Unexpected system shutdown under load: bad caps or power supply issues?

[Lord Zyll]

Greetings! I'm the owner of an aging Asus A7N8X-E Deluxe system. Usually this system is left running 24/7, but some time ago I had turned it off for an extended period, and I encountered all sorts of stability issues trying to bring it back online afterwards.

Principally, this manifested itself with problems getting it to boot, and unexpected shutdowns when it eventually managed to boot. Eventually, when no amount of coaxing could cause it to power on again, I was convinced that the power supply had died (it was nearly to the MTBF spec anyway). I ordered and installed a new power supply (had to use a 520W ATX2.x instead of the old 480W ATX1.x), and it boots once again... but the system still powers itself off under load. It is quite repeatable using memtest86+.

Does this sound like a potential symptom of bad caps, or does it seem possible that I got a dud power supply, or perhaps PS isn't supplying enough voltage on 3.3 and 5v lines? One of the caps near the VRM definitely shows signs of bloating, and I'm willing to take a crack at replacing them instead of throwing the motherboard in the bin, but I wanted to run the failure scenario by you guys for confirmation beforehand.

To try to eliminate the possibility that the power supply was overloaded (due to the difference in voltage distribution between ATX1.x and ATX2.x), I unplugged 4 out of my 5 HDDs - but got the same behavior once I started up memtest86. The spec sheet on the new power supply only claim 130W on the 3.3 and 5v lines (480W on the 12v lines). I wonder if this is enough for an old motherboard designed around the ATX1.x spec?

Looking forward to hearing your thoughts.

-z

[mariushm]

Re: Unexpected system shutdown under load: bad caps or power supply issues?

I'd check the heatsink on the northbridge, make sure it's properly seated, making good contact.

If I remember correctly, they do stay hot usually, so it wouldn't be impossible for the electrolytic capacitors close to the heatsink to go bad.

As you say one of those is bloating, then that's your problem right there, replace all the VRM caps.

The board is still designed around 5v and 12v is used very little, you'd be better to use an older style psu with more amps on 5v.

Check the old power supply, especially the capacitors. I'd suggest you keep using the old power supply if it keeps the system stable after replacing the vrm capacitors and the ones by the AGP slot.

[orbiter]

Re: Unexpected system shutdown under load: bad caps or power supply issues?

As you say the system is rather old and it's been left running for very long periods. It's possible that during this time the CMOS battery has died on you (as they do over time) without you noticing any problems due to the system being awake all the time.

Also as you also mention that these problems started following a shutdown.. I would again suspect the CMOS battery. This little coin cell (located on the mainboard) supplies power to the BIOS chip during system-off periods, this retains necessary data for the system to initially start up.

If the battery has died the system may be booting in a default BIOS configuration, which it may not be happy with, especially if it's had the BIOS set-up in a particular way by the system builder etc. In this case you will need to replace the battery and get a knowledgeable computer friend to help you configure the BIOS to keep the system running at it's best.

The battery is a very cheap thing to try first and is worth a go. If that doesn't resolve the problem then it's time to start looking at further issues. Caps are a definite possibility, especially around the CPU area. If these are bulging then stability can definitely be affected. Replacing them is quite and easy job if you're good with de-soldering & soldering tools.

Just make sure that if you do start trying to replace caps that you get the old caps solder properly hot before removal and installation, as motherboards are more than just one layer thick (some up to 6 - 8 layers these days)

Use a hot iron, around 400*C.. Heat up the solder on each of the two joints for around 2 seconds, then switch between both solder joints with your iron whilst at the same time lightly wiggling the capacitor from side to side, it should start to loosen, and once the solder is molten on both joints it will just pull out.

When installing the new caps.. Observe their polarity (this is often marked on the board & is also marked on the caps themselves)

Hope that helps

[mariushm]

Re: Unexpected system shutdown under load: bad caps or power supply issues?

If the battery was dead, the BIOS would complain about it. But I agree, changing the battery is easy and cheap, it's a plain standard CR2032 battery.

I disagree with the 400C temperature, that's a bit too much and you may burn traces on the back of the board.

The lead free solder melts at about 220c but the board copper planes will cool down the solder iron tip relatively fast, so about 300-350c tip temperature is a reasonable value.
If set to 400c it's quite possible you burn up all the flux in the solder wire before it gets a chance to do its magic on the board.

Basically, you need a somewhat better solder iron than those 10$ things you plug directly into the wall socket, with no way to adjust the temperature. Those don't store enough heat in the tip to be able to properly desolder and solder capacitors on a motherboard.

Heat up the solder points of the capacitors, add some 60/40 63/37 lead solder to those points and gently wiggle the capacitors pulling them out.
Push the new capacitors in, while heating up the solder on the hole tips.

Alternatively, use solder wick to completely remove solder from the holes clearing them (use steel pins/dental picks if you want to clear the holes) but if you don't do this often you may not have access to these things.

PAY ATTENTION to the polarity, because Asus (and Asrock) like to draw the polarity the other way on the boards (the part marked with the filled section or minus is actually plus on capacitors)

[delaware74b]

Re: Unexpected system shutdown under load: bad caps or power supply issues?

To the OP, what brand, series and value of the swollen cap in the VRM section of the motherboard? Example: Nichicon HM 1500uF 6.3 volts. This cap I listed just happens to be a bad batch Nichicon used in the era of your mobo. If it is this cap, you should replace all of them and you can use the same brand and series since the problem has been resolved for 5 years.

BTW, I had one of these boards until it got Antec'd. Smoked the northbridge with 8 volts on the 5vSB. This board does run hot but the one I had was very stable and passed 48 hours continuous memtest with no problems, Barton 2600 core CPU.

[Lord Zyll]

Re: Unexpected system shutdown under load: bad caps or power supply issues?

Thanks for all of the replies. I made time to get out the multimeter and take some voltage measurements directly at the 20-pin motherboard power supply connector (with all system loads connected). I'm not an expert in interpreting these things, but it seems like the +12V was reading a bit high, and the +5V was reading a bit low. The +12V tended to creep higher during the memtest, and the +5V tended to creep lower, but I recorded the observed voltage reading when the system actually failed.

+5V ON-FAIL 4.75v
+12V ON-FAIL 12.68v
+3.3V ON-FAIL 3.34v
+5VSB OFF 5.02v / ON-FAIL 5.01v
PS_ON#: OFF 3.80v / ON-FAIL 0.05v
PWR_OK: ON-FAIL 4.56v

Now I'm not sure if this could be due to a crap power-supply (it is brand new, though that doesn't mean anything), incompatibility between the ATX 2.x vs ATX 1.x spec, or if the voltage skews aren't indicative of anything. The problem is still 100% reproducible, with the system shutting itself off within a few minutes after starting a memtest cycle.

To answer a few of your questions:

I'm pretty sure that the battery is actually dead now, as the BIOS complains about being reset to default settings, but I don't believe that is the root cause of my the power-off under load behavior. I'll replace it if I can get the system back to being relatively stable.

I no longer have the original power supply to attempt a re-capping, though now that I am struggling to confirm that the new ATX 2.x power supply is actually compatible with my old motherboard, I find myself wishing that I had held onto it.

So I guess my question is, now that I have some voltages recorded, should I keep this new power supply, or attempt to return it for something with more +5V power? I would like to get that issue sorted before I turn my attention to the caps on the motherboard.

Thanks for the help!