Help! Before I throw the damn computer out the window...

[scroggs]

Any of you hardware wizards out there able to help me on this one?

My desktop computer has developed a habit of freezing completely, and occasionally even doing its own hard reset, for no apparent reason. There is no consistent prompt for this behaviour; it can be when there's nothing on the desktop, or when there's all sorts going on. With or without sound, 3D graphics, whatever, it just stops.

The hardware is a Athlon T'bird 1.2G on an Abit KT7A(133) motherboard, running a single 256mb 133Mhz SDRAM. The i/o cards are a Creative Soundblaster Live Platinum 5.1, Creative GeForce GTS 32mb video, and Motorola V90 Speakerphone internal modem. The drives are two Maxtor HDDs, one 40Gb, one 4Gb, a generic DVD, and an internal EIDE Zip 100mb drive. External peripherals include a UMAX 600P scanner, Epson 900 usb printer, and a Gateway Vivitron 15 monitor. The OS is Win98SE with IE6.

Any useful thoughts..... please!! <img src="frown.gif" border="0">

Edited to mention that I did a full format and reload of most of my software today - no change! Thank God I've got this laptop, or I'd be Ppruneless!! <img src="smile.gif" border="0">

[ 10 January 2002: Message edited by: Scroggs ]</p>

[FL310]

Tell me which window and which building...will be there to collect.... :) :) :)

No, this sounds like
a. power interruption from the power supply
- a lose cable
- a bad cable connection on these cheap plugs
- the fan has stopped and overheating sensor stops the unit

b. surge in the internal power system due to high load....

c. software screw up from a driver

d. a virus which you did NOT delete while re-formatting (try to erase the partition with FDISK and than format....)

e. the motherboard is touching the cabinet somewhere, thus you have a shortage occasionally

f. fan of CPU is not running and CPU overheating

g. CPU is not ok....

so...now guess...

[Man-on-the-fence]

I had symptoms very similar to this.
Get inside and mae sure that the IDE cables are very firmly pushed into both motherboard and the Maxtor drives.....sounds like it could be socket creep resulting in lose cables.

[WhiteSail]

I don't think you will ever cure this totally, but make sure you have the latest VIA 4 in 1 drivers (NOT BETA) and the very latest drivers for your video card.

Its an on going problem with VIA chipsets.

If you ever find a definative cure, I would be delighted to know!

[Evo7]

Did you build it yourself?

Given the freeze/reset you describe I'd go for

[quote]
f. fan of CPU is not running and CPU overheating
<hr></blockquote>

although as there is a fair amount of kit inside the case the CPU fan may be working OK but is not getting enough air circulating to do a good job cooling the CPU. Also, is the power supply well within capacity for the kit it is running??

However, I wouldn't dismiss WhiteSail's comment about VIA either. There have been 'issues' for a while with their chipsets, and I wouldn't be surprised if that turned out to be the cause - although I'd still go with my guess.

[scroggs]

Thanks for your suggestions, people. I've taken the thing completely apart and reseated every component, and I'm running it with the case open at the moment. I have been monitoring the CPU temp, and it's never approached overheating, and all the internal fans (4!) are working correctly. The power supply suggestion is interesting; this case has a 145W supply which may be a bit marginal.
I forgot to mention that the computer rarely starts up first time, hanging before any beep codes are output. Pressing the reset switch always cures that problem - till next time.
I did build the computer myself, or at least have continuously updated it since it began life as a Gateway P166. I think I'll have to try replacing each component in turn to see if I can isolate the faulty item, although I'll try the new drivers first. I'll also try new HDD cables, as the ones I've got look a little tatty - these new 80-wire cables seem quite fragile!

[ 12 January 2002: Message edited by: Scroggs ]</p>

[fobotcso]

Re: Evo7's "fair amount of kit". It's been said that poorly routed ribbon cables can be the cause of component overheat because they obstruct the airflow. So much so that some manufacturers are now selling bundled ribbon cables where the cross-sections are more rectangular than flat. There's even a site with detailed instructions on how to slice a conventional ribbon cable lengthways(usually one of the IDE cables)and tie the five sets of eight(?) wires one on top of the other. However, this is not recommended by serious gurus as electrical interference could result from crosstalk.

{Edit: Oops, Scroggs, our posts crossed and mine is now not very relevant but I'll leave it there for posterity. Are you using 80 wire IDE ribbon with a 5400 rpm drive? Should really only be necessary with 7200 rpm drives. Why not try the old ribbons if your drives are 5400?}

[ 12 January 2002: Message edited by: fobotcso ]</p>

[WhiteSail]

If you are confident you have the correct figure for your power supply (145w), you are seriously under powered.

No way would AMD recommend such a low power supply for a Thunderbird chip.

