something, it actually gave you an idea on what might be the problem.
I would recommend QAPLUS if they had an up-to-date version that booted from a CD and had it's own KNOWN GOOD drivers for hardware. A Plus would be some sort of modular technology that would allow you to add drivers for more hardware in the future.

Low Cost

A low-cost alternative is a bootable copy of Knoppix [knopper.net], escpecially usefull if equipped with a virus scanner - like
http://www.linuxforum.com/linux_wallpapers_full/ 93.php>Knopicillin - sorry no ISO Image found - it was once in the C'T magazine [heise.de]...

Re:Low Cost

I totally agree. And the Gentoo [gentoo.org] LiveCDs all contain the excellent mentest86 program (type memtest on the boot command line) And also, the ISO is a lot smaller than the knoppix one (if you choose the bare-scraped one you'll just need 60 megs of downloading)

Re:Low Cost

Uhhh... no.

CDRW's don't boot in every CD-ROM player in existence. If you know you will be restricted to modern enough CD-ROM drives, fine, but if you want the broadest support, you have to burn on CD-R (even if you have to consider them as disposable items)

Re:Low Cost

If it can't get the hardware to work correctly, there must be a problem. In my experience, its difficult to tell the difference between a software and hardware problem on a windows machine because there are just strange app crashes. Whereas if there is _any_ problem with a linux machine, I assume its hardware. Also, the drivers in linux are much more verbose when they have a problem.

Since you mentioned Dell

If you like dell diagnostics, then you should probably buy PCdoctor. Atleast some of the diagnostics in the earlier versions of Dell servers were sublicensed from PCdoctor. Just go into the installation folder and liik at the DLL names, or read a config file.

Excellent Software

The Troubleshooter by SmartCertify direct. It comes as a bootable floppy, with a couple of dongles and a CD-ROM to test ports while in diagnostic mode. This has worked excellently for us...we were able to diagnose some odd, random computer issues as being caused by bad video RAM

Re:Excellent Software

I've used The Troubleshooter on numberous occasions and can attest to its ability to pinpoint most hardware related problems. It's a great program for anyone that does a lot of hardware maintenance.

Re:Excellent Software

I used troubleshooter for a couple years at a local hardware shop and we finally quit renewing our product license. There were just too many false-positives and missed errors. I don't know if that makes this one worse than the others, but we went back to simple by-hand replacement testing which worked quite a bit better. Of course we still used memtest86 and harddrive manufacturers' testers, those are invaluable.

Also of note, troubleshooter comes on one floppy disk and one backup floppy disk; it cannot be duplicated by any means I could discover, even for backups. "Not that there's anything wrong with that!"

They are all basically useless...

Nothing beats experience and a supply of 'known good' replacement parts. I have been out of the repair and troubleshooting business for years, but I always remember being frustrated at useless memory and system testing software that could not find anything wrong with memory chips, etcs, that were obviously bad. Even most hardware units (like ram testers) were almost useless. If the POST testing didn't find anything wrong, it seemed almost nothing else would either, most of the time. If you think the part is bad, swap it out with an equivalent and see if the problem goes away.

Re:They are all basically useless...

I don't disagree, but I can't help but think of the Jargon File entry for field circus [catb.org]:

Q: How can you recognize a field circus engineer with a flat tire?
A: He's changing one tire at a time to see which one is flat.

Q: How can you recognize a field circus engineer who is out of gas?
A: He's changing one tire at a time to see which one is flat.

Q: How can you tell it's your field circus engineer?
A: The spare is flat, too.

Biggest cause of failure in old PC: Bad contacts

Agreed: Try known good parts.

To prove to yourself that it really is fixed, use a memory tester, the hard disk manufacturer's disk diagnostics, and either a program that reboots an OS 20 seconds after it is loaded (on Windows XP, Wizmo [grc.com] from GRC.com and Sleep.exe from the resource kit) or, even better, some Linux or BSD build process that takes several hours.

The biggest cause of failure in an old PC: Bad contacts. Just move every card and connector 2 millimeters out and in again. The rubbing of metal to metal creates fresh contact surfaces. Renewing the contacts should be the first step in fixing any PC.

The biggest cause of real failure in a new PC: Infant failure. Components are more than 100 times more likely to fail in the first week than they are in the 100 weeks after that.

No kidding

I used to write diagnostic software for computer companies in the early 80s. If the computer was together enough to run my software there was not that much that could be wrong with it and very little I could find.

