a reminder to consider hardware failure when troubleshooting

[molleh 451]

so recently i've been getting back into MSFS (having not really been too interested in it since launch until some good aircraft addons were released) and one problem i had was the sim randomly crashing. it wouldn't give an error message, blue screen, or any other alert, it would simply close and leave me with no clue as to what happened except for application faults listed in Event Viewer. the recorded events would typically list flightsimulator.exe, nvwgf2umx_cfg.dll (part of the Nvidia driver package) or a couple other system-related DLLs. the sim would otherwise run perfectly, sometimes for more than an hour at a time, but the crashing was definitely frequent enough to make any serious flying a no-go.

at first i assumed it must be that MSFS is just an unstable, buggy platform as when i searched forums for similar crash reports, i found quite a few (right down to some with the nvwgf2umx_cfg.dll listed sometimes) which seemed to confirm my assumption. however, the other day i was playing a game (Assetto Corsa Competizione, a racing sim) and i had a crash there, too - at which point it occurred to me i'd actually been seeing random crashes pretty frequently (though not obnoxiously frequent) for the past several weeks, maybe even two months or more.

that's when it dawned on me i should probably start checking my PC hardware for faults, including RAM, CPU, GPU, and storage. the first thing i did was run memtest, a free and simple tool that tests your RAM. lo and behold, running Memtest showed errors within 30 minutes (and a properly working PC should be able to run Memtest for literally days on end without error.) immediately i knew i found the actual culprit behind my crashing symptoms. i had four sticks of RAM so i tested two at a time, and in each pair of slots, and they produced errors every time. thankfully i was able to borrow some known good RAM of similar specs which i swapped in and Memtest ran for 12 hours without errors. MSFS no longer crashes.

just a little reminder for y'all to keep hardware in mind whenever you're faced with troublesome crash-to-desktop errors that seem to have no cause. in my case it was easy to find the issue, but it could've been my GPU for example, which perhaps could be pinpointed with a stress-testing tool like Furmark.

MSFS is probably one of the most system-intensive games/sims on the market right now - it hits your CPU, GPU, RAM, storage devices and even your network/internet access quite hard, and it'll find the weakpoints in your PC in ways that most other games won't. also, this isn't the first time i've had RAM fail on me, but it's definitely the first time that the issue was pretty difficult to spot at first. i think Windows 10 and 11 are quite stable compared to what we had prior, and what probably would have been BSODs or hard system locks that require manual reboots were instead benign CTDs without error messages. it's a good thing overall, as it keeps the system running despite component failure, but it makes the core problem a little harder to diagnose.

Edited May 2, 2022 by molleh
typos

[irocx 292]

3 hours ago, molleh said:

MSFS is probably one of the most system-intensive games/sims on the market right now - it hits your CPU, GPU, RAM, storage devices and even your network/internet access quite hard, and it'll find the weakpoints in your PC in ways that most other games won't.

Great points.  Asobo gets considerable bashing (unfairly imho) about CTD's when I suspect much of the issues are actually hardware related.  If they could incorporate a better crash logging system that could help users/developers identify problems, it would be interesting to see how much of that is actually true.  Conversely, they are still having CTD problems on the closed/universal XBox architecture, so it indicates there are some problems in the code.   

[AnkH 2,267]

This and the fact that many users have such a messy system with bloatware all over and the need to have a zillion things open in parallel. And then they wonder, what could go possibly wrong... 

[Humble Jim 51]

Remember, too, to check for dust clog on the filters fitted to your PC case. I do this at about 6 month intervals.

Whatever your CPU cooling method, MSFS will generate heat in your case and will exercise case fans.  A lot of dust on filters will often cause those fans to work even harder...and more dust builds up.

Don't just stop at the main case fans front and back. Most tower type PC cases have a filter on the underside of the case (below the PSU).  Many don't know it's even there. If you have one, keep that clean too.

Aside from the basic cooling benefit to your system as a whole, filter cleaning can help keep the general PC noise down.

Edited May 3, 2022 by Humble Jim

[cianpars 1,165]

I spent ages trying to diagnose and fix my CTD issues with MSFS only to find after considerable amounts af ballache that it was a failing PSU.  In the past, when a PSU has failed it has gone suddenly, but the last one was failing slowly over several months.

[wiler 394]

6 hours ago, irocx said:

Great points.  Asobo gets considerable bashing (unfairly imho) about CTD's when I suspect much of the issues are actually hardware related.  If they could incorporate a better crash logging system that could help users/developers identify problems, it would be interesting to see how much of that is actually true.  Conversely, they are still having CTD problems on the closed/universal XBox architecture, so it indicates there are some problems in the code.   

