WHEA UNCORRECTABLE ERROR

[jimmy]

So I have been having some issues with one of my vgx xl100 machines, I am on Windows 10 and have got the r7 240 amd card installed. I have been getting the error WHEA_UNCORRECTABLE_ERROR on a Windows blue screen. Is there anything I can do to figure out what is going on? Logs I can check etc?

edit: I have noticed that Google chrome contributes to the crashing quite often. Im reading also that it may be to do with voltage applied to motherboard or CPU, is it possible to log these things to check?
This particular vgx is used alot and left running could it have caused this error to happen?

This happened on windows 7 also, so i clone drived my good vgx with the same hardware and plugged into the other vgx, seemed to work great until i was streaming online soccer and it cut to blue screen.

[jimmy]

For anyone else who may get this error in future, i 'think' i have solved it. First of all i deleted google chrome, turned updates as off as possible (ask for install) disabled away mode, CD drive, all unknown hardware or not working devices like marvell wifi (doesnt work for me in win 10), Installed known non-issue RAM.

I left the computer on for 4 hours and watched some films went online through internet explorer also ran a hardware monitor to see how high temps would go (50 degrees most)

[jimmy]

Nope... Still happening

I think this is a failing power supply, what are my options folks?

[Oliver]

VGX-XLs have a custom sized power supply, So if you do need to replace it you need another VGX-XL power supply, unless you are able to mod another type of power supply to fit.
I'm not sure if you can refurbish/repair your current PSU, but if you do need a new unit, you can sometimes pick then up on eBay, also try asking on here in the VGX-XLshop section.

Best of luck

[jimmy]

I have opened it up, and it is full of dust and dirt. Looks like some of the solder joints have are being joined by dirt so a short is possible, also there is a join between 2 legs on a board which doesnt look right, i will desolder and correct this aswell as hitting the unit with an air line.

[jimmy]

No fix yet still crapping out, over the weekend i will dismantle the 2 fans at the back of the board and clean, I'm trying to update the bios and also find out if the cpu has been overclocked..

[jimmy]

Could really dried out CPU paste cause this? And what is the heat sinned device across from the CPU?

[Oliver]

Jan 26, 2016 21:22:13 GMT jimmy said:

Could really dried out CPU paste cause this? And what is the heat sinned device across from the CPU?

Did you check for overheating? When you said about all the dust I wondered if heat could be the issue.
I guess clearing all the dust and changing the thermal paste can't hurt.
I use a program called CoreTemp to periodically check my machine's temperatures .
I was worried about temps after upgrading my CPU and graphics card, but I've not had any issues, in fact my VAIO runs pretty cool.

[IT Troll]

This is most likely a hardware problem with your CPU. It could be overheating and so the first step should be to check this with CoreTemp as Oliver suggested. The reads the temperature from inside the CPU itself.

However more likely is that this is due to incorrect or insufficient voltage being supplied to the CPU. This could be down the to the PSU, but it could also be the motherboard failing to regulate voltage correctly when under load, or could even be a fault with the CPU itself. I don't think your CPU will have been overclocked as the VGX BIOS doesn't have any options for doing this.

This sort of hardware system crash can be very bad for the O/S installation and can easily corrupt system files and registry. So following a system crash you may start to see other secondary O/S and application errors. To rule out the O/S completely it might be an idea to run MemTest86 or a Live Linux CD so you can stress test the hardware without Windows.

[jimmy]

I ran a cpu temp program which also applied a load test, it crashed straight away. I have 2 of these systems so i am going to buy some thermal paste and swap the ecu's if it still happens then i can rule out the ecu itself. Thanks for your help so far, anything further i find out i will continue to update this thread.

[IT Troll]

I suspect it is going to be the PSU / Motherboard which is failing to supply stable voltage when the CPU is under load. You should find that some diagnostic software like CPU-Z or SiSoftware Sandra can give you a reading of voltages in your system. If you set these to log to file then apply load you can then check if the voltage droops just before the crash.

I've used this technique before to troubleshoot overclocking problems which are quite similar in symptom. However on a motherboard which supports overclocking you can manually adjust the voltages.

[jimmy]

I ran intels cpu diagnostic test and it failed under heat. this is after i have re seated the cpu with golf thermal paste so i have just ordered 2 d945 cpus off ebay, in the mean time i will try the same test on my other machine and swap the cpus to see if i can narrow things down.

[jimmy]

Ok still working on this i have swapped and replaced thermal paste on the two machines i have, swapped PSU, CPU, RAM and HDD, each in series. So i was getting the BSOD with every swapped device throwing the same WHEA ERROR until i swapped the HDD i streamed soccer and the Connor Mcgregor press conference on youtube. Machine ran for about 3 hours then BSOD with the code DPC_WATCHDOG_VIOLATION, windows 10 OS.

Any ideas folks?

edit: just tried it again and got the crash BSOD KMODE_EXCEPTION_NOT_HANDLED