Keep in mind typical diagnostic sofware back then would test for things like memory not really being there (bad address or data line problem) or interrupts stuck on or not happening when they should or can't talk to the disk drive.

None of this crap really helps is you have a bad scsi cable (ouch, that was a long drawn out pig) or a bad cable or the wrong cache controller chip (ouch) or a bad power supply or wrong speed RAM any of which will cause a system to beheva erratically and in a - and this is the bad part - non repeatable way.

Back then almost every part was $8000, these days the answer to "how do I fix a flakey computer" is "buy a new one".

swapping out as a diagnosis technique

swapping out can really suck as a diagnostic technique. More precisely, it can really suck if not done very thoroughly. Horror story follows.

There was this little mom and pop computer shop, where I bought some stuff, and my buddy bought a whole computer, circa 1993-94. My buddy paid $800 for 16 megs of RAM. And he paid $250 for a honking big tower case, with wheels, and hinged panels, and almost a dozen external bays.

That honking big case was a honking big mistake. He never populated those bays, and it was too bulky to carry on public transit. Well, I had a car, and he didn't. So, when he had hardware problems he would beg me for a lift.

I can't remember how many trips we made to this store. A bunch of them concerned his flaky hard drives. The fellow replaced the drive, at least once.

Well, one time my buddy asked me to take his computer to this store for him. And this time I watched the owner's diagnostic technique. The first thing he did was take the drive out of the big honking case, and put it in his test rig. He also confessed to me that he wasn't replacing the drives any more, he would just test it, and if it was okay he was telling my buddy he had replaced it.

When I got the computer back to my buddy's place, I opened it up myself. I found that one of the pins on one of the power connectors had come loose. So it was only making intermittent connection. And this was causing intermittent problems.

Diagnosis through swapping out components failed here.

Re:swapping out as a diagnosis technique

Sounds more like they failed to swap components, not that the swapping of components failed.

WHat?!

I sure hope you are trying to troll, because I just don't get that at all.

If you come to a computer that needs work and you boot into a Knoppix CD you get...a desktop. On the other hand if you boot on to something like the ultimate boot cd [ultimatebootcd.com] you get a nice menu broken down by category of things like "File system utilities", "Memory Tests", and "Hard drive cloning". I just don't see how Knoppix can compare to that. I'm willing to listen though, if you care to explain what makes it so great for this.

Not many PC maintenance people use any diagnostics

And that's why... Most technicians do it by instinct and years of experience. If this peticular thing is happening, you know it could be one of x, y, or z.

That's always worked better for me than anything else. Although it would be nice to have something tell me what's wrong :)

Magic Smoke

PC diagnostics are really two-part work. Part one is to fix the PC, and part two is to convince the luser that the machine is actually fixed. That's where the diagnostic software comes in handy. Run the software, show the luser the "problem", fix the PC, and show the luser that the problem is "fixed". I've found that a piece of paper in a luser's hand is worth hours and hours of post-fix re-diagnostics because "something" has changed.

Sandra

Sandra is a good info/benchmark util.

For windows machines, I found a little app called RegSupreme which actually does a good job of cleaning/fixing keys in the registry.

Best "tool" for tech support is a good working knowledge of the PC. If you're looking for a piece of software to do support for you, then I'm sure the rest of the self proclaimed "IT Guru's" here at slashdot will warm a spot for you in the unemployment line.

I use make buildworld

This is slightly OT, but I've found one of the best ways to test (as opposed diagnose) hardware is to install FreeBSD then run "make buildworld" on it... If it completes with no problems, it's a pretty good indication that the hardware is in good condition.

Another vote for SiSoft Sandra

Rather irritatingly, the free version has lots of menu icons that don't work, a bit like blanked-off switches in base model automobiles. However, it's a good tool for basic performance measurements and reassures you that your RAM/CPU etc are working at their rated speed.

A tip: run it as Administrator or you'll get limited information out of the BIOS. And if you're using *nix, you'll have to look elsewhere.

Re:Another vote for SiSoft Sandra

Any diagnostic / troubleshooting software that requires that the computer boot into the operating system first is largely useless. It does nothing to help you fix a problem which is preventing the OS from booting.

Generic tools like memtestX86 are a good start, but there is a limit to what you can do with generic tools. Ultimately, you need hardware-specific tests. Hardware manufactures need to do a better job of providing diagnostic tools for the things they sell. Yeah, you can test the gross functionality of any sound card by playing music, but to do a complete test you need something designed for that specific card.

Re:I've found these useful.

