Problem with fresh PC system build - unusual instability

[PartingShot]

Hey all, I have a problem with a brand new PC system I built last week. As this is a very helpful group, I figured I'd see what the community thinks.

Last week, I bought fresh parts for a PC build. Here are the components:

CPU: AMD Athlon II X2 250
MB: Asus M4A88TD-V EVO/USB3
MEM: Crucial DDR3-10600 2x1GB
PSU: Cooler Master GX Series 750W ATX Power Supply

With the above parts, I added some Seagate SATA 300 HDDs (1 for OS + 2 for storage) and a Pioneer A109 DVD burner. The PC case and accompanying case fans are older (used) as well.

I integrated all these parts and was able to power up, set up BIOS, and install WinXP Pro 32-bit with no issues whatsoever. I did do something a bit "exotic" during install in that I slipstreamed the AMD AHCI driver into the WinXP Pro setup CD (also slipstreamed SP2 in), and completed the WinXP installation so that all SATA drives are running under WinXP in AHCI mode, not IDE/ATAPI.

Next, I installed the MB drivers (for LAN, onboard video, onboard audio, etc.), applied SP3 and all critical updates from MS, and installed some programs.

All throughout installation, updating, and configuring to suite taste/needs there were no issues. I'd say the system was on for around 6 or 7 hours, and was stable and solid.

The system ran fine for the first couple days, then an intermittant failure began to show up.

In the first noted instance of failure, the PC did not boot. Turning on the power button activated all the fans (CPU, PSU, and 4 case fans), but nothing was being sent to the display. BTW, for video, I'm using the onboard ATI graphics, connected thru VGA to an old Samsung SyncMaster 13" CRT monitor. It was clear that the graphics hardware was outputting no signal on the VGA (the Samsung monitor has an LED that blinks when receiving no VGA signal). Additionally, the system would not respond to keypresses, and holding the power button for >4sec failed to shut it off (I held the button down for around 20 seconds, and nothing). However, pressing "reset" worked to restart the system.

I read that when the PSU will not shut down while the power button is pressed and held, it's possible that a short exists on the MB card or between the MB card and the PC case. So I took the PC apart, extracted the MB, blew about half a can of compressed air all over it (in case of a short due to a metal shard or loose solder ball). Then I put insulators on all the mounting standoffs on the case (used to screw down the MB into the case) to electrically isolate the MB and the PC case.

I reassembled the PC, and it seemed to work fine... for 1 day.

The next instance of instability occured after getting into Windows. While browsing the web (IE8 with all the latest patches), the screen went black. No video being sent to the Samsung monitor, just like before. CTRL+ALT+DEL doesn't work. And, once again, pressing and holding the power button did not work. But again, pressing "reset" does restart the system.

I went to the retailer that sold me the MB, but they don't have any other units in stock at the moment, so an exchange is not possible yet. I just want to see if the community here thinks that a faulty MB is the problem. It doesn't seem to be a Windows-related issue. I may try downloading and installing WhoCrashed to see if any additional information can be obtained. The fact that the problem is intermittant makes it difficult to troubleshoot, as you can imagine

Any response would be very much appreciated. Thanks in advance.

[CobraPilot]

I've had a suspiciously similar problem, but the only commonality is that we're both using AMD CPUs. Mine, however, is a much older Socket A single-core, running at 1.8 ghz.

Every so often, after the video goes into "Energy Star" mode, it simply will not come back on, and I have to do a cold turn-off-the-power-strip reboot, which usually (but not always) results in a good reboot. After one of these perplexing crashes, with the power off, I usually open the case, check for firm seating of all components and connectors, etc. Then I turn the PC on and watch all the cooling fans to see if heat is a problem. So far, all fans spin okay.

I've removed the heat sink, reapplied thermal paste, and reinstalled the fan and CPU.

While I'm having longer periods between crashes than you are, I suspect I'll continue to have them unless I can isolate the problem.

I'm hoping my "I feel your pain" response, while not helping you, will serve as a message "bump" to keep this question active.

