[SOLVED] "Clock Watchdog Timeout" on new build during software installs ?

[paul892]

Hello Everyone.

I really hope someone can advise me on this. I've built a number of basic systems for CAD software in the past, and decided it was time to upgrade again. But I've never had any issues before! This is my new kit:-

Intel Core i7-12700KF
MSI MPG Z690 Edge Wifi DDR4
Corsair Vengeance RGB RT DDR4 2x16GB 4600MHz
(kept old graphics card NVidia Quadro FX K600)
Western Digital Blue 500GB SSD
500W CiT 500U PSU
3 x Seagate 4TB HDD - raid 5 configured in BIOS

The issue I had is that after installing Windows 10 Pro (64) and booting, I obviously needed to install the motherboard drivers. Left along it had no issues, but as I started to install the drivers I had a couple of cases of the BSOD with a clock watchdog timeout error, though it froze at 0% rather than rebooting itself. In these cases it was always just after I had executed the driver install. I originally assumed this was due to the lack of drivers, and that once all installed it would be fine. This seemed to be the case to begin with, and I started transferring files / emails settings over with everything looking good, but then I went to install some PLC programming software (GX Works 2), which froze everytime it reached a USB comms driver install. This time it simply froze though - no BSOD, just no mouse, keyboard not responding, only holding the power button got me out of it.

It feels like it could be a hardware issue - but it only seems to fail when I'm trying to install something. I don't know much about fault finding this kind of thing - I usually Google until I find something that works - but this seems more complicated. So I really need the help of some experts! The fact that I can make it fail every time, and it seems like only some kind of install process causes the problem. I had a bit of a look in Event Viewer, but all I can see are things 1-2 minutes before the crash, things like DistributedCOM 10016 and Kernel-Processor-Power (Microsoft-Windows-Kernel-Processor-Power) 37.

Any ideas or advice on how to go about finding the issue?

Thanks all!

 

[JohnBonhamsGhost]

Any ideas or advice on how to go about finding the issue?

most of the times i've run into CLOCK_WATCHDOG_TIMEOUT errors they were related to insufficient voltage on a CPU.

if you notice it happening more often during software installation then this could be a sign of that since the CPU has ramped up a bit for unpacking compressed install packages.

run some CPU benchmarks and see if it repeatedly happens while the CPU is in these high usage scenarios.

 

[paul892]

Thanks for the input. I started to run a benchtest this morning, and had to leave it running while I dropped my daughter at school. When I got back the PC was switched off and completely unresponsive to the power button! I unplugged it for 10 minutes and then tried again and I've got nothing! Could this just be the PSU?

 

[paul892]

I had an old 300W PSU lying around which I used and managed to boot the machine back up again, and then I tried the original one again and it still wouldn't switch on, so I assume that means the PSU has failed?

One thing I have noticed whilst swapping things around is that the PSU I was using only has a 4-pin CPU power connector, but the motherboard has 2 x 8-pin. I believe the 4-pin supplies about 140W so enough for the base 125W of the processor, but not the "maximum turbo power" of 190W. Could this have been causing my issues?

Do I need to buy a PSU with 2 x 8-pin CPU connectors? Or would a 4-pin + 4-pin be enough to satisfy the power requirements?

 

[Deleted member 14196]

Yes. Get a proper psu with the connections you need

 

[JohnBonhamsGhost]

Do I need to buy a PSU with 2 x 8-pin CPU connectors? Or would a 4-pin + 4-pin be enough to satisfy the power requirements?

the majority of new motherboards, at least up to Z590 / X570, have either 2x 8pin or 1x 8 & 1x4.
only a single 8pin connection is needed for regular use or even regular overclocking.
the additional 4pin or additional 8pin is usually only there for EXTREME overclocking.

check in your manual for the mandatory power requirements for your CPU and use that.
it is possible that your Z690 + 12th gen actually needs the combination and that is why it was crashing while heavy processing was going on.

if you're using a modular PSU and don't have the proper cables, contact their customer support and tell them you need an updated package of cables for Intel 12th gen use.
it's possible they will send you a new kit for free, many manufacturers were offering the new PCIe RTX 3000 FE GPU cables for free also.