Uh, yeah.. No, about half our customers' problems at our Support firm are over Norton Utilities or SystemWorks, or some other Symantec product breaking and screwing up the OS. Happens daily. They don't learn anything, but how to manually uninstall a Norton Product. [symantec.com]

I am looking at my sagging bookshelves that are

stressing under the weight of years worth of hardware & software Bibles. If there is such a miracle as a good diagnostic program that can keep up with the daily onslaught of new hardware, protocols and standards, then I have wasted a lot of my time but wish you the best of luck on your quest.

memtest86

The most common problems I encounter with PCs are memory related.

The best tools for checking memory are memtest86 [memtest86.com] and the follow-up memtest86+ [memtest.org].

Both of these are free to download and use. I usually leave them running for roughly 24hrs for a reasonable level of confidence. You should also burn-in the other major components too but memory is the best place to start.

#1-TuffTEST Pro, cheap, bootable

#1 TuffTEST pro is a cheap, bootable, hardware-only diagnostic. It supports all current x86 processors. It does not work on top of DOS or Windows or anything, so it's convenient for eliminating the hardware as a problem. Works great, I use it all the time. As a side note, if you use it on Dell machines, Dell seems to have an internal loopback on the serial and parallel ports. It will report the ports are OK even if they're not. http://www.tufftest.com/ [tufftest.com]

ultimatebootcd

I use the ulitmatebootcd [ultimatebootcd.com]. It consolidates several good boot floppy images onto one cd, including many free hardware diagnosis programs.

Forget expensive software

It is not possible to diagnose hardware by running software on it. At best you can determine if there is a hardware failure, but no software will be able to nail it down to a specific component all of the time.

Consider a motherboard failure for instance - a failing motherboard can in effect emulate any other hardware failure - ide controller bad? Your software may blame the hard drive. Bus problems can cause memory checks to fail.

I recommend you carry a simple bootable cdrom that loads the entire system (disk i/o, memory i/o and cpu load) and checks for errors. When a system fails these checks all it tells you is the problem is definately hardware and not a buggy driver or other software issue.

See BartPE [nu2.nu] for a good free solution.

The state of PCs

PC hardware is shit. Made as cheaply as possible, knocked out by the million. Nothing gets repaired - nothing is repairable. If it's broken, buy a new one.

PC software is shit. Software is still in the dark ages. No qualifications to show who has the first clue about quality, security, extensibility etc.

If you get any problem you can't fix in 30 mins, best to make sure you've backed up everything important (naturally you never need to ask anyone whether this is the case, because everybody always backs up their important data on a daily basis, right?), then just format or ghost the fucking disk. End of problem, and no tedious troubleshooting what happens when you try and get a LameSoft2000 graphics card working with a ShysterTronics printer.

Re:The state of PCs

The parent may be modded as "Funny", but speaking as a veteran of the support wars, it's actually about as "Insightful" as you can get.

Re:The state of PCs

Except that the origianl post was inquiring about HARDWARE problems. No amount of re-installing will cuse a faulty peice of hardware.

Re:The state of PCs

This is modded funny, but there's many a true word spoken in jest. Half the time, having a good backup of valuable files means precisely not losing hours to recover your machine. I've lost countless hours bringing my machine back purely because my backup set was too old and I wanted files back.

F.I.R.E.

Forensic and Incident Responce Environment [dmzs.com]

FIRE is a portable bootable cdrom based distribution with the goal of providing an immediate environment to perform forensic analysis, incident response, data recovery, virus scanning and vulnerability assessment.

Also provides necessary tools for live forensics/analysis on win32, sparc solaris and x86 linux hosts just by mounting the cdrom and using trusted static binaries available in /statbins.

diag software

...the best software i've ever encountered was inside my skull....

Barking up the wrong tree...

I have yet to find diagnostic software that is more reliable than my own expereince/instincts. I haven't really done an exhaustive search, but the handful that I have used tend not to work well (and take...too...long...).

Most OEM's are fairly accomodating if you describe problems in a decent amount of detail (and the machine is under warranty).

If these are white boxes, you're probably better off keeping a pile of spare parts around. A quick swap can get a machine up and running quickly.

Good Luck!

Hardware Specific

I will resort to diagnostics only when other troubleshooting is unrevealing, but the diagnostics for whichever hardware you have are usually provided by the manufacturer. For example, each hard drive manufacturer will have its own diagnositcs and if you expect warranty returns, you will have to run their program and tell them what fails.

