Salvaging a Failed Computer

English Open Chat
1

Last weekend one of our computer’s died. This was my 32-bit AMD Athlon-2800 w/2GB RAM (Asus A7N8X Deluxe motherboard) w/ PCI nVidia GeForce 8400GS graphics [age ~7 years] running 32-bit openSUSE-11.3 KDE-4.4.4 (and also a version of SLED-11.1 SP1 that I boot to occasionally). This is my PC for vdpau testing. I also had just installed in another partition openSUSE-11.4 RC1 on this PC.

I was very sick last weekend, so I simply left the dead PC, until I was well enough to look at it. Today (one week later) I feel better (albeit I still have a wicked cough).

**Symptom **

The symptom of ‘death’ for the PC was to push the power switch, and no noticeable response. No lights, and no detectable noise. I suspected Power Supply or Motherboard.

So I removed the PC from our hardware KVM switch and opened it up. There were no noticeable power scorch marks. A small amount of dust (of course) but nothing deadly there.

I plugged in the power cable again, switched the master power switch on the back of the PC to ON, and pressed the power ON. Nothing. BUT when I switched the master power switch on the back to OFF, I heard the distinct (but very brief) sound of a hard drive click. I repeated this (ON/OFF) and heard the same. Clearly some sort of power reached one of the hard drives.

With the PC electric cable plugged in but the PC OFF (with the master power switch on the back ON) I noted that if I plugged a wired Ethernet cable into one of the mother board’s Ethernet jacks the motherboard light by that jack would light up (as expected in a healthy motherboard) BUT if I plugged the same wired Ethernet cable into the mother board’s other Ethernet jack, the light by that 2nd jack did NOT light up.

My suspicion now is the power supply works, but the motherboard is dead. I can’t prove this (yet).

This PC is old. I think I’ll just throw it out, cannibalizing the parts best I can.

keep Graphic card

It has a nVidia 8400GS PCI card which is a slow card (using the PCI bus) that has superb High Definition Video playback capabilities using Pure Video (in winXP) or VDPAU (in Linux). I’ll keep that card although I don’t have a PC to use it in (yet). The power supply in my old athlon-1100 is not strong enough to power this card.

**keep the RAM **

The expired athlon-2800 has 2GB of RAM in 3 DDRs: Two 512MB DDR-400’s and one 1GB DDR-400. I ‘might’ be able to use those. My even older athlon-1100 PC (w/MSI KT3 Ultra motherboard) can accept DDR200, DDR266 and DDR333 and it has only 1GB of RAM (with 2x256MB and 1x512MB DDR333s). I might remove all those DDR333 and try all the DDR400s in that old athlon-1100 MSI KT3 Ultra motherboard. My research suggests there is a possibility a DDR400 DIMM will work in a DDR333 slot, albeit only at DDR333 speeds. Still that would be ok, as 2GB of RAM is superior to 1GB of RAM.

**other equipment **

The dead PC has 2 hard drives: an 80GB and a 300GB. Both worth keeping. It has 1xdouble-layer DVD read/writer (worth keeping) and 1 x DVD reader (worth keeping). There is also an old internal 2.5" floppy drive that I might keep or throw out. No one uses floppies any more. And of course the nominal assortment of cables and screws and mounting brackets that I can salvage.

There is a dusty old CPU fan. I don’t know if I will clean up and keep that. For the few cents its worth its possibly not worth the effort to clean up.

**salvage/test the power supply ? **

Finally there is the power supply. A Chieftec HPC-360-302. Its old. Its dusty. BUT if it still works, it is a superior power supply over the make of power supply in my old athlon-1100. It has me thinking I could try that power supply in my athlon-1100 (although moving all the cables around could be a major time waster and source of irritation). Putting this power supply in my athlon-1100 PC ‘might’ prove for certain if the problem with the ‘athlon-2800’ was the mother board or power supply. The risk is dependent on the athlon-2800 PC failure, if it was the power supply then putting a failed part in the athlon-1100 could destroy the motherboard on the athlon-1100. I’m not sure I want to go down that investigative road. BUT if it works, then I could possibly put the nVidia 8400GS in the athlon-1100’s MSI KT3 Ultra motherboard give this old athlon-1100 high definition video playback capability - although there is a downside to that - that 8400GS PCI card is SLOW for all else, and the current graphic card in this athlon-1100 (w/MSI KT3 Ultra) is an AGP FX5200 nVidia card which is actually faster in all other respects than the 8400GS PCI card (due to the superiority of the AGP bus). Note I am referring to a PCI and NOT a PCI-e bus.

