IME, random resets in recent years have most often been the result of dodgy caps - often in the power conditioning parts of the motherboard. These can also have knock on effects for the PSU creating mutually induced problems.
(If you want a new motherboard to try, drop me a line - I have just acquired a bunch of similar spec MSI boards after doing some upgrades for clients - yours for the cost of postage).
In message , Roger Mills writes
The two tone siren usually indicates something (the CPU most likely) is overheating.
As soon as it restarts can you access the BIOS and check the PC Health status for the CPU temperatures etc?
Given the age of it, I'd suspect the CPU fan is clogged with dust and likely the VRM capacitors on the motherboard are stuffed. Possibly the new PSU is also failing and I'd check *all* the power connections to the motherboard are properly fitted.
[snip]
You don't say what the motherboard is. I had just this with an Asus one. Basically, it decided the main processor was getting too hot (it wasn't) and shut things down. It had a prog supplied with the motherboard software which showed the temperature of various things and fan speeds etc too. And that told of the non existent overheating. I never found out how the temperature was measured (a sensor in the chip or elsewhere?) but tried cleaning and reseating the chip. Which appeared to work for a short time, but back came the fault. Replaced the motherboard but kept the main processor and memory and it's been fine ever since. Of course it could also have been a software fault in the BIOS.
In article , Dave Plowman (News) scribeth thus
Its right underneath the chip usually, sort of springing up a little bit on long "ish" leads...
Course now you may have the change the whole lot depending on the mother board requirements. We've got this problem in that you can't find IDE drives with large capacities there're all SATA...
That looks to be about the right age to have capacitor rot. There was a batch of duff capacitors at about that time. Look for bulging/leaking capacitors, particularly near the processor socket.
Check the mobo manual, what does it say about the two-tone beeps?
Don't let that trick you, solder cracks are progressive. They're caused by a combination of mechanical stress and thermal cycling (which is mechanical stress as well I guess).
Nothing like that on mine.
I've had a close look at the caps, and can't *see* anything untoward - but that doesn't necessarily mean they're ok.
That's a very generous offer - thanks! I'll give it some urgent thought - but may opt for an upgrade anyway, since the whole thing's getting a bit long in the tooth by computer standards.
Anyone know how to remove the heatsink from the processor? It's held on with some sort of clip which fits in a deep groove across the middle - but I can't work out how to undo it, and don't want to use too much brute force and ignorance!
There will (should!) be a little handle/lever at one end of the clip, ease that away from the heatsink and it should release tension in the clip, which will then unhook at each end.
Ignore the slot in the middle, that's just to pass the leaf spring part of the clamp. The clip pulls down under the force of the leaf spring in the middle with rectangular holes in the end pieces engaging on nipples moulded into the processor socket. One end of the clip will be plain, the other will have either a thumb tab to apply pressure to release the clip or a small tab in which you insert a flat bladed screwdriver to do same. Push down and move the tab to release it from the nipple on the base.
Note the way the heatsink goes on for replacement.
Some don't have a tab but just have a recess where the tip of a flat bladed screwdriver can be inserted. You then use a bit of downward and lateral pressure to pop the clip of the lugs on the side of the socket.
In article , Dave Plowman (News) scribeth thus
Must be doing it another way or prolly don't actually do it. Three boards here have got them..
I've got two desktops here - and of course had the original faulty motherboard. All Asus with Athlon processors. None of them have any obvious sensor under the processor or anywhere near it. All will read out the various microprocessor temperatures and fan speeds etc in the BIOS or by using PC Probe on the desktop. And the faulty board would sometimes shut down just after startup before the processor had a chance to get even remotely warm. With the BIOS showing it overheating. I assumed therefore the sensor was in the processor. But the same processor in a new motherboard has been fine. So I enquired no further. ;-)
Chip temperature is determined by measuring the voltage across an on- chip reference diode.
MBQ
Right - that makes sense. But didn't explain the fault on my motherboard since the same chip is working fine on the new one. I did re-flow the solder to the socket with no improvement. So it could have been perhaps a software thingie - or broken track.
Ah, that's the one - simple when you know how! Thanks.
Not sure it's done me much good though. I removed the heatsink, re-distributed the conductive paste, replaced the heatsink and switched on. [No monitor connected, so I can't see exactly what it's doing - but wanted to see how long it would run for] After about 4 minutes - by which time it would barely have finished loading all the start-up processes - it switched itself off with the usual very quiet two-tone siren (which I've discoved is a little speaker attached to the chassis connected by a thin wire to the motherboard).
I felt all round the heatsink as soon as it happened, and it was very cool. So if it *is* an over-temperature alarm of some sort, the sensor is pretty certainly lying!
Does your machine give the temperatures etc in the BIOS setup pages? If so go to these at startup - on mine you press 'delete' just after switch on and see what it says. It will update the temp as it increases. I also have PC Probe which does the same on the desktop. But that may be an Asus specific prog. Although I'd expect other makers to have similar.
Yes, according the booklet which came with the motherboard, you can set a critical CPU temperature, but I don't know whether it's enabled or not. There's also a PC Health Check page which displays temperatures, voltages, fan speeds, etc. The system is in bits at the moment, but I may put it back together and have a look at that later - assuming it will run for long enough to do so!
There's a software program called PC Alert4 which does the same - and sounds a loud alarm if any of the monitored parameters go outside their set limits. I've had that on the screen all the time during the recent problems - and it's not shown anything unusual, or produced any alarms - the PC just stops dead with no warning! The CPU temperatures displayed are actually quite a low lower than they used to be before I fitted a new CPU fan a year or so ago - so it seems unlikely that the problems are due to having a lower temperature set in the BIOS - because this would have caused problems before the new fan was fitted.
In article , Dave Plowman (News) scribeth thus
The ones I looked at are indeed quite old, thats why there're doing not a lot at the moment;!. The on chop diode sensor sounds a very good way of doing that job...
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.