[redwudz]

You don't want to put insulators under the MB mount screws. That's part of the ground plane for the motherboard. I would unplug and replug all cables, cards and RAM and see if you just have a bad connection. Then I would probably try a different power supply if you have a spare. You could also substitute the RAM, though that doesn't sound like the problem. If you still have the same problems, then it may be the motherboard.

EDIT: The front panel power switch shorts to ground to turn on the power supply. There is a transistor switch on the MB that does the actual switching, it shorts pin 15 on the PS to ground to turn it on. Software turns that same switch off when shutting down. Or the front panel switch sends the signal to the MB for shutdown.

[lacywest]

How about using a real video card instead of the one built in ... to see if the error still occurs.

And ... re-install ... Win XP ... without the slip stream involved ... and see if the error still occurs.

I recently had problems with a PC I was building last night [for a customer] ... I kept losing video to the monitor.

I thought maybe the Power Supply was just not stookum enuff for the video card ... ATI X1300 AGP video card.

So I opened up a box and put a new Power Supply in ... nope same errors occured.

So ... I thought maybe the video card was overheating ... so I added a different heatsink / fan to it ... I did get nice temperatures after doing this but the error still kept happening. So I changed the video card. All is well now. I swapped a ATI X1300 ... for a ATI X800 ... I really didn't want to do that but hey !! ... its not doing anymore blue screens anymore.

I use the games ... Crysis or Crysis WarHead for testing a PC. If the game plays well and doesn't stutter or freeze up ... the parts I'm using are good enuff.

A few months ago ... this customer I'm building this PC for ... he brought me this Pentium 3 PC to try and get it to go faster. The MB will only support 512 megs of ram. Its not even SD DDR ram but the stuff with 2 slots on the sticks. UGH !!

So ... I did what I could and 2 weeks ago ... he brought it back saying it is running slow ... he he he ... I told him it was slow to begin with ... " well see what you can do ... and clean it up ... so it can run faster ".

Well ... the tower is a good one ... no USB ports in front or audio jacks in front but the tower itself is the kind you can build on ... so ... he now has a AMD 2.3GHZ dual core - MSI K9MM-V system with 2 Gigs of DDR2 Ram ... now to see if he is so cheap ... that he wont want to pay me ... $80 bucks for speeding it up. And this guy is the manager at a Shell Station convenience store on Main Street ... in the center of the town I live in. As for his nationality ... Do you watch ... The Simpsons ?? Need I say anymore ??

How about skipping all those extra HDs and just have one inside the tower ... that is all I have ... one simple 80GB ... dual boot ... WINXP and Win 7. But I have six or more sometimes 500GB USB 2.0 hard drives connected. My games are installed on the one I have designated as the D Drive.

Plays ... Crysis and Crysis Warhead ... Aliens Vs Predator ... Legendary ... Singularity ... Mass Effect 2 ... F.E.A.R. 2 ... Left 4 Dead 2 ... BioShock 2 ... just fine.

I'm looking forward to ... Crysis 2 ... coming out in next few months ... the previews look great

[TimA-C]

Have just had similar issues with a similar spec PC (not quite such a nice ASUS mobo though) and 64-bit Windows 7. I suspected it was a memory problem but didn't have access to any other DDR3 RAM to substitute. After eliminating everything else apart from the RAM and the motherboard I got the RAM replaced and, touch wood, that seems to have sorted it.

If the on-board graphics is using system RAM then your's could be a RAM issue as well. Have you tried using just one stick of RAM? That was the clincher for me. System would work fine with one stick but would nearly always BSOD during cold boot or within a minute or two of browsing in IE8 after a cold boot (well, a very cold boot as the problem would only happen after the PC had been off for several hours.)

Other suggestions I found offered to people with similar problems and similar spec PCs on the ASUS and Corsair forums were;-

- Manually enter the memory's timing in the BIOS in case the motherboard had wrongly identified the RAM.

- Up the voltage to the memory from 1.5v to 1.6v (NOTE: any over-clocking or over-voltaging(?) is done entirely at your own risk and would probably invalidate any warranties etc.)

- Power supply not supplying enough current to the memory until it had warmed up a bit.