… I’m not sure what I’ll do wrt the power supply.

2

That’s a terrible loss oldcpu. Here in Hellas many computers die from high voltage.

3

Sorry to hear that your athlon 2800 died oldcpu. A few weeks a go my testing computer (2GB RAM, Pentium 4 3.2GHz, NVIDIA 8400GS PCI-E) died but I revived it with another power supply from a older computer I had lying around. Trouble is this replacement power supply I had put in does not have any SATA power ports so I could not plug in the SATA drive so I had to go hunting around the house for an old IDE hard drive which I will continue using until I buy a Molex to SATA power cable. Age of this computer is >7 years old.

4

Sorry to read you could not use the SATA power ports, but its good to read the replacement power supply had sufficient power to run your nVidia 8400GS PCI-E card. Its also good to read your testing computer has a PCI-e bus. My athlon-2800’s A7N8X Deluxe motherboard was simply too old to have the PCI-e bus.

old PCs are great for testing, but when they fail, one needs to ponder carefully the question as to whether it is worth throwing money at them. I still have not decided to salvage this old dead PC of mine for parts, or try to repair, as I am not 100% certain its the motherboard that failed (it could be the power supply - albeit the evidence appears to point to the motherboard). Still, this PC is so old its likely not worth the cost/effort to try and revive it if I don’t have the spare parts handy (which I don’t). I know a PC that is this age is hard to even give away as most people don’t want a PC that old.

5

Salvage is probably the best approach in your situation Lee.
Now you have an excuse to build a nice new one.
I always find, I’m reluctant to throw load tack in to a new build though.

Fun time ahead though :smiley:

6

On 02/19/2011 11:36 AM, oldcpu wrote:
> SALVAGE/TEST THE POWER SUPPLY ?
>
> Finally there is the power supply.

of course, i’ve been wrong before but i think what you have is a bad
power supply (anyway, statistically it is more likely to fail than the
motherboard–unless you see swollen, distorted or leaking capacitors
<http://en.wikipedia.org/wiki/Capacitor_plague>)…it provides
several different voltage/amperage levels as well as detects system
the draw…

for example, in modern computers the main “on” switch only momentarily
puts a small load on the power supply which “detects” the load and
then supplies power to certain of the leads to kickoff the bootstrap…

wake-on-lan can work without a finger present because it too can
tickle that load…

anyway, if you have a known good power supply of sufficient capability
you can sub in, i’d suspect your old friend will serve some more…

–
DenverD
CAVEAT: http://is.gd/bpoMD
[NNTP posted w/openSUSE 11.3, KDE4.5.5, Thunderbird3.0.11, nVidia
173.14.28 3D, Athlon 64 3000+]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11

7

If you have a multimeter, test the RTC cell, will only take you a moment. Often a computer cannot cold boot because the RTC is not running.

Next thing is to test the power supply by borrowing one from another computer. PSUs are under high stress.

8

The Real Time Clock in this old (now dead) computer has been acting up for some time, giving the wrong time. Not sure what the “RTC cell” is but I think you are refering to the Real time Clock.

Mind you I’m getting NO power indication on this computer. Push the power button and there is nothing. However when the main power switch on the back is switched OFF, one can hear a distinct ‘click’ from one of the internal hard drives (even thou there is no sound when power applied).

I don’t have one of these lying around. I supose I could pull the power supply from our athlon-1100 and put it in the athlon-2800. All the re-wiring would be a major pain thou. I need to decide if I wish to spend the time. :slight_smile:

Thanks for the suggestions thou.

9

I was actually very suprised that the NVIDIA card worked with this very old replacement PSU.

10

It’s the (usually) CR2032 lithium button cell in a holder.

11

My wife thinks we have a spare battery, so we may try to replace that and see if it makes any difference. But I confess it suprises me that a dead battery would cause a completely dead PC indication. We will also try your suggestion to measure the voltage in that area, as we plan to borrow a volt-meter off of a friend.

I took another close look at the motherboard / power supply with my wife, and it turns out two eyes are better than one.