Look for approved AMD power supplies. Someone like Enermax make well specified units. If I were you, I would go for at least 350w, which will give you a little room for manoeuvre.

I am amazed it is running at all on 145!

[carbheatcold]

I would replace the power supply first, as WHITESAIL has pointed out, it is very underpowered. I imagine that if the PSU is running at max chat and beyond, if it is old, then there is likely to be AC ripple on the DC voltages which are creating all kinds of problems for the other components (memory,CPU etc). Unfortunately unless you have access to an oscilloscope you will not be able to easily confirm this.
With all modern computers this ripple can cause all manner of weird and wonderful things to occur.
Have you added anything recently that may have contributed to the problem?

[ 12 January 2002: Message edited by: carbheatcold ] Edited for spelling and grammar

[ 12 January 2002: Message edited by: carbheatcold ]</p>

[Mac the Knife]

Agree about power supply - 145W isn't marginal for a Thunderbird, just plain insufficient.

A good power supply makes a surprising amount of difference to stability so spend a bit on it.

Good luck!

[scroggs]

Thanks guys, again. The 40Gb HDD is a 7200rpm drive, and the smaller one is the secondary on the same EIDE slot. None of the ribbon cables are obstructing airflow. I am confident that 145W is correct - its an ASTEC power supply with quite a lot of info on it, with MAX POWER OUTPUT 145W in large letters on the side of it! I was surprised it was so small as well, so I guess that's the first option. I have, incidentally, loaded the VIA 4in1, new NVIDIA, Soundblaster, and Iomega drivers.
I reckon it's time to retire the old Gateway tower case! I was getting fed up with the inconvenient access to the internals, anyway.
Thanks again.

[HugMonster]

Scroggs, with an Athlon CPU you have access to all their monitoring software - I think you've already made reference to it. Have you also checked the power usage on it? ASUS's website has quite a lot of diagnostic software for my processor, which admittedly is a bit later than yours (but not much). I get the impression they're still slightly nervous about taking on Intel...

On first reading your post, "Power Supply" was my first thought. I still suspect that's where your prob lies. Go on, you old skinflint - you can afford a new case/motherboard!

All the best

[FL310]

NEVER EVER RUN THE MACHINE OPEN FOR A PROLONGED TIME...as strange as it sounds...like in any aeroplane engine, the airflow is disturbed and you will have trouble...believe me...and put at least 250 Watts in....cheap cabinets have cheap power supplies...that is the difference...
Forgot something...your starting problem indicates the power problem and as a second solution to this, do not run the monitor over the power-out adaptor, use a separate plug to mains, the surge may have worked on your switch already and also does not allow the power available to start-up the power supply while the monitor draws all the current...oin a second attempt, the monitor is already on and not drawing so much current anymore thus allowing the power supply to start "normal" but under a heavy strain from all your bits and pieces installed...

[ 12 January 2002: Message edited by: FL310 ]</p>

[Evo7]

For what it's worth, the IBM P4-1.6GHz that I've got here is using a 350W power supply to run the motherboard, two hard disks, a CD-RW and a DVD-R. 145W for a roughly equivalent set of hardware in your machine does sound too low.

A couple of years ago, I built myself a dual Pentium-II box that suffered from similarly random lock-ups. That turned out to be an underpowered power supply - 250W IIRC.

[ 13 January 2002: Message edited by: Evo7 ]</p>

[scroggs]

Hugs,
yes, I can probably stretch to a new case <img src="wink.gif" border="0"> but I seriously hope that a new motherboard isn't required - this one is only 4 months old! Haven't found a power-consumption meter within the Abit software, but I'm sure there is one - I'll look.

FL310
You're probably right about cheap cases, but this one's problem is age, not price - the system it surrounded cost me over £2000 from Gateway a few years ago! Incidentally, the monitor is independently powered. I question your refernce to in-case airflow. I see no evidence that any thought has been applied to the airflow within this case. In any case, the profusion of on-chip fans that accompany the kit now on board are likely to have completely rearranged the standing flow patterns within the case! Running with the case open has consisitently reduced both the CPU and system temps by 2-3 degrees (40/37 and 33/30 respectively).

[Mac the Knife]

Hmm. I'm surprised that you managed to fit a modern motherboard into an old case. As regards the airflow, you are right, the AT and Baby AT cases were not designed with any particular airflow in mind. Modern ATX cases are a better, especially the tower designs, which allows more sensible airflow patterns anyway. Basically the air comes in at the lower front and exits at upper back. I bought a huge server tower case for my PIII/850. The case itself has 3 fans, one for the power supply at upper back (exhaust), one just above the PS (exhaust too) and one lower front (suck) and this keeps it cool. It's also a lot easier to work on stuff in there. And there are plenty of extra holes for removable drives/DVDs/CD-RW's/tape streamers etc.
Only thing is that you may end up having to get special long 80 conductor IDE cables to reach 'em (hard to find and pricey) - btw, max length of IDE cables is supposed to be 18", but I'm using 24" without problems.