It's not something really stupid like the on-board graphics card suddenly deciding that your 13" CRT can run at some ludicrous resolution or refresh rate and the monitor just going "Nope!" and going to sleep?

[Bjs]

So many stories already ?

Q: Turning on the power button activated all the fans (CPU, PSU, and 4 case fans), but nothing was being sent to the display

The case is OLD and so are those buttons ... it is cheaper to grab a new case and try the setup again... you may be suffering intermittent startup button contact failure.

As for Q is generally a sign an issue exists with the PSU ... Ive seen this a few times before ... but first try a new case before swapping the PSU.

[nic2k4]

More than likely it's a bad motherboard, if it was RAM it would still turn on and beep like crazy. It should do the same when no video hardware is present, but not sure if that applies to on-board video. It sounds like a case of bad solder joints, usually with a BGA chip (chip has no visible pins); BGA is total crap and unreliable in the long run, it's the industry's dirty little secret.

Try turning the PC on it's side with the board on the bottom, if the problem goes away that's what it is. Might not do anything, but that doesn't mean the problem is something else. A little extra pressure might be needed, like adding a small stack of quarters on top of the chipset (careful you don't make it fall with the power on).

I had 2 K7S5A pro go that way after years of reliable service. One would only work on it's side, so I bought a desktop case, put it in and it never turned on again.

Don't wait for your supplier to have more boards to return it, most stores have a limited time to exchange DOA parts. After that you only have the manufacturer's warranty i.e. you ship it to Asus they test it and IF they find a defect they fix it. Takes a long time and you might have to do it more than once with an intermittent problem.

BTW, ctrl-alt-del doesn't reboot a PC anymore (with NT), it only brings up task manager. Ctrl-alt-del, alt, u, u or r will shutdown or reboot the PC. With Windows running, pushing the power button once will shutdown the PC. Reset always works, it's a hardwired switch to the chipset and CPU reset pin among other things.

[Nelson37]

Fascinating.

The power supply runs correctly, at least part of the time, the system does boot and run regularly. Power supply failures are rarely intermittant or changeable, while there are exceptions it is most common that when they go, they just don't work anymore. If you have an extra one laying around, this is always a quick and easy diagnostic check.

No beep codes on failure or lockup, so the error is low-level. A dead short is always a possibility, these can be intermittant for a variety of causes.

Since the expensive board is under warranty and replaceable, I would do so at the store for the closest immediately available substitute, before the in-store-replacement deadline expires.

The power switch mechanically is fine, otherwise no fans would come on. Reset works, and power does not. What component controls their function? Mobo.

Next diagnostic step would be remove the board from the case, set it on a box or pad, one stick or ram, boot to the bios and let it sit. Flaky board will reboot here, eventually. Good board will stay up indefinitely. Different power supply and RAM, both KNOWN GOOD, will nail it.

[ranchhand]

Don't forget that lately, say the last year or so, there have been a increasing number of bad capacitors being reported on mainboards as well as LCD Tvs; when a cap is in the process of failing it can begin giving problems before the electrolyte actually starts to bulge or crack the container. Someone would never know that was the problem and spend countless hours trying different stuff with no success. My suggestion is to return the mobo if it is still under warranty, while you still can. IF it starts doing it with the new board at least you will have eliminated the board as the problem.

[redwudz]

All good points. I had forgotten that the on board video uses system RAM, so worthwhile checking the RAM.

Most times bad MB caps show after quite a bit of use of the MB, but anything is possible.

Replacing the case because of a bad power switch seems a bit extreme. You can trade the reset switch terminals for the power switch terminals at the MB front panel connection as a easy test of the power switch. They both operate the same way.

AHCI isn't native to XP, but if set up right on XP should work. Easy enough to disable it in BIOS most times to eliminate it as a possible cause.

But this does sound more like a hardware problem than software. I'd probably exchange the MB and if the problem persists, you've at least narrowed down the causes.

[TBoneit]

My guess is that the power supply has bad capacitors. That will make it flakey.

I just replaced four bad capacitors in a 650 watt power supply that were bulging today. 3 x 1000uf 10V and one 2200uf 16volt. Most power supplies get crud brands used in them, Antec included.