Aida32

Aida32 [aida32.hu] is a great windows-based system information application. It's wonderfully complete --- I've used it countless times to find drivers for those 'unknown devices' (modems, sound cards, network cards) that windows can't recognize, and that I didn't want to take out of the machine. However, it doesn't really do diagnostics, such as checking memory or serial ports.

Free for personal use, businesses must register. Well worth it.

SpinRite and Memtest

The only effective hardware tests I've used in the ten years I've been supporting all kinds of hardware and software have been SpinRite [spinrite.com] and Memtest86 [memtest86.com]. Between these two, I can check for the most insidious and hard-to-detect hardware problems; i.e. flakey hard drives and RAM. A cheap $20 POST card is highly useful for dead machines. You don't need all the extra features the Microscope card gives you unless you are designing motherboards or doing some other such serious work. No software will replace your own experience and ability to know where a problem is forming based on the specific failure of the machine. All the rest of the so-called diagnostic software is more or less useless from a practical perspective, aside from testing serial ports with loopback plugs and printing cute certification reports for anal customers. This is detective work. You have to suss out the exact problem based on clues left by the failure of the system. Learn how the hardware works, and it's easier than you think.

DisplayMate

for tweaking your display. An absolute must have.

Flawed Concept.

I have tried to find this, but the problem with getting flawed hardware to run software to detect it is flawed is just that in concept.

First of all, if it is an issue with hardware, the machine may not boot at all. If it is a ram issue, the diagnostic software may generate errors.

Second, even if it highlights an error in a configuration, it could be generated with the analysis software.

Third is that failed hardware often will not register as failed unless it is operating. Such as, a failed modem will not become noticable until it is used and then it may lock the computer up which could stop the software diagnosing the issue.

Your best bet is to use a cause and effect analysis. Then trial and error. The machine won't boot, find every possible cause of it not booting and eliminate each one as a possible cause. Continue on this until the issue is completely solved. Make a checklist for yourself so you don't forget anything.

It is how I do freelance repairs and it has proven bulletproof compared to the Voodoo Computer Repair Experts that try random things in the hope that it fixes the issue. (Install drivers, reinstall OS, Check CPU)

Hard Drive Diagnostics

I've heard and read great things about Spinrite from Steve Gibson, for hard drive diagnostics and I have friends that swear by it. The only drawback to his current version is that it won't do a thing for NTFS partitions. That being said, he is working on Spinrite 6.0 that WILL read NTFS partitions. This new edition is due out any time now. If you buy Spinrite 5 you get a free upgrade to 6 when it becomes available. As far as a general software based, pc diagnostic tool, I've yet to see any that actually work with any degree of certainty. SiSoft Sandra, as others have mentioned, is pretty good at letting you know what hardware is in the box. That along with some available "known good parts" is what I do to diagnose issues. I have also found the book Upgrading and Repairing PC's to be particularly valuable and a must have in my collection.

Some products

As others mentions, memtest86 and knoppix are invaluable tools.

Other tools you might be interested in;
Aida32 [aida32.hu] basically lists all of your devices, drivers, wmi software entries, pci devices, etc. for windows - needs an install, though.

OnTrack [ontrack.com] sell Easy Recovery Professional; the "file repair" options are pretty crappy, but for serious, near-forensic recovery on fscked up filesystems, ERP is a fine tool. Some of OnTrack's software (i.e. SMART tests, usually) may be licensed by the manufacturer of your harddrive, so check those pages out.

SiSoft Sandra is recommended a lot, but I don't find it offers a lot of diagnostics, though it is prone to crashing.

On windows, you might want to check out the Event Viewer, hidden in the Computer Management section of the (classic) Control Panel -- it will list all sorts of errors and notifications, kind of like /var/log/messages ;-)

OT, but I have a tough hardware problem...

I've taken an ailing HP Omnibook 4150 laptop to my local computer repair place where I was told (in this order) :

1. I needed a new power supply.
2. I needed a new motherboard.
3. They didn't really know what was wrong with it.

It suffers from intermittent power failure, otherwise it runs fine. I wish I knew how to locate the trouble or if I'm just wasting my time thinking this machine can be fixed. I'm loathe to take it to another repairman, I'm already out some $$ that got me no closer to a real solution. I hope this is an appropriate question to ask, 'cause I like that machine and would rather not junk it. Any civil advice will be vastly appreciated (including suggested URLs for diagnostic tools such as those mentioned in the original article). TIA!