Running with the case open is fine unless you're seriously overclocking, but by then you get into the realms of huge forced airflow fans, baffles, Peltier cooling (even water-cooling!). Gets very complicated and atrociously noisy.

[FL310]

if you have a look at your power supply you will easily find out how the airflow should go...in case you leave the machine open, you actually disturb this airflow and believe it or not, you will face problems with the unit as it overheats. Further, the machine is easily able to pick up more dust as it already usually does, have a look at your fans and you see how much they have already accumulated. Don't tell me that this little bit will not interfere....

You asked for help...well my over 25 years experience in building and maintaining these nasty machines speaks here. There is still a company in a very hot subcontinent which I run and there we face dust and heat more than you may imagine. There is a little knowledge behind my words....

But anyway, with this power unit you will never be able to run a smooth machine. Or do you think that Boeing is sitting on some nice 737-200 engines and considers them to fit on the 777?
Adjustment and selection to combine suitable parts into a proper working machine, that is the task which even most manufacturers (not only in the computer business) are unable to perform..

[scroggs]

OK, new AOpen H600A case ordered, with 300w power supply and a suitable selection of fans, so that should sort it, hopefully.
A310, I'm not having a go at your knowledge, just commenting that in my Gateway case (one of the first ATX cases, from 1995) there is no obvious thought applied to the airflow. The only reason I've run the system with the case open is to attempt to diagnose the problems I referred to in my first post! Once I'd discovered that new drivers wouldn't cure the problem, the system's been closed down.
Anyway, I'll let you know if the new case/PSU doesn't sort it out!

[FL310]

all the best and good luck with it...i am very sure that this decision will cure most of your problems.

[Rollingthunder]

At this point I'd be chucking the bastard machine and starting anew, cheaper and with more bells etc., pre-installed at reputable point of purchase.If you can find a point of purchase with real support integrity.

[Evo7]

[quote]
If you can find a point of purchase with real support integrity.
<hr></blockquote>

They certainly exist, but the majority of home users will go elsewhere. Good support costs, and given the choice of two systems costing the same most people will pick the one with a better specification rather than the one with better support. Can't have both.

Buying on specification is fine if you're skilled enough to deal with problems yourself (which I can) - and more importantly if you have the time to do it (which I don't). Most people aren't able to fix their own problems, of course, but would still prefer to pay for a faster CPU they almost certainly don't need (2GHz vs. 1.6GHz - how many people can really tell the difference?? Some, sure, but not many) rather than the support they almost certainly do....

<img src="rolleyes.gif" border="0">

[scroggs]

Well, new case and power supply now enclose and feed my system and.... I still have the random freezing problem, although the unprompted hard reset problem has gone.. .All the suggestions you've come up with so far have been carefully tried and eliminated other than changing the motherboard and/or CPU - and the diagnostics I've tried can't identify any problems there.. .Any other thoughts?

PS: I've disabled all the junk in the systray - no difference.

[ 21 January 2002: Message edited by: Scroggs ]</p>

[Evo7]

Scroggs

Might me a long shot, but there has just been some news about an Athlon bug.e.g.

<a href="http://www.theregister.co.uk/content/3/23749.html" target="_blank">http://www.theregister.co.uk/content/3/23749.html</a>

<a href="http://linuxtoday.com/news_story.php3?ltsn=2002-01-21-001-20-NW-KN" target="_blank">http://linuxtoday.com/news_story.php3?ltsn=2002-01-21-001-20-NW-KN</a>

Symptoms involve random lockups when using AGP - sound familiar?

[quote]. .Here are the details. As you may know, x86 systems have traditionally managed memory using 4K pages. However, with the introduction of the Pentium processor, Intel added a new feature called extended paging, which allows 4Mb pages to be used instead.

&lt;snipped&gt;

Many Athlon and Duron CPUs experience memory corruption when extended paging is used in conjunction with AGP. . .<hr></blockquote>

It is mainly reported as affecting Linux users, but that is because a Windows fix already exits - however, it is a CPU bug rather than a bug in an OS.

From The Register:. . [quote]. .AMD issued a patch for Windows users in Fall 2000. .<hr></blockquote>

Are you patched?

Alternatively, if you can, try swapping in a PCI graphics card and see if that helps - the problem occurs only with AGP.

All the best...