 

[paul892]

Thanks guys. Ordered a new PSU with the right connections so should be able to test this properly tomorrow. Hopefully I'll be back with good news.

Cheers!

 

[paul892]

UPDATE!!

Thoroughly confused and rapidly losing faith….

New PSU arrives, plugged it in loose on the floor, powered it up and went to install the program and driver which had been making it freeze. Both worked!! Hurrah I thought, and set about dismantling and reassembling the machine with the new PSU, carefully running everything neatly and tidily as I was done.

After doing this I decided that to make sure there were no issues with the windows installation after all the freezes by reinstalling windows from scratch. When I came to start reinstalling the motherboard drivers all was going well until it froze again. Same clock timeout error. To begin with I hoped this was just because all the drivers weren’t yet installed, but bizarrely I can’t actually get all the drivers installed because this keeps happening. I have one serial driver left uninstalled and the software which installed fine in my first test now won’t install again.

Please can someone offer some further advice or suggestions? I’m starting to lose the plot here, I thought this would be easy!

Thank you.

 

[JohnBonhamsGhost]

start with the manufacturer rated CPU voltage.
sometimes "Auto" settings don't play well.
each freeze or crash you experience raise the voltage .005v at a time.
hopefully you can find a stable voltage and don't have to troubleshoot further.

also try disabling any Multi-core Enhancement, Thermal Voltage Boost, or any auto-overclock type settings.

and also make sure your memory is running at manufacturer rated specs.
if there are multiple XMP profiles available some may have some overclocked settings.

 

[Deleted member 14196]

first do everything suggested above my post. then make sure to use memtest86 to test your ram sticks, you want to see zero errors in 4 passes

memtest86 makes a bootable usb and runs from that so no windows freezing to test ram fully. i would test each stick individually as well.

MemTest86 - Official Site of the x86 and ARM Memory Testing Tool

 

[paul892]

I ran a Prime95 stress test before your comments above, and ran for about 35 minutes without a flicker. Surely these kind of tests would be more stressful to the CPU than a basic software installer?

I also ran a benchmark test on UserBenchmark and got great results on everything apart from the graphics card (which I know is old):-

MSI MPG Z690 EDGE WIFI DDR4 (MS-7D31) Performance Results - UserBenchmark

I then got your message and tried increasing the CPU voltage, but I've got up to about 1.31V now and nothing is any better. I really don't understand why it worked when I first plugged it in this morning but it won't work now?

Is there something I can have running when I try the installs to monitor what is happening just prior to the crashes?

 

[JohnBonhamsGhost]

I've got up to about 1.31V now

i actually saw one report of a user needing to set it to 1.38v to run stable.
whether this was an issue with the motherboard or just a lower quality chip, ???

 

[paul892]

UPDATE - SOLVED! - (I think)

I can't believe what fixed this. Being old school I always put a DVD drive in my PC builds, and this was no exception. After another day of toiling yesterday I tried to think what was different about my "open floor" test which worked, and my finished setup. One was it was a completely cold test (it had been off for 10 hours), the other was I only connected the SDD main drive. This means I hadn't connected the 3 RAID5 hard disks or the DVD drive. So first thing this morning I unplugged the DVD and tried again. It worked. Put the drive back in - failed again. So then I thought is it the drive? the cable? or something else. The motherboard has 6 SATA ports, A, B, 5, 6, 7, 8. The raid ones are 5, 6, 7 and 8 so I had the SDD in A, the DVD in B, and 5, 6 and 7 the raid drives. I switched the DVD into 8 and it still worked. So it turns out that the motherboard does not like a DVD (or this DVD drive) in sata B port? I don't know why - there's nothing in the manual about this. It just describes them all as 6GB/s ports.

Anyone know what the difference is between A/B and 5/6/7/8?

Thanks anyway for all your input!