Btw, the repair house told me that their "diagnostics" consisted of letting the machine run for a day or two. I paid them their bench fee and swore I'd never take another machine there again.

AIDA & PC-Config & NSSI

A great, free, Windows program is AIDA32 [aida32.hu]. It gives lots of valuable info. It's not perfect, but it's constantly in development and improvement. Up until about a week ago there was a DOS/16-bit version available, but due to lack of demand it was discontinued (sadly). Another ok program is PC-Config [holin.com], which is no longer being worked on, but it's pretty good. And NSSI [navsoft.cz] is really pretty nice as well.

Check those out.

My advice

Knoppix - verify the part under another OS.

Tomsrtbt - I forget if Knoppix has badblocks or not. If it doesn't, Tomsrtbt does.

memtest86 - Memory tester.

Spare HDD - good for having a clean install of windows to check things on.

Spare low-density memory.

Spare older computer for testing daughtercards.

That's about it.

Of course, sooner or later you *will* get the machine from hell with an intermediate fault that ends up locking windows for no damn good reason every so often. Then life will suck. But that's why they call it work.

Always pay with a credit card

If you pay with a credit card and the vendor tries to screw you over you can contest the billing and return the unsatisfactory product, In fact I am in the process of cancelling an order i placed for Serif's Photoplus 7 because the site selling it never mentioned that it was old and shitty, but they were perfectly happy to tell me how bad it was when they called me asking if i wanted to pay $50-$90 to upgrade to version 8 or 9... obviously contesting the charge is a last resort but overall I feel much more confident buying software online because i know that i am not out the money untill i actually pay the credit card bill (for internet purchases there is no signature)

All those tools suck

I was a PC repairman for many years. I say from experience that all those software diagnostics suck.

Microscope from Micro2000 is actually the best of the bunch, but that's not saying much. If your computer won't even post, non of these tools will do you any good. (They do have some good training materials for those wanting an A+ or N+ COMPTIA cert.)

The PCI cards that display diagnostic codes are better than the software in those cases. They still aren't very helpful though. Basically they will tell you there is a problem with the memory, or the parallel port, etc., but they won't tell you exactly what's wrong so they aren't of much use either.

Here's my advice:
1. Get the power supply tester from PC Power & Cooling. It's $20, and in my experience most of the time the reason a computer won't even post is because the cheapass power supply that came with the case died.
2. Carry a bunch of known good parts: an AGP and a PCI video card you know work, a PCI network card and PCI modem, some known good RAM (PC 100 and DDR), and a good hard drive. Ideally, these are all in a fully working computer you've brought to the site so you can swap between the working computer and the not-working computer and narrow down the problem. Resist the temptation to fix the system with your known good parts; make them buy new, name-brand components with a warranty.
3. Bring a USB keyboard and mouse. I've seen lots of 3+ year old computers have their PS/2 connections short out or stop working but their USB ports are just fine. They can solve input problems.
4. Have a Knoppix CD in your kit. The linux forensic toolkit can be of great use recovering files and finding problems.

It has been my experience

That the free stuff is just as good (if not better) than high-priced diag software.

Allow me to give you some background:

I have done IT work for 4.5 years. I work with Novell, RedHat, all (disgusting) flavours of Windows, BeOS, Sun, SGI, Apple (Mac) and QNX. I support everyone from Joe Grandma to major Universities and Medical Colleges.

I have several CDs worth of useful tools at my disposal, all of them free:

Ad-Aware: I consider this to be my single best resource in the fight against Windows NT (and up) flakery.

www.trendmicro.com does an on-line virus scan. Not perfect, but usually finds the major ones.

Demos of Anti-Trojan. Again, good enough for the closing of trojan ports left open.

AVG Anti-Virus software. Good, free AV software, if Norton isn't available.

Winzip: Obviously a good thing, many many drivers come zipped.

A CD full of the most common NIC drivers from the biggest vendors.

nVidia and ATI drivers.
Via drivers

All the latest browsers on another CD.

MemTestx86 (as you have found): Allow me one point further int he favor of it, major memory makers will accept their RAM bad, no questions asked (in my experience) if you tell them it was checked and found bad, via MemTextx86.

SiSoft Sandra, if for nothing else than the CPU-Burn wizard. If the CPU is bad, Sandra will find out.

Emergency Boot disks and cd-rom access disks (sadly, the Win98 boot disk is pretty handy)

A live Linux and live BeOS CD (very handy for recovering data of hosed systems)

And last, but not least, a good Google search. Another thing that has saved my skin time and again is to input exact error messages and see what Google turns up.