These days when I am prepping a used computer for resale I check the motherboard and power supply for bad capacitors. We give a one year warranty so I'm being proactive that way. You'd be amazed how many I change that way.

[TimA-C]

If you want to take Windows and it's drivers out of the equation then you could download a linux live CD and boot from that (you can run off the CD/DVD without installing onto a hard drive). I'd look at either Linux Mint or Ubuntu. If the PC is stable with Linux then it's possible that your slipstreamed Windows disk is the cause. It still sounds like a hardware problem to me but this might help to confirm it without having to re-install Windows again.

[Constant Gardener]

First thing I'd do would be try running with just one stick of memory. If no change, try the other stick. I always suspect memory when weird things happen, particularly when the system is new; also, it's an easy test.

Boot up with Memtest and let it run for a few hours. Anything other than zero errors is unacceptable - however it may not be the memory at fault. One time I encountered a poorly-mounted CPU heat sink that looked like bad memory. Fortunately I fixed the heat sink before I considered replacing the memory.

[PartingShot]

Hi everyone. Thanks a bunch to all who have shared their ideas and experiences. Some new information on the problem I'm having has come to light.

For one thing, the problem I am experiencing seems to only take place after a "very cold" boot, meaning that if it happens, it only happens when the system is started up after being off for several hours. And when the problem actually happens, it always happens during the first several minutes of system uptime. After several minutes of system uptime have elapsed, the problem never appears. So if the system has been "warmed up," the problem is eliminated.

For another thing, the problem has been observed with this system powered by two differnt PSUs. So I think the PSU can be ruled out as the source of the problem.

When I last went to the retailer that sold me the mobo, their customer service department flagged my purchase as "exchangeable after 30 days" while I'm waiting on their new shipment for the mobo. (They are currently out of stock.) I definitely want to swap out the mobo while I'm able to do so through the retailer.
I'll continue to monitor this thread, and will post updates. Thanks again to everyone for all your helpful contributions

[TimA-C]

This is sounding more and more like the problem I had. If the mobo is still in your system, does the problem go away with only one stick of RAM in the system?

[PartingShot]

This is sounding more and more like the problem I had. If the mobo is still in your system, does the problem go away with only one stick of RAM in the system?

Actually, the problem seems to have vanished altogether. I've written an update, which follows below:

Update...

Since my last post (on 9/4), I have introduced a stop-gap measure to protect the WinXP Pro OS from potential damage or corruption due to having to press "reset" regularly.

I put in a boot CD-ROM, which runs the Acronis True Image backup program (2009 version). This boot CD is Linux based. When the PC is turned on, it first boots from the CD drive. The Acronis program launches, and the first thing it does is present the user with a welcome screen which gives him/her 3 choices:

1. Start the Acronis program
2. Continue to Windows
3. Power off the PC

This screen never times out, so the user has to make a choice. I figure this is a good state to be in if the PC goes into its bad state (on cold start up) requiring a "reset" button press, rather than being in Windows and having to press reset (which I read can lead to OS damage or corruption).

Since I put in that provision, however, the problem has never shown up since. I turn off the machine at night before bed, and start up again the following morning (giving about 8 hours of off time). For the last 4 days, the problem has never occured, even on a cold startup.

So today, I checked my retailer's stock on the mobo, and they show (online) that 4 new units are available. But if the problem is no longer there, should I go through the trouble of exchanging it at all? What would you fine folks do in my situation? Thanks again for your continued input.

[ranchhand]

Parting Shot said:

For one thing, the problem I am experiencing seems to only take place after a "very cold" boot, meaning that if it happens, it only happens when the system is started up after being off for several hours. And when the problem actually happens, it always happens during the first several minutes of system uptime. After several minutes of system uptime have elapsed, the problem never appears. So if the system has been "warmed up," the problem is eliminated.

That is a common characteristic of caps in the process of failing. I have seen the exact same symptoms in so many LCD TVs and computers I can't begin to count. All had bad caps, at least what I can remember. In fact, my Samsung 40" LCD (1.5 yrs old) is doing the exact same thing presently and I have to replace the caps-4 of them are bulging and cracked on one small board inside. Sammy has a class action they are looking at because they didn't take care of the problem.

