I've had a stable install of the latest version of P3D 4.5 for over a year now, use it nearly daily, at the same config settings you see in my forum signature below:  [email protected] all cores, HT disabled.  Core temps are very low, I run P3D w/ quite low autogen/vege/distance settings and rarely see the main thread go over 70% in the NGXu or PMDG 777.  And GPU is rare over 60% of peak.  This is the way it's been for over a year now.

All of a sudden, I'm getting a consistent CTD quite quickly after loading a flight plan I just created.  I've tried a couple different plans involving different departure and arrival airports and it's always the same.  The plane loads at my departure gate setup in the plan.  I begin to start setting up my flight instrumentation as per usual.  I use ASN4, FSCaptain, REX Sky Force, and that's pretty much it.  Within just a few minutes, maybe around 10, up pops the blue screen of Win 10 letting my know something is wrong and in x seconds it will reboot.  If I'm not mistaken here there have been different error codes displayed.  Then during reboot DOS-type white text suggests hitting Alt-L to attempt the last successful boot.  It appears to look like an overclock failure which I've not had previously.  In fact after getting these I used Intel Extreme OC utility to stress test CPU and RAM, and 30 minutes of each at 5.0Ghz were passed.  The CPU runs very cool using Coretemp, rarely over 50C but often low 40's at [email protected], and that is true in the stress test as well as while flying P3D or MSFS.

I have a dual boot system, and the 2nd and identical NVme drive contains a cloned bootable copy of the main drive, done about 8 months ago.  It contains my complete P3D & addons, except a couple of months ago I purchased and updated FMS data w/ Navigraph and FSAerodata about 2 months ago on the main drive, and did not re-clone to the 2nd drive, so it does not contain those updates.  Nor does the cloned drive contain the most recent Win 10 updates as I just booted to it this morning and it has not yet demanded Win 10 updates.

Here's the weird part:  I can boot to the clone, do a complete flight in P3D 4.5, and all's well--no CTD.  Certainly hints the hardware is fine.  I might just continue to boot to the cloned drive for P3D now, but would rather not at the moment and would like to find out what the hey happened to the first drive and P3D.  Of note, I can run MSFS in my main drive w/o a CTD.  So this has something to do w/ P3D in particular.  I don't want to re-clone from drive 2 to drive 1, in case something fails in the clone operation, plus on the main drive I have MSFS and a few steam titles and those are not on the clone.

One possible clue:  i have to disconnect my Saitek Yoke after bootup else the aileron axis does not center.  I did this while P3D was open, actually have done it maybe 3 or 4 times because I forget to disconnect/reconnect the yoke prior to loading P3D, and a buzzing, hardware-related CTD happened 2 of those times from I think it said something about an IRQ conflict.   That may have played a role in future CTD I'm not sure but one of these happened I think just before the CTD began occuring, and thusfar it's always a CTD w/ any attempts at running P3D on the main drive now.  Anyway, I also moved the yoke to a different USB port but still I can not run P3D w/o a CTD very quickly after starting a flight, within about 10 minutes it's over.  

What would you try to troubleshoot this oddity?  Should I reinstall the Client component of P3D perhaps, overtop of the current install?  I would rather not uninstall all of P3D and it's addons if I don't have to, but I could.  

Thanks in advance!

 

A 30-min run of the Intel burn-in utility isn't enough to rule much out.  I'd start with a 8-12 hour overnight run of MemTest86+ to see if you have a memory stick going bad.  Memory problems are common enough that it's the first thing I try to rule out when a previously stable system starts throwing blue screen errors.

An IRQ conflict on a USB port is pretty strange.  Are the system drivers and BIOS up to date?

A diagnostic run on your drive is also a good idea...could well be that your drive is headed for the bit bucket in the sky.  I use Samsung SSDs exclusively on my machines, and Samsung's Magician utility is a pretty good tool for scanning those devices.  Perhaps Sabrent has some similar utility, or some third-party software is available.  Just be sure it's designed for SSDs, as disc-scanning software designed for HDDs is somewhere between useless and even harmful when used on an SSD.

If you know what chipset your SSD uses, the manufacturer will have a utility for it. Sabrent will just have badged this as their own (he says literally as he installs a Sabrent NVMe drive).

I'd echo the advice to do a longer test of your hardware components. However, the fact that your cloned image runs OK suggests to me that it's a Windows issue related to a recent update or change that you've made, rather than P3D - unless you've made config changes or had add-ons on the main image that you don't on the other. That MSFS runs is not necessarily diagnostic - P3D uses the hardware and OS differently. One thing I'd do is boot each drive and write down the list of KBs installed on each (and make sure Windows Update is off or the PC is disconnected so it can't start updating the cloned image). Compare and see what's been added, then Google the KBs that you only have on the main image. See if anyone's seem similar issues.

 

7 minutes ago, w6kd said:

A 30-min run of the Intel burn-in utility isn't enough to rule much out.  I'd start with a 8-12 hour overnight run of MemTest86+ to see if you have a memory stick going bad.  Memory problems are common enough that it's the first thing I try to rule out when a previously stable system starts throwing blue screen errors.

An IRQ conflict on a USB port is pretty strange.  Are the system drivers and BIOS up to date?

A diagnostic run on your drive is also a good idea...could well be that your drive is headed for the bit bucket in the sky.  I use Samsung SSDs exclusively on my machines, and Samsung's Magician utility is a pretty good tool for scanning those devices.  Perhaps Sabrent has some similar utility, or some third-party software is available.  Just be sure it's designed for SSDs, as disc-scanning software designed for HDDs is somewhere between useless and even harmful when used on an SSD.