This whole cd-wallet has set me back perhaps $20, and does far more than "professional" diag tools can hope to accomplish.

Best hardware diagnostic

In my experiance, the best hardware for diagnosing PC hardware problems is another PC that you know works. Quite frankly, the built in diagnostic capabilities of PC hardware is extremely poor, the only reliable way to discover if a part if bad is to replace it with a known good part and see if that fixes the problem. Don't blindly trust new or especially refurbished parts either, they need to be tested and known good. With experiance you will know what parts to try first, but it is still important to verify that replacing the failed component with a new part fixes the problem.

I remember back at school where the tech support guys were ripping their hair out because a lot of the school issued PCs were coming back with random crashing problems. (I had a roommate who's machine would crash everytime the screensaver kicked in). They were replacing parts left and right and it never seemed to fix the problem. My roommate had everything replaced except the case at one part, and it still crashed on a vanilla Win3.1 install. It took them awhile to realize that most of the machines had bad memory, and the vendor supplied replacement memory for the systems was usually bad as well. I eventually loaned my roommate my memory sticks, and when his system didn't crash he went back to the PC guys and told them exactly what the problem was and made them continue swapping in RAM sticks until they finally found one that worked (apparently the RAM was OK in their hardware RAM tester, but failed once it was actually put in PCs. They suspected the same thing was happening at the vendors end. They would get bad memory, test it and not discover a problem, then ship it right back to the school.

IBM'S PC Doctor

We are an IBM shop (By state mandate) so I am a bit biased. You said manufacturer versions are model specific, but I doubt IBM's is, because they use the same PC doctor program for the last 10 years worth of laptops, workstations and x86 based Servers. It comes on a bootable floopy (or on servers a bootable partition just for it). It does full HD, floppy, CD, and memory scans. It will test the standard AT external ports (Parallel, Serial, PS/2) It runs VGA mode scans, and will test some features of the AGP port. Admittably it won't test stuff like sound cards, or many built-ins, but I doubt any Paid program would either.

To find it go to IBM support pages and start looking for Diagnostics, you will eventually find it.

Alternative medicine

Has anyone found something that works for you?

I have found variations of the Hopi rain dance to be effective in solving many PC hardware problems. Tibetan chanting also work well.

Re:Alternative medicine

A friend's (now ex) girlfriend had an issue with her computer a few years back. She'd had somebody try to fix some problem, which he fixed, but in the process managed to make the cd-rom and soundcard unuseable. So he and I set about trying to make everything work. A few hours in, we decided our brains needed a break, so we went swimming. Laying in the sun made us thirsty and lazy, so we started doing shots of tequila. Hours later, we drunkenly went back to work and decided wicca was the only solution. When we sobered up we had ashtrays full of burnt down incense sticks, candles melted into puddles in a circle around us, and a computer with a fully functional cd-rom and soundcard - and no fucking clue how it got that way.

Ultimate Boot CD

www.ultimatebootcd.com

Free and includes loads of software. .::Dread

For exhaustive memory testing, Suntory Memtest86

memtest86 [memtest86.com] gives you a bootable floppy that will scan *all* RAM in a system, and turns up the most obscure memory errors. Some errors are not consistent, they only appear in transitions from one bit pattern to another, for example. Or adjacent bit cells may bias the bit in question.

That's all it does, but it's good. And it's free. One other point, systems with mismatched parts (designed for different bus speeds or timings), and overclocked systems, may generate memory errors. Since I started using memtest86, I've stopped overclocking, as every single overclocked system I've checked has shown errors under memtest86!

QuickTech, OnTrack ESR and SpinRite

I'm lead tech at a smallish computer shop. Over the years we've used just about everything to aid us in diagnosing hardware issues. The ones we use most are:

QuickTech Pro - This is great for testing memory, serial ports, video RAM, and just about every other problem that you can experience. In the past we've had so many problems with recieving bad RAM from our vendors that we now run the quick memory test from QuickTech on every PC before it leaves our shop.

OnTrack is simply the best when it comes to data recovery. They aren't cheap, but for a small-medium business their software is well worth the price. We've also purchased some of the MS Office add-ins to help us recover corrupted Word/Excel documents better. This program really works, and if it can't recover something off a failing drive (or something that a user deleted by accident) then Ontrack will almost certainly be able to recover the data at their facilities (for a steep price of course).