I made a thread about CTD diagnosis here yesterday and not one reply or at least comment. shows you how much people really care about that topic. I did find my CTDs were related to the program settings itself, by clearing the ultra/high/med graphics settings to low and saving and then putting them back where they were. CTDs seems to go away , at least for now.  A file that tells u which .dll or file caused the conflict like in xplane11 would be nice ....

[JSmith2112 317]

To add to @molleh's post, recently I started getting system reboots while in FS2020 - no bluescreen, just a pause in the sim followed by a reboot. At first it was only in FS020 where I would get these reboots after about 30 minutes but then it started affecting other games as well and the reboots would come within minutes after starting any game.

Memtest didn't find any errors and after some troubleshooting I decided that the motherboard was probably the culprit. An upgrade to a Z690 board and i5-12600KF later and FS2020 is running great agan.

While waiting for the parts to arrive I removed my GPU fan shroud, cleaned out the fans and heatsink fins and cleaned out the case fans.

The upgrade didn't make any really noticeable difference in sim performance but subjectively it does seem smoother. My photography and video editing apps seem much faster than before though. 

[IcedOutN8 6]

Great post OP! Testing against other flight simulators would be a great test to run on one's PC but who's to say if everyone has a copy of FSX, X-Plane, and/or P3D. Perhaps if there are any other gaming titles installed on the machine, a test of running those (at maximum settings to re-create the intensity of MSFS) maybe a good comparison to determine hardware faults.

[Fiorentoni 4,234]

Hardware failures aren't that common after all. In most cases it's a coding/driver/software error, which can - to be fair- also be caused by the user itself (Nvidia CP settings all over the place, anti-virus tools).  MSFS is notoriously bad at tracking the cause for CTDs, as already mentioned, and this is something Asobo has to be blamed for as much as possible. The crash logging (or lack thereof) is not acceptable.

That said I think most CTDs are caused by addons. I don't know how often I've read "CTD, community folder empty!" and 10 posts later the person adds "Oh well I am using the google maps mod, I didn't think this was a problem". Sometimes people hesitate to really check their mods because they do not want to have to abandon them (I can relate to that).

[kerosene31 349]

Basically, MSFS is not good for testing, because it does have lots of software CTDs.  CTDs are the #1 issue on Asobo's buglist and have been since day 1.  It isn't always hardware.  Once we start dumping dozens of add-ons into the sim, tracking down CTDs can be difficult if not impossible (there's a dedicated CTD forum here for a reason). 

That's why it is important to find another game or program to stress test your PC.  If you get instability in other games, then you know hardware is the likely cause.  If your PC has zero other issues in any other games, you probably don't want to be ripping out all your hardware. 

[Bob Scott 7,721]

The point of the OPs post--and a good one--is not that most CTDs are caused by hardware, but that some are, and if you're plagued by multiple crashes, especially those that don't share some commonality w/r/t what's running at the time, don't summarily discount hardware as a potential cause.

Memory failure or instability is actually not that uncommon, especially when running binned and highly overclocked DIMMs (which generally use DRAM ICs that are rated by the chip manufacturers at much lower speeds).

I've had several instances where RAM sticks either went bad or became unstable at their manufacturer-overclocked SPD settings.  Also, if there are bad addresses that are way up high in the address space, when running other games/software that don't use as much RAM you may never encounter the bad or iffy segment(s) of memory.  So "it works for other programs but not for MSFS" does not necessarily point to MSFS as the culprit.

One trick though--if you do have a DIMM suddenly start throwing errors--especially fast DIMMs heavily overclocked by the manufacturer, quite often just adding just a little voltage to the Vdimm settings or reducing frequency/relaxing timings in the BIOS can clear that up.  I've actually had DIMMs that wouldn't run at the manufacturer's spec voltage right out of the box, but worked fine with a couple hundredths of a volt bump up from rated Vdimm.  That could be explained by inaccuracies in the voltage actually applied by the motherboard--the testing equipment used to bin memory in the factory is maintained to a high level of precision, but motherboard voltage control circuitry is generally not nearly as accurate.

[bean_sprout 1,201]

Also the team can analyze and correct software errors but not so much with your hardware CTDs.

ns

[rayharris108 259]

This thread makes some excellent points. I can confirm I had memory errors and constant CTDs. I eventually found my Motherboard was faulty.

since upgrading to a new PC zero CTDs   Including SU9 beta and SU9, mostly running on HIGH TO ULTRA at 2 k. 
 

whilst the sim can certainly be unstable at times I really think that many CTDs are not primarily the fault of the sim but more likely mods, add ons, Windows driver or hardware errors. Otherwise we’d all be getting similar CTDs like in SU5 wouldn’t we?