Thank you for that.  Of note though, I can run P3D on the 2nd drive no problem so in terms of troubleshooting an overclock why would that not suffice?  After all, that's all I've ever done is set my PCs up so they could run a stable P3D, versus Prime35 or the better stress tests.  I have 1000's of stable hours on that.

So perhaps as you say the drive is going which seems plausible, so I will look for an SSD diagnostic.  Perhaps because the installs aren't identical on main and clone, very likely different memory might be involved, so MemTest86+ sounds like the right idea.

I will follow up on those suggestions.  Thanks

5 minutes ago, neilhewitt said:

If you know what chipset your SSD uses, the manufacturer will have a utility for it. Sabrent will just have badged this as their own (he says literally as he installs a Sabrent NVMe drive).

I'd echo the advice to do a longer test of your hardware components. However, the fact that your cloned image runs OK suggests to me that it's a Windows issue related to a recent update or change that you've made, rather than P3D - unless you've made config changes or had add-ons on the main image that you don't on the other. That MSFS runs is not necessarily diagnostic - P3D uses the hardware and OS differently. One thing I'd do is boot each drive and write down the list of KBs installed on each (and make sure Windows Update is off or the PC is disconnected so it can't start updating the cloned image). Compare and see what's been added, then Google the KBs that you only have on the main image. See if anyone's seem similar issues.

 

Thank you Neil.  I see Windows Update was not disabled (I didn't think that was possible, is it?), so when I turned the cloned OS off, I just hit the power button when initiates the SHUTDOWN sequence.  Because I have a dual boot I then am able to boot back to the main drive.  I guess I could ask Sabrent what chipset my particular two NVMe drives use.

There was a recent Windows update also prior to this.  I did a system restore from about 10 days prior when everything was fine.  It didn't fix our problem though.  

I should also try rolling back the overclock to 4.7Ghz and see if it happens.

50 minutes ago, Noel said:

This is the way it's been for over a year now. All of a sudden, I'm getting a consistent CTD quite quickly after loading a flight plan I just created.

The first thing I'd do is bring it outside or garage and compress it.

Blowing it out is a good idea, but always be careful with fans...you don't want them spinning up with an air blast, could damage the bearing(s).  If you hold the fan with your fingertip or be very cautious with the air it should be ok.

The other suggestions are where I would start, too.  The only thing I might add has to do with AVX instructions...I'm not sure to what extent P3D uses AVX instructions, but it may be an added stress/load factor that may (or may not) have been used with your stress test software.   I *think* I remember that P3D doesn't make much use of AVX instructions, and so I don't normally test a rig with that, BUT it's been a little while since I configured a new system, so it might be something to look into.

21 minutes ago, Mace said:

The other suggestions are where I would start, too.  The only thing I might add has to do with AVX instructions...I'm not sure to what extent P3D uses AVX instructions, but it may be an added stress/load factor that may (or may not) have been used with your stress test software.   I *think* I remember that P3D doesn't make much use of AVX instructions, and so I don't normally test a rig with that, BUT it's been a little while since I configured a new system, so it might be something to look into.

P3D does indeed make use of AVX...enough so that the CPU will run with any AVX offset applied if set in the BIOS.

2 hours ago, Mace said:

Blowing it out is a good idea, but always be careful with fans...you don't want them spinning up with an air blast, could damage the bearing(s).  If you hold the fan with your fingertip or be very cautious with the air it should be ok.

The other suggestions are where I would start, too.  The only thing I might add has to do with AVX instructions...I'm not sure to what extent P3D uses AVX instructions, but it may be an added stress/load factor that may (or may not) have been used with your stress test software.   I *think* I remember that P3D doesn't make much use of AVX instructions, and so I don't normally test a rig with that, BUT it's been a little while since I configured a new system, so it might be something to look into.

Thank you Rhett, someone here named Martin mentioned this when I built the PC a year ago, a different stresstest that does AVX instructions.  I have quite a few hours on this, it runs super cool, but maybe it needs a little more voltage.  I've been getting away with [email protected] but maybe it's degraded a bit too.      I will go ahead and blow it out.  Again, it's only P3D that is doing this, nothing else, so far.

Well, apparently it had a bad case of 24h flu or PC-covid.  I thought I'd give it a try today, P3D from KLAS to KSEA in the NGXu and all is well now after a good 60min.  I really don't know what happened but I tend to think it's hardware related--perhaps a bad audio jack, or something unexpected like that.

I used Win 10's memory diagnostic which is very deep when you choose the Extended option which takes many hours to perform and is done outside of Win 10 like MemText86+ is, and it passed that test.  I did not find a SSD tester but used Sabrent's Rocket Control Panel which displays 'health' and reports it at '100%'; and displays Smart results as well and all seem perfect. This may not be a sufficient test so we'll wait and see what happens going forward.

Doing more CPU stress testing seemed not particularly relavent since all other apps, including Intel Extreme Tuning stress test were fine, temps cool, etc.  Even so, I used RealBench which does evaluate AVX and after 30 minutes of maximum stress as well for my 32GB of ram, and all was well again.  Max temp for the entire test was 55C.

What I did not do was take the box out and blow it out.  I will go ahead and do that today.  If it screws up again after that, perhaps I do have something like an audio jack or some other component that is not stable in its use.  I have bumped into the cable that connect my headset and put a little stress on the jack, but not much.

So really, it's kind of a 24h flu of sorts.  I did multiple attempts at flying again after crashes and reboots , moved USB ports for my yoke, system restore, and things like that and nothing helped.  Seems well right now!  Weird!

Thanks all.  If it screws up again after blowing it out I will report back. 

Cheers :o)