[Nelson37]

The problem IS still there, you just have a functional workaround. The mobo is suspect, you can replace it for free. This is a no-brainer. Eliminate the new boot-up procedure and verify if the problem still exists or not.

However, your workaround adds some additional, interesting info. The only thing you are eliminating is booting an OS from a cold hard drive. Try a similar, but slightly different test by immediately entering BIOS upon boot-up and allowing the PC to "warm-up" without any other changes. AND/or, test this theory by booting a Linux disk and IMMEDIATELY doing several hard-drive intensive operations, such as performance and read-write tests, a defrag, copying a large set of files, etc.

It may be that it is not the mobo itself that need to "warm-up", but the hard drive itself, and/or the drive interface circuitry. Or it may be that the basic boot you are doing is not working the faulty mobo component to a failure point.

I would still replace the board while a free replacement is available, it won't be forever.

[Webster]

Is the hard drive passing the S.M.A.R.T test?

I'll do what Nelson suggested. Download a liveCD Linux distro (Ubuntu, Mint, MEPIS etc...) and boot from CD to rule out Window boot problem. You may have a corrupted file in Window boot up sequence.

[Nelson37]

Some other thoughts.

Disconnect the two storage drives to lessen power load. Test. The fact that you need a cold boot somewhat complicates all testing, but such is life.

Remove the first two, and reformat one of the storage drives as a boot drive. Eliminate the "exotic" AHCI slipstream and install normally. Test this thoroughly on cold boots, then do it again with the AHCI setup, assuming the first test works.

Also, check the Event viewer and SMART logs, there may be some useful info there.

I'm thinking you may have a bad hard drive. Also, it is ALWAYS possible that there is more than one problem.

When replacing the board, often if you talk to a manager and somewhat enhance the problems this defective equipment has caused to such a loyal customer, they may offer you a significant discount on an upgrade mobo. Or some other goodie, nothing major but better than nothing.

[lacywest]

I recently had problems with a PC I was building last night [for a customer] ... I kept losing video to the monitor.

I thought maybe the Power Supply was just not stookum enuff for the video card ... ATI X1300 AGP video card.

So I opened up a box and put a new Power Supply in ... nope same errors occured.

So ... I thought maybe the video card was overheating ... so I added a different heatsink / fan to it ... I did get nice temperatures after doing this but the error still kept happening. So I changed the video card. All is well now. I swapped a ATI X1300 ... for a ATI X800 ... I really didn't want to do that but hey !! ... its not doing anymore blue screens anymore.

I wanted to mention ... the problem up above ... I did a Internet search ... and found that in the ATI Control Panel ... a feature called VPU Recovery ... was perhaps causing the problem ... another person had similar problem.

I disabled the feature in WINXP but couldn't find the feature in the ATI Control Panel in the WIN 7 version that ATI installs.

And the customer was anxious to get the PC back home ... his wife was pressuring him to get it from me ... his wife is pretty cool ... I've met her ... she has a nice smile.

I did have shortcut links to Crysis and Crysis Warhead in XP ... even tho it was installed on WIN 7 ... and the game played but ... gameplay was different ... in the level where I had to fight the enemies wearing nano suits in the Cemetary ... after I defeated them ... the next level was to begin ... nope it did not happen ... I was like ... lost in limbo.

So I was trying to install ... Singularity ... on the PC ... when the customer arrived ... he was hovering over me ... bugging me and he was concerned about the fan I had mounted on the bottom of the tower blowing air on the video card. He wanted me to put extra silicone rubber glue on all the legs holding the fan in place ... so here I was sticking my hands in the PC moving wires around ... to reach in ... POOF !!! ... BSOD ... CRAP.

He then said ... "Okay ... I got to take it now" ... I told him ... "wait ... I got to check it and see if it will boot up proper "

So I rebooted into WIN 7 ......... the game of course ... was screwed up and would not play ... so I deleted the install folder.