I'm not a huge fan of the guy that wrote SpinRite, but if you've got a FAT(16/32) partitioned drive that seems to be failing, this tool is great. It will recover bad sectors off most drives (not seagate), especially the ones that happen when Win9x doesn't get shut down properly due to a power surge or the like. This software is free and takes a very long time to run at its highest level.

Other than these 3 programs and a few other niche utilities we generally diagnose all other problems by having known good hardware and swapping out to see if the problem still exists. It is still, without a doubt, the best way to diagnose a hardware problem other than Technician's intiuition.

My favorite PC repair tool?

These wonderful utilities [amazon.com] are versatile, come in many sizes, and assure me that my @!$%@&! PC never, ever gives me an error message again!

OT: Get it in writing

Anytime you accept something with the option of returning it, get it in WRITING..

swerving back on topic: perhaps its time for a set of OSS tools for diagnostics.. some parts exist now, such as memtest86.....

CheckIt Diagnostics by Smith Micro Software

I'm surprised that no one has mentioned CheckIt Diagnostics from Smith Micro Software (http://www.smithmicro.com). You will find it on their website listed under the Utilities section. CheckIt Diagnostics only costs $70. I bought my copy at a local computer store a few years ago.

What I like about it is that in the same box you get both a version that boots and runs from a single DOS boot floppy, as well as a Windows version on a CD ROM. I almost never use the Windows version because the DOS version is very comprehensive. The only complaint I have is that I can't find a way in the DOS version to make the diagnostics run "n" number of times. Their memory test is very thorough, but I'd like to leave the whole thing running for a few days. Maybe they've corrected that in newer versions?

They don't have an eval version on their web site, but maybe you can contact their sales department about getting one. Please note that they now have a product called CheckIt86 which has nothing to do with the DOS version (e.g. 8086) but is an ad-blocker.

Free Microsoft Windows Memory Diag

Microsoft offers a freebie memory tester [microsoft.com]

Supposedly pretty good, it boots off a floppy or CD so it looks like it will run under Linux. I haven't had a chance to try it myself yet.

Is one tool at a time OK?

Google for memtest86.

robust, reputable, and Free

Have you tried using the Linux Test Project [sourceforge.net]? I administer a cluster of Linux machines, and use LTP as a pretty comprehensive test framework. Many of the tests are software related, but you can shut those off if they're not useful to you.

I suggest you make a Knoppix [knoppix.org] CD with LTP installed. With a little configuration, that will take care of all of all your tests for the memory, disk, IO, and CPU. You might want to install America's Army [americasarmy.com] or something to test the video subsystem.

If you put a little effort into it, you'll have a test suite as good as, and likely better than, anything you could pay money for. If you want to buy something, you can make a donation to the LTP and Knoppix projects.

There are also simpler tools, like Memtest86 [memtest86.com]. I find this tool to be invaluable when I try to salvage old hardwar. I can't begin to tell you how much time it's saved me that I would have spent aimlessly swapping components around.

Cerberus

Go get Cerberus [sf.net]. It's the burnin and diagnostic suite for Linux (even if you don't intend to keep Linux). It's been a de facto for VA Linux and for the Linux kernel developers. Just install a complete development environment, including a compilable kernel in /usr/src/linux (make a symlink to linux-2.4), type './newburn', and walk away for about a week ideally. At least 8 hours. The longer the better.

Just... listen...... to the screams......

The SMOKE test

Turn it on and see if it SMOKES. If it does its a hardware problem.

Best utility is detective work

I've worked as a technician for about five years, both on-site and in-house, and I really haven't found any good all-around utilities. Most of the diganostic programs mentioned here tend to test things that hardly ever test bad in reality, such as serial and parallel ports, timer chips, address lines, etc. The best thing to have around in case of a rather obvious failure (failure to power on, blank screen, beeping) is a collection of known good parts that can be swapped in to isolate the problem. They can also be useful in cases involving more subtle or intermittent problems, such as a CD burner that fails half the time or a computer that locks up after a few hours. Heat problems can be subtle -- I remember a friend telling me that the computer he shop didn't have heat for a while during the winter because of a bad relay, so there was a case where a computer with a slow fan worked at the shop, but didn't work when the customer took it back to their warm home. Sometimes a USB device a customer installed won't work because they plugged it in before installing drivers, and now it needs to be removed from the registry and installed again. In Windows, sometimes certain programs don't play well together, and experience will tell you what programs to avoid, and which processes in your task list don't belong. The point is, certain utilities may have their place, but no utility can replace the ability to diagnose a problem through experience and process of elimination. A memory test utility may show that you have bad memory, but experience may tell you that it's bad anyway, even when the program says it's okay. The good news is that I believe quality is improving again, after hitting a low in the K6-2 era, when cheap motherboard production proliferated, and the problem was compounded by the large increase in cooling needed for those chips (I'll never forget PC100 mainboards (yes, that was the company name, not the memory speed)). It's also pretty cheap nowadays to have spare parts on hand for quick checks.