When we plug in the PC and with the master power switch on the back of the PC ON, there is a light that comes on the motherboard (its called “the onboard LED (Green)” and according to the motherboard manual, the “Green LED indciates the system is ON, in sleep mode or in soft-off mode, not powered off”). My wife noticed that when I push the front “ON” switch on the PC, while nothing happens, if we then switch OFF the master ON switch on the back of the PC, the CPU fan suddenly spins for a few seconds in addition to the hard drive noise I heard.

Here is a sketch (from manual) of motherboard, with arrow pointing to LED that is ON :
http://thumbnails3.imagebam.com/12020/71c336120194925.jpg](http://www.imagebam.com/image/71c336120194925)
[click on image for larger view]

and an image of the motherobard (from manual) where item-19 is the LED that is lit :
http://thumbnails40.imagebam.com/12020/8a43e9120194928.jpg](ImageBam)
[click on image for larger view]

My wife’s assessment matches that of DenverD’s, that the powersupply is at fault, where she thinks the motherboard is only delivering part power to part of, but not all of the motherboard. Me? I’m not so sure. Hopefully when we get our hands on the volt meter we can figure out the best place to measure the voltage. Clearly part of the motherboard is getting power, so its still not clear to me if this is power supply or motherboard at fault.

12

We removed the old RTC cell (battery) and checked it with our friend’s volt meter. 0 volts. We then pulled our our replacement battery of same type and measured it (almost 3 volts) and put it in the computer. The computer is still dead (same symptoms) but that is at least one thing we can (1) confirm is now fixed and (2) note not a direct cause of PC failure.

We then connected the volt meter to one of the CD/DVD drives power input. 0 volts even though power is being applied. But what was of interest was as soon as we switched OFF the master power switch on the back of the PC (ie on the power supply) there was a brief few second power surge that was immediately measured on the volt meter. This does appear more and more to be abnormal power supply behaviour.

We have run out of PC tinkering time for the weekend, but I may purchase a replacement 430 watt (or so) power supply if I can find one that is inexpensive, and replace this old Chieftek power supply with a replacement. Our local PC show is selling a 430 Watt beQuiet power supply for 57 euros. Thats a bit much for an old PC, and I have not yet decided to do a purchase such as that, but it may be the course of action to follow (assuming I want to throw money at a 7 year old PC).

13

On 02/19/2011 02:06 PM, oldcpu wrote:
>
> Hopefully when we get our hands on the volt meter we can figure out the
> best place to measure the voltage. Clearly part of the motherboard is
> getting power, so its still not clear to me if this is power supply or
> motherboard at fault.

unless you know how, even with a volt/ohm meter it is just a shot in
the dark to try to measure voltage output at the ends of any of the
many wires/cables coming out of the power supply unit (PSU)…

because, as mentioned part of the thing the PSU does is meter out
power according to need (which it senses)…that means if your volt
meter can not draw power, it won’t be sent power and you will
therefore not learn much with a volt meter unless you do it exactly
correctly, see
http://pcsupport.about.com/od/toolsofthetrade/ht/power-supply-test-multimeter.htm

much easier, safer and better is to use a PSU tester
<http://about.pricegrabber.com/search_attrib.php/topcat_id=1/form_keyword=power%20supply%20tester/st=query>

and, if you do opt for a new power supply i highly recommend you buy
one about 50% more powerful than the one you are replacing, and i find
the ‘quiet’ versions well worth the extra few euro (in my book)…

–
DenverD
CAVEAT: http://is.gd/bpoMD
[NNTP posted w/openSUSE 11.3, KDE4.5.5, Thunderbird3.0.11, nVidia
173.14.28 3D, Athlon 64 3000+]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11

14

The probabilities favour the PSU being the problem. Since ATX onwards, the PSU is always powered if the master switch is on, supplies a standby current (hence the lit LED), and there is a soft on switch on the front panel which goes to one of the PSU pins. I’d say that the PSU has aged and is unable to fire up the main power section.

Usually I check by borrowing a PSU from another machine. It’s not too hard, you only have to remove the 4 screws holding it to the case and disconnect a few power connectors, all of which are keyed and impossible to plug in the wrong way. There may be some subtleties whether you have a 20 pin or 24 pin connector though.