I few days later went by ... I see the customer at the store he manages ... he gave me the rest of the money almost ... he stills owes me $10 bucks.

And I find out ... they let their son use WINXP and the games installed in WINXP and they use WIN 7 for them selves. So the boy doesn't even get to play the Crysis Games I installed because they are installed on WIN7.

Oh and he brought me another computer to work on ... this one is a HP Pavilion 752n ... old PC ... and the Intel mother board doesn't even have a AGP socket for adding a video card to it ... but oh ... oh ... oh ... the Intel Mobo chipset can provide up 48 MB of expanded video ram. Geez ... where do these people find these computers ... are they hand me downs or are they getting them from yard sales ?

[lacywest]

My guess is that the power supply has bad capacitors. That will make it flakey.

I just replaced four bad capacitors in a 650 watt power supply that were bulging today. 3 x 1000uf 10V and one 2200uf 16volt. Most power supplies get crud brands used in them, Antec included.

These days when I am prepping a used computer for resale I check the motherboard and power supply for bad capacitors. We give a one year warranty so I'm being proactive that way. You'd be amazed how many I change that way.

That is a real bummer about Antec ... their PSs look great ... nice aluminum cases ... nice lettering and blah blah ... but they go bad quick. I use to get them at PC Club store in Fresno but not anymore. Before PC Club ended their business ... [a new store is now offering similar services in the building PC Club was at] I quit buying Antec and started getting Thermaltake Power Supplies ... so far the three I've bought ... are still working ... it's been a few years.

[lacywest]

Something to try ... don't mount the mobo into the tower ... put it on a cardboard box ... check underneath for any metal shavings or debris that could cause a problem and hook up everything in plain sight. Get a small screwdriver and use it touch the pins that turn on the PC.

When PC Club was still in business ... their technician had his own area to work on stuff ... I had to return a Mobo once and before I could get my money back ... they took the mobo out of the box and he quickly laid it down on this small spot where he had a power supply and stuff and when he observed that the mobo posted ... I got my money back.

A few months ago ... I did the same thing ... I cleared a area out on my workbench and checked different mobos for any errors. They were mobos bought cheap on Ebay.

[PartingShot]

Thanks everyone for continuing this discussion.

Something to try ... don't mount the mobo into the tower ... put it on a cardboard box ... check underneath for any metal shavings or debris that could cause a problem and hook up everything in plain sight. Get a small screwdriver and use it touch the pins that turn on the PC.

It turned out that my PC tower environment was the problem. Did almost exactly what lacywest outlined in the above quote, and the critical difference between a working system and a non-working system was the front panel connector for the RESET button. When it was connected to the mobo, it failed to POST. When it was not connected to the mobo, the system would POST. So I've opted to run this system without a RESET button for now. I'll probably replace the switch on the front panel sometime later.

It's been frustrating as hell, but it's finally resolved, thank goodness. Once again, my sincerest appreciation to the members of this forum who have offered their ideas and suggestions. This forum is simply one of the best out there

[lacywest]

Thanks everyone for continuing this discussion.

Something to try ... don't mount the mobo into the tower ... put it on a cardboard box ... check underneath for any metal shavings or debris that could cause a problem and hook up everything in plain sight. Get a small screwdriver and use it touch the pins that turn on the PC.

It turned out that my PC tower environment was the problem. Did almost exactly what lacywest outlined in the above quote, and the critical difference between a working system and a non-working system was the front panel connector for the RESET button. When it was connected to the mobo, it failed to POST. When it was not connected to the mobo, the system would POST. So I've opted to run this system without a RESET button for now. I'll probably replace the switch on the front panel sometime later.

It's been frustrating as hell, but it's finally resolved, thank goodness. Once again, my sincerest appreciation to the members of this forum who have offered their ideas and suggestions. This forum is simply one of the best out there

Okay ... get a Voltmeter / Ohm meter ... and check for contuninity in the wiring for the reset connection. If the needle pings over or shows activity with connecting the leads to the reset connection ... it is somehow shorting out to the main frame.

You could use a simple switch ... push button ... momentary touch type ... and mount it where it is easy to get to. Just push the button and it ... reboots !!