Calling all Master computer technicians. Never seen this before!

[Spitfire7]

I have been building computers for over four years now and I have never seen this. The day I upgraded my CPU and mobo I have been receiving an error that I have never seen before.

The Problem:

When my computer sits over night off and I go to start it in the morning both HHD light and Power light come on and glow yellow. Then it sits there like that for about 30 seconds with nothing on screen. Also on the mobo the code is 21 which is a power management problem. Then after those 30 seconds I finally get a post beep and the first window begins to load up, but stops with a warning saying... "Warning! You are now is Safemode. You need to reset your CPU and memory configuration in CMOS, hit F1 to continue." Then I hit F1 and everything seems to be ok. This happens every time my computer sits off for more then an hour. If I just have to do a restart or PC is off for half an hour or less, the PC comes on without any problems. Also, when I am in windows and I put the PC in system standby I have problems getting back into windows. It will either just shut itself off, or will come back into windows, but everything is frozen.


Things I have done to fix the issue:

-First of all, I sent back the CPU for a new one. No change.
-I sent back my Mobo for a new one three times. No change.
-I reformatted HHD each time I got new hardware and did a fresh windows XP install, no change.
-I got a new PSU. No change.
-Upgraded RAM to two 2GB sticks totaling 4GB. No change.
-Got new GPU no change.
-I have set all settings to default in BIOS, no change. No overclocking.
-Also installed the latest BIOS drivers as well as all hardware BIOS drivers.

I have also broken down my system to build piece by piece to see if I could come across the problem. First thing I did was just plug in the Mobo power and the CPU power to get my post beep. I did. Then building on top of that, one stick of Ram. I got my post beep. Then GPU. Got my post beep just fine. Then HHD, still got post beep just fine and windows started up normally. Then so on and so on and everything worked fine. It then occurred to me that my issue only happens like I mentioned up above if my computer sits for more than an hour or so. In other words a cold start. Like I said, if I just shut down right now to off and restarted, everything would work fine. I only see the problem when I turn off and then start back up the next morning or go into system standby. Not sure what else to do. Please help.

My Specs:
Quad Core Q9550 3.83GHz
XFX GTX 260 core 216 Black Edition (702 core / 1215 mem)
4GB Corssair XMS2 DDR2 800 (4-4-4-12) 2.1 volts
EVGA 780i Mobo
700w OCZ GamerXStream PSU
320GB Hard drive 16MB cache 7200RPM
Sound Blaster Audigy SE
Windows XP

 

[MRFS]

I infer that you did not replace the RAM, the hard drive, or the Sound Blaster, correct?

In such situations, I would recommend that you strip your machine down to a "bare bones" i.e.:
PSU/MOBO/CPU/HSF/RAM/KVM (keyboard/video/mouse).

Remove the audio card and disconnect all other peripherals.

If you have on-board video, remove your video card and revert to integrated video.

Then, try booting into the BIOS.

If that works, tab over to Hardware and monitor temps, voltages and fan speeds.

Because you have already replaced the PSU, CPU and the motherboard,
the prime suspect is your RAM (not much else that can go wrong with such a "bare bones").

If your bare bones system is stable for 20-30 minutes, then try connecting
peripherals, one at a time.

It may be something really simple like a bad cable too.

This method should help you further isolate your problem.


MRFS

 

[jonpaul37]

oh man, this happened to me! it was a totally different situation tho, i bought a Biostar Motherboard that was used for benchmarking an E8400 and i put in a Pentium 4 630 chip when it arrived, so needless to say, the error message was legit. if i remember correctly, the change i made was a small one in the BIOS, though i'm not sure what the change was, i believe it had to do with quick POST or something like that, possibly even changing the boot partition back and forth. Unfortunately i'm at work right now, but i may have written down the changes i made, (as i sometimes do) if i did, i will get back to you...


EDIT*** i remember that flashing the BIOS did not have any effect either...

 

[Spitfire7]

Thank you guys. I apologize, I actually did change the RAM as well. I will correct that above.

 

[Granite3]

Remove the sound blaster and delete the drivers.

You have to turn off the onboard audio in the bios to use the SB, so that may be causing the problem.

Also make sure you have all 4 cores enabled in the bios. Default is only 2 cores enabled if I remember correctly. I am on the 790i, so same bios options.

 

[DXRick]

Did you set the BIOS correctly for the RAM (2.1V)?

 

[Spitfire7]

Remove the sound blaster and delete the drivers.

You have to turn off the onboard audio in the bios to use the SB, so that may be causing the problem.

Also make sure you have all 4 cores enabled in the bios. Default is only 2 cores enabled if I remember correctly. I am on the 790i, so same bios options.

Already got the onboard HD audio disabled. Also all four cores are correctly enabled. Quick question, you liking the 790i? What advances over the 780i do you have?

 

[Spitfire7]

Did you set the BIOS correctly for the RAM (2.1V)?