How to tshoot most PC's in 15 minutes or less

This will work 90% of the time:

Prerequisites:

(1) GHOST'ed hard drive with Windows 2000 installed, and the Sysprep utility enabled on it

(4) Processors: a Socket 7, a Slot A, a Socket A, and a Socket 478

(4*2) RAM modules: a couple of 72 pin SIMM's, a couple of PC-66 SDRAM's, a couple of PC-133's, and a couple of DDR's

(3) Power supplies, an AT, an ATX, and one of those new ATX'es

(2) Motherboards, 1 AT, 1 ATX

(2) video cards, 1 PCI and 1 AGP

Step 1:

Replace user's HD with your SYSprep'd one. Boot. If you can boot, Win2K will do it's hardware detection routine. If it finishes, and boots to a 2K desktop, your problem is probably software and you have to narrow down from there. If it's using Windows, boot with user's hard drive in safe mode. If problem does not occur, disable taskbar lint under HKEY_LOCAL_MACHINE/Software/Microsoft/Windows/Curr entVersion/Run and from user's Startup folder. Reboot. If the problem re-occurs, probably virus'd / spyware'd or corrupt Windows install. Re install Windows. Fixed. If it isn't fixed, follow step 2

Step 2:

If the user's machine did not go past the Win2K hardware detection routine, then the problem is hardware. Go into the PC's CMOS and reset to Setup Defaults. If that doesn't work, CMOS is OK, it's a hardware component. The list of problematic components in a typical PC, listed in desending order of probability is:

• RAM
• CPU
• Power supply (yes, power supply)
• Motherboard
• Add on cards, including video

Go into the CMOS, disable L2 cache on the motherboard (common problem and you don't have to swap out anything). Try Win2K detection routine. If it works, motherboard L2 cache is the problem. If it doesn't, swap the ram. If it works, it's RAM. If it doesn't, swap the processor. If it works, it's chip. If it doesn't, swap the PSU. If it works, it's PSU. If it doesn't, take out all add-on cards and swap the video cards. If it works, swap back each add-on card one by one until the problem re-occurs. When it re-occurs, the last card you swapped back in was the problem. Replace it, and you are good to go.

Using this technique, I can troubleshoot 90% of PC's in 15 minutes or less, 90% of the time. HTH.

QuickTech from UXD

http://www.uxd.com/qtpro.shtml

Our company swears by it... every machine before it goes on the bench and after it leaves goes through fully automated and comprehensive "Burn-In" process. I'm not sure how much it costs, but its definitely up there around $400 maybe? and comes with bunch of loopback interfaces.

it's pretty up to date too, last version we have has serial ATA, athlon 64 etc support

I can't even count how much time we saved when after initial burn-in we realized that the memory was set to CAS 2.5 instead of 3 causing timer problems and subsequent weird application crashes.

A tool like this is very important if you have many crap machines coming from the street and you dont know who worked on them before (Joe Sixpack thought that lower CAS will make his computer go faster, but his el cheapo memory modules just wont take it)

excellent (free) windows utility

http://www.zhangduo.com/udi.html

unknown device identifier, it will give you the PnP info on your hardware, so youll know what sound card is in there *without* having to pop the case.

& its only 800k, so you can send it to someone & have them run it

while sleep 1; do make clean;make -j zImage; done

No, seriously, the best non-synthetic test of a system's datapath IMHO is to build the Linux kernel repeatedly. GCC is quite RAM and disk intensive, and so stresses the most flakiness-prone parts of a machine. Plus it gets the CPU pretty hot. This, in my experience, will make a machine with marginal memory, clock settings, or even heat dissipation fall over.

Re:to start you off

Memtest is a great tool, however it is specificaly mentioned in the article itself.

If you read the article, RebornData is looking for something more comprehensive than memtest offers. (ie: more than just a memory test. I assume to include disk, bios, video, cpu information, and a variety of other system tests and checks.)

I myself question the need for much more than a disk-surface-scan tool and a copy of memtest, but it's what RebornData is looking for.