15

ken yap wrote:

> Usually I check by borrowing a PSU from another machine. It’s not too
> hard, you only have to remove the 4 screws holding it to the case and
> disconnect a few power connectors, all of which are keyed and impossible
> to plug in the wrong way. There may be some subtleties whether you have
> a 20 pin or 24 pin connector though.

You have led a sheltered life, it appears. I can even recall a flight line
mechanic using a pocket knife to remove the “ears” on a power cart plug to
make it fit - backwards. Pilot hit the battery switch and that cart did a
flip - amazing what a generator rated to source 24 volts/2200 amps will do
when reverse connected to batteries with a rated charge rate of 1000 amps
each…

I hate to switch PSUs myself - my fingers are too big and stiff anymore to
reach some of those neat places where they tuck conectors.

–
Will Honea

16

Well PCs are consumer electronics and it’s generally bad business to kill the customers. :wink:

That’s why I switched to software. With software you can put the smoke back into the “chips”. :slight_smile:

The last time my PSU died I swapped in a PSU from another machine. I bought a replacement a week later, but it was months before I got around to putting the replacement in. It would have been even easier to do the swap if I hadn’t wanted to avoid disconnecting all the wires from the back panel and slide the tower out of its cart. :wink:

17

On 02/20/2011 03:40 AM, Will Honea wrote:

> that cart did a flip

heh…heard that story before! torque is an amazing thing!

–
DenverD
CAVEAT: http://is.gd/bpoMD
[NNTP posted w/openSUSE 11.3, KDE4.5.5, Thunderbird3.0.11, nVidia
173.14.28 3D, Athlon 64 3000+]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11

18

I may simply roll the dice and buy a new power supply unit (PSU) to avoid the risk of injury. … let explain …

The PC I have with a PSU that I could try is my older athlon-1100. We have the PC neatly located under the desk connected to a hardware KVM. To remove it requires crawling under the desk, disconnecting all the connectors, and then pulling the athlon-1100 PC out. But I’m not young any more, and crawling under the desk is murder on my back (which is still sore from pulling out the dead athlon-2800). Then once the athlon-1100 PC is out and open, disconnecting the power connectors and removing all my wife’s carefully made tie wraps (as we put this old athlon-1100 together ourselves). Then when the test is over, doing all the above again in reverse and crawling back under the desk. Plus all the cables under the desk then need to be neatly re-arranged again which require one twist and move about in a confined space.

That reads like a lot of pain …

The athlon-1100 was the last PC my wife and I put together, and the main reason it was the last is both of us found the connector part of the assembly (trying to wrap the wires so they were neatly located inside the PC while still provide superb access for one’s hands to do various upgrades … etc … ) the most annoying part of the assembly. We told ourselves afterward, never again. We would rather pay someone 50 euros to assemble a PC for us than do that again. … Well when I look at the price of a new PSU and the less expensive ones cost around 50 euros.

Now if I could only find that legendary fountain of youth, I might be able to reconsider my approach to this. :slight_smile:

19

DenverD wrote:

> On 02/20/2011 03:40 AM, Will Honea wrote:
>
>> that cart did a flip
>
> heh…heard that story before! torque is an amazing thing!

At one time, the photos and incident report were part of the UPT training
program. I was especially impressed after I had a regulator stick in a
generator (when lead acid batteries were used before NiCads became common)
and the resulting battery explosion blew a good chunk of the nose off the
a/c.

I wish I had a nickle for every minute I’ve spent watching some tech try and
connect D-style connectors upside down :wink:

–
Will Honea

20

On 02/20/2011 09:36 AM, oldcpu wrote:
>
> (trying to wrap the wires so they were neatly located
> inside the PC while still provide superb access for one’s hands to do
> various upgrades … etc … )

jeeeeezeee…i bet you two are the ‘type’ that even puts the case
covers back on!! [and has a place where the vacuum cleaner is always
stored] :slight_smile:

i lost my vacuum cleaner once…for a couple of months…
then, one day i tripped on the cord, and found it… :wink:

–
DenverD
CAVEAT: http://is.gd/bpoMD
[NNTP posted w/openSUSE 11.3, KDE4.5.5, Thunderbird3.0.11, nVidia
173.14.28 3D, Athlon 64 3000+]
“It is far easier to read, understand and follow the instructions than
to undo the problems caused by not.” DD 23 Jan 11