Yep I have been sitting with 4-4-4-12 2.1v for quite some time and just went back to default to see if that makes any difference.

 

[Spitfire7]

oh man, this happened to me! it was a totally different situation tho, i bought a Biostar Motherboard that was used for benchmarking an E8400 and i put in a Pentium 4 630 chip when it arrived, so needless to say, the error message was legit. if i remember correctly, the change i made was a small one in the BIOS, though i'm not sure what the change was, i believe it had to do with quick POST or something like that, possibly even changing the boot partition back and forth. Unfortunately i'm at work right now, but i may have written down the changes i made, (as i sometimes do) if i did, i will get back to you...


EDIT*** i remember that flashing the BIOS did not have any effect either...

When you find out what you did, please let me know.

 

[Muster]

If it is a power issue and the BIOS is resetting, I would first check the CMOS battery (reseat it). If you still get the issue after the computer has been off for a while then it may be a damaged motherboard. Perhaps a solder point broke to a necessary capacitor. This may have happened while messing with the cpu and ram.

 

[turk_1000]

I vote hard drive. I've seen some strange problems with cold ones.

 

[Spitfire7]

If it is a power issue and the BIOS is resetting, I would first check the CMOS battery (reseat it). If you still get the issue after the computer has been off for a while then it may be a damaged motherboard. Perhaps a solder point broke to a necessary capacitor. This may have happened while messing with the cpu and ram.

I will check the battery, but I have cleared the CMOS many times by pulling out the battery and firmly placing it back in and that didn't seem to change anything. Like the others have said, I replaced the 780i mobo 3 different times. So I dont think its the mobo.

 

[Spitfire7]

Muster, he said he replaced the mob 3 times dude. It pays to read the whole post.
Have you had your outlet tested by an electrician?

I brought my computer to my parents house for the Holidays and in their outlet the problem still happened.

 

[roadrunner197069]

Bad outlet?

 

[Spitfire7]

I vote hard drive. I've seen some strange problems with cold ones.

I hate to say this, because i dont want everyone to rule everything else out, but I dont have another SATA HHD to test this out with, but I did use an old IDE HHD and the problem was still there. Its a hard thing to test because like I said, I have to keep the PC off for a long time like over night. The HHD is the only thing I haven't replaced since the problem has began, but I did do the SeaTools scan drive test and it past perfected with quick test and long test.

 

[the last resort]

when you tryed the barebones thing mentioned above, was that in the case or out? if it was still in the case, try booting it cold outside the case on a piece of cardboard or something non conductive. because if everything has been replaced, then something that hasn't must be the problem, like the case.

 

[Jim_L9]

Have you tried Memtest86+ and memory that runs at 1.8V as a test?

 

[MattC]

This is a tough one!! You are on the right path, though - you're just not testing the right thing. We just need to devise a few more tests for you ... and more importantly you may need to re-do a few of your tests in a different manner.

First off, redoing the tests: if you didn't the first time, make sure that each of the aforementioned tests you conducted was done in a manner that would actually replicate the problem - there is no point testing something for 30 minutes, then adding another stick of ram (or whatever) and retesting immediately, as one poster suggested - you will have to wait for 30+ minutes to do each test so that the machine can cool down (or do whatever it is that causes the problem after being off for a while.
As another poster said, you should do this piece-by-piece, but for now let's just go to the bare minimum and see if you can replicate it. No sound card/cd drive/, reset the bios settings to the default, and then ... don't do anything with it for a moment. Instead, I'd read the motherboard manual, cover to cover. Just to be safe.

Now, here are some new tests to conduct after you've done that:
(1) In the morning, when you expect to have the problem, do not turn your computer on. Instead, unplug it from the wall and THEN plug it back in and turn it on. Humor me on this one - it's probably not going to do anything odd, but there are components which remain powered up while the computer is off but plugged in.

(2) Again in the morning (next day?) when you expect the problem, before turning your system on try placing a space-heater on it. Don't heat it up too much, but get the case to be a little warm to the touch (even after the heater has been off for a few minutes). The best thing to do would be to heat up the entire room for an hour or so, or keep the space heater on the computer for a good while. The goal is to get the inside of the HDD and everything else up to at least 30 degrees C or so. THEN try booting up.

(3) Unplug all case fans from the system and instead use a completely separate house fan with the side panel off. Unplug the various case-cables that are not needed, such as the hdd activity light, the reset button, the pc speaker, and all usb ports. See if the problem still occurs, but remember that you will need to test it in the morning when the problem normally occurs.

(4) Increase the voltage on your cpu and memory. Try 2.2 for the memory and something slightly higher than normal for your cpu. Remember that each of these tests is separate, so you should have your fans all plugged in, but for this one add an external fan and keep the side panel off. It's not needed but it can't hurt. Try to turn the computer on in the morning after it's been off overnight. Try this test as an absolute last resort though as it's probably not going to tell you anything and you should be careful about over-volting components.