[scroggs]

Evo,. .thanks for that. I'll try a PCI graphics card (I have an old Matrox G100 around somewhere). The CPU was made in about July 2001 (just before the Athlon XP came on line), so hopefully that bug shouldn't be present - but I may well ask AMD for their thoughts.. .I've heard elsewhere that the VIA/Athlon T'bird/Win98SE combination is beset with problems, so I may try Win XP to see if things improve.

[allthatglitters]

Several colleges had Athlon machines and had problems of freezing.. .The first located a news group and found a very long list of people with the same motherboard and the same problems, (confirmed today an ABIT KT7A) board replaced, function checked satis.. .The second returned it to the point of sale and brought a P4 machine.. .I, on the other hand am more than happy with my home built Athlon.

This may be of some help:-http://www.viahardware.com/faq/kt7/faqinstability.html#Why does my KT7 freeze every 5 seconds?

[ 29 January 2002: Message edited by: All That Glitters ]</p>

[ShyTorque]

Tim,

I had similar problems about 4 yrs ago, when I built a new ATX system with an AMD CPU (the system I am still using).

It turned out to be a combination of a poorly performing power supply and a faulty CPU.

I sorted the power supply quite quickly but the faulty CPU nearly drove me mad and cost a lot of money as I had replaced nearly everything else until I found it. I finally exchanged it for a new Pentium and the problem was solved immediately.

I suggest you clock down your CPU using the BIOS or board jumpers to see if it reduces the number of crashes and freezes. If it does, it's a good sign that the CPU is bad.

Good luck - always glad to help an old BFTS and UAS mucker! <img src="wink.gif" border="0">

[allthatglitters]

Also another thought, do you have on the m/board cd, a programme for looking at the temp of the cpu while it's running. With a previous machine, 2 years ago, I had many hours of frustration with a simular lockup, rebooting system, only to find the power supply fan was insufficient to cool the machine, (the top of the case was almost untouchable), only recirculating the warm air, case fan installed, functioned satis.

[scroggs]

Shyto,. .replaced the power supply, and I've tried under-clocking at 900Mhz. No difference! PS: who dat den?. .AtG:. .I mentioned my CPU and system temps in an earlier post - they're fine, as you'll see if you check back. Thanks for the thought, though.

[ORAC]

Another possibility is faulty RAM.

If you have 2 chips pull each in turn and see if the problem goes away. Replacing it would be cheap. 256Mb is about £30.

[rotorfan]

ignore this post, just experimenting........... . <img src="smile.gif" border="0"> . . <img src="wink.gif" border="0"> . .{@). .:&lt;. .:&gt;. .:}. .| <img src="frown.gif" border="0"> . .%). . <img src="redface.gif" border="0"> . . <img src="redface.gif" border="0"> (. . <img src="redface.gif" border="0"> )

[ShyTorque]

Well in that case:

http://www.stopstart.fsnet.co.uk/mica/tossometer.gif

[scroggs]

All that Glitters. .thanks for the link. A lot to go through there; I'll let you know how I get on!

[Blacksheep]

I know it won't make you feel any better but I run an office that is full of nice shiny IBM PCs all under two years old, some brand new. They all have the same symptoms as you describe and our IT Department have been unable to do anything with them. I just mention this so you won't imagine that going out and buying a preloaded fully integrated super-dooper new machine from a reputable manufacturer will be a panacea. I spend half my day soothing enraged staff whose entire morning's output has just been trashed and preventing them throwing their PCs out of the window. Deep inside I'd like to hold the window open for them while they get on with it...

**********************************. .Through difficulties to the cinema

[allthatglitters]

I'm hoping you have read the ammendment to my first post of the 22nd Jan, as I have spoken with my college and confimed the problem m/board was an ABIT KT7A.. .Like everyone else, over the past few days having problems to connect or post, today appears to be greatly improved.

[ 03 February 2002: Message edited by: All That Glitters ]</p>

[scroggs]

As I've referred to this thread on another, I thought I'd update you with the results.

The problem turned out to be a design snag within the Abit KT7 motherboard. Somehow or other, its AGP architecture won't interface properly with some graphics cards. My GeForce2 GTS was only occasionally problemmatic; when I replaced it with a GeForce 4 MX440, the computer wouldn't start at all!

Neither Abit nor anyone else has found a definitive cause, but the only real cure is to go back to a PCI graphics card. Incidentally, the power consumption tests I carried out suggested that 145 watts was sufficient, if marginal - as I originally suspected. The 300w power supply that system now runs on made no difference at all.

My system is now on a Gigabyte K7 Triton 400 mobo, running an Athlon 2000+ XP and the aforementioned GeForce 4 MX440, 512 mb DDR ram and Win XP. It works beautifully!

[Ennie]

Get rid of that shocking m/board and your probably on a winner!!!:D