(5) Use a different cable wherever possible. Switch out the SATA data cable and use a HDD SATA Power cable that is on a different rail than the one you have been using previously.

(6) Check your methodology. You say you have been building computers for a while, but we should still consider the possibility that you've just been lucky in the past - make sure you always either (a) wear and anti-static wrist band, and properly ... or (b) are constantly touching a grounded source, such as your power supply metal casing WHILE It is plugged in (or get a metal toaster that has a three-prong adapter ... that should work too). You should not so much as move your toes without touching a grounded source. Don't wear polyester! Well, that last one is a joke, but who knows.

(7) Even though you have tested at another site, we cannot rule out power-outlet related problems. Buy a kill-o-watt tester, at the least, and do some research about this. If you are having bad spikes, you may be damaging components (which you then bring to your parents house ... where they'll remain damaged). This one seems like a long-shot though.

(8) I'm stretching here, but use a different keyboard and mouse and don't plug your system into any network ports for the test.

(9) disable the option to speed up your POST; disable any other features that do not appear absolutely necessary. Keep onboard audio disabled even though your SB card is not in the system.

(10) This one relates to the power issue. Buy a decent UPS. You should have one anyway, you'll want it for either your computer or, if you have notebooks, for your modem/router so that you can use the internet for hours even in a power outage.

Post your results here!

 

[Outlander_04]

It seems that your mb isnt handling the overclock .

Remove the cmos battery from the mb for a couple of minutes . Make it 15 minutes .

Install it again and then use the pc at default fsb and multipliers to see if the problem persists .

 

[Spitfire7]

when you tryed the barebones thing mentioned above, was that in the case or out? if it was still in the case, try booting it cold outside the case on a piece of cardboard or something non conductive. because if everything has been replaced, then something that hasn't must be the problem, like the case.

I actually tired it both ways the first couple of times. At that point, I just sent the mobo back for a new one because the EVGA tech said the board sounded like it was fried.

 

[Spitfire7]

Have you tried Memtest86+ and memory that runs at 1.8V as a test?

I have done the memtest yes but not with a 1.8v memory stick. Im not sure I can do that because I dont have any around.

 

[Spitfire7]

It seems that your mb isnt handling the overclock .

Remove the cmos battery from the mb for a couple of minutes . Make it 15 minutes .

Install it again and then use the pc at default fsb and multipliers to see if the problem persists .

I have my specs listed as overclocked, but currently I have everything on default clock speeds and the problem persists.

 

[MRFS]

I think someone already mentioned this above, in the context of removing
the motherboard for doing certain tests; so please forgive the dupe here:

sometimes, I've seen an HSF backing plate short the motherboard against
the motherboard tray/case.

Just another thing to check!


MRFS

 

[Spitfire7]

Man that is the biggest hassle, but now I am having to look in every nook and cranny. So, I will give it a shot too. Be patient with me lads, I will be doing all of these things and will keep you updated.


p.s. I did leave my CMOS battery out as well unplugged the PSU for about two hours while away and then put it back in and still stuck on error 21 without a post beep until about 30 seconds later.

 

[Spitfire7]

This is a tough one!! You are on the right path, though - you're just not testing the right thing. We just need to devise a few more tests for you ... and more importantly you may need to re-do a few of your tests in a different manner.

Hey Mattc,

So you are never going to believe what just happen, or maybe you will and it was expected. So, before turning on my PC in the morning, I wanted to see if my case was causing anything to be shorting out or causing any problems. So once again I built it on a cardboard box. I triple checked everything to make sure everything was correctly installed. I only had the bare bones mobo, CPU, ram, and GPU. As soon as i turned on the PC, it took a second still with that error code 21 sitting with a black screen and then... poooooooofffffff, smoke from the rear panel near the USB ports and right near the on board sound card came out. I immediately powered off by turing off the PSU and then unplugged. I gave it a second, crossing my fingers and in disbelief and turned it on again. It seemed to start up fine all the way into windows. I even browsed around for a bit. Weird. I shut down and rebuilt the system back into the case. I turned it on and once again error 21 for 30 seconds and then post beep and then started up windows. As I browsed around for a bit and was about to write this letter, the whole system shuts off. Oh shoot, not good. I tried to power on again and I got a small burst of power just enough to spin a fan one cycle and then nothing. I unplugged and did it again. Still just that small burst of power to spin a fan and then dead. So EVGA is cross shipping a new mobo free of charge right away. I dont know whats going on here, but I dont think its the mobo, or maybe it is and I just got three unlucky bad boards. Highly unlikely. Break it to me straight. Is just the board fried (hopefully) or is my PSU fried, or even worse? remember I did get the puff of smoke from the rear panel section. Kind of near the North Bridge top heat sink. I am 85% sure it wasn't the CPU. What do you think is going on here?