Black Screens, Fans 100 Percent Behaviour During DCS

[LucShep]

8 hours ago, nephilimborn said:

Thanks LucShep, ordered the wireview. 

Also check 4 different power supply calculators, the majority of which seem to indicate that i am at the edge or above 80% power draw for the MSI 1000G PSU. I might just also pull the trigger on a new PSU; considering the XPG CORE REACTOR II 1200W which, according to Hardware Busters has very good 12Volt rail properties https://hwbusters.com/psus/xpg-core-reactor-ii-1200w-psu-review/ and is assessed as top tier on the cultist psu list https://cultists.network/140/psu-tier-list/

I tried undervolting with GPU Tweak III and got mixed black screen behaviour when running the Unigine Superposition stability test - i ended up just ignorantly setting the power target to 90% for now. Will put more effort into tuning the VF curve when the thermal grizzly wireview arrives.

This sh*tshow is driving me into mental instability - replacing the PSU is driven more by desperation rather than any evidence that the MSI PSU is the root cause.

 

If black screens occur with undervolt then, yeah, matter of optimizing curve (a little less clock perhaps).  But if 90% is doing ok, then whatever works best for you.

And yes, Cultist's PSU Tier List and especially anything HWBusters informs is reliable (as good as the good old JohnnyGuru website, IMO).
So, any recommendations there can be trusted.

BTW, just my opinion but, if you ever consider replacing the MSI 1000G (already a great PSU), then might as well go the extra mile and pick a good 1250W+ PSU.
Pretty expensive and somewhat overkill, yes, but with a high-end system like yours, I personally think it pays up in the longer term (may even be reused again on next system).
https://hwbusters.com/best_picks/best-atxv3-pcie5-ready-psus-picks-hardware-busters/7/

All that said, I do agree with @kksnowbear above.
Also, it looks to me as well that the PSU Cable seems to have a bit of a tight bend there, right before the GPU connector.
Could be it(?). 

As I read that you've ordered the thermal grizzly wireview (nice one!), it will perhaps also help aliviate that common issue.

As side note, I know it's not everybody having the problem but... if it's enough of an issue for so many, then it shows to be an issue with the concept (not of user handling).
Not so sure I'm alone when I say that the 12VHPWR connector is among the worst things Nvidia did in many years... 
 

Edited August 21, 2024 by LucShep

[kksnowbear]

45 minutes ago, LucShep said:

As side note, I know it's not everybody having the problem but... if it's enough of an issue for so many, then it shows to be an issue with the concept (not of user handling).

Of course, I see your point, and it's reasonable (at least initially). But...

TBH, what this actually reflects more than anything is that computers are (and have always been) somewhat complex devices.  Somewhere along the line - I can only imagine to increase sales - the industry convinced consumers it's a totally "DIY" matter.  They simply sell a *lot* more if everyone believes it's easy to build your own.

I strongly disagree, and always have.

The problem is, it actually has gotten fairly easy.  Pretty much anyone can walk into a computer store (or buy on Amazon) and check out with an entire machine worth of parts.  They don't have to know crap - and boy, does it show sometimes.

They're not going to stop selling this way just because I don't like it, of course.  But I do spend a considerable amount of time working to overcome mistakes that were made by other, less experienced builders.  And let me promise you, some of the stuff I see is horrendous from a technical perspective (if not dangerous outright).

I spent several years pursuing formal education in this field.  More still on advanced technical training (like micro-mini electronics repair in US Navy avionics; arguably the best in the world).  And decades honing a skill set.

I'm sorry, but you just cannot replace that with a box of parts from Amazon.  End users are responsible for more problems with hardware and building by far than anything associated with design.

For some reason, everyone seems to understand that doctors exist for a reason.  And we all know what they say about a fool and his lawyer.  But for some reason, some people seem to think that spending money on parts and/or paying others to build things for them somehow makes them an expert.

But, somehow, people are still paying me to fix their f*ckups.

I would respectfully submit that what's happening (and has) - at least in part - is that these people are in over their heads.  They simply lack a level of skill and training appropriate to the stuff they're messing with.

To contend anything else is to dismiss the lifetime of training and experience earned by professionals just like me.

TBH the smart ones are the ones who seek out and commission qualified expert help, with verifiable references.  If nothing else, they'll stand behind their work.

As I saw once on a plumbing truck:  "Do it yourself first, then when it breaks, call us."

(Now this is straying a bit off topic and I'll apologize; I am only responding to someone else's comment)

Edited August 21, 2024 by kksnowbear

[LucShep]

10 hours ago, kksnowbear said:

Of course, I see your point, and it's reasonable (at least initially). But...

TBH, what this actually reflects more than anything is that computers are (and have always been) somewhat complex devices.  Somewhere along the line - I can only imagine to increase sales - the industry convinced consumers it's a totally "DIY" matter.  They simply sell a *lot* more if everyone believes it's easy to build your own.

I strongly disagree, and always have.

The problem is, it actually has gotten fairly easy.  Pretty much anyone can walk into a computer store (or buy on Amazon) and check out with an entire machine worth of parts.  They don't have to know crap - and boy, does it show sometimes.

They're not going to stop selling this way just because I don't like it, of course.  But I do spend a considerable amount of time working to overcome mistakes that were made by other, less experienced builders.  And let me promise you, some of the stuff I see is horrendous from a technical perspective (if not dangerous outright).

I spent several years pursuing formal education in this field.  More still on advanced technical training (like micro-mini electronics repair in US Navy avionics; arguably the best in the world).  And many more years honing a skill set.

I'm sorry, but you just cannot replace that with a box of parts from Amazon.  End users are responsible for more problems with hardware and building by far than anything associated with design.

For some reason, everyone seems to understand that doctors exist for a reason.  And we all know what they say about a fool and his lawyer.  But for some reason, some people seem to think that spending money on parts and/or paying others to build things for them somehow makes them an expert.

But, somehow, people are still paying me to fix their f*ckups.

I would respectfully submit that what's happening (and has) - at least in part - is that these people are in over their heads.  They simply lack an appropriate level of skill and training to contend with the stuff they're messing with.

To contend anything else is to dismiss the lifetime of training and experience earned by professionals just like me.

TBH the smart ones are the ones who seek out and commission qualified expert help, with verifiable references.  If nothing else, they'll stand behind their work.

As I saw once on a plumbing truck:  "Do it yourself first, then when it breaks, call us."

 

Not entirely disagreeing, but my point is -  there was absolutely nothing wrong with PCIe 6+2 connectors.
How many issues you've seen with them?  (I never seen any in 20+ years that connector has been used)

Also, the location of the 12VHPWR in the GPU itself, for ALL of the RTX4090s in the market, is wrong. 
It should never have been within/below/above the PCB, but at the end (i.e, at the side) of the PCB.
Because the connector, as it is currently, is then being heated by the fans exhausts in addition to all the massive electric current already going in that single connector. 

If only EVGA never pulled the "we quit" action, and went ahead with their own idea of how a RTX4090 should be (my guess is they knew all along), I think half of the issues would've never ocurred....  

 

That's their RTX4090 FTW3 prototype, which never went through production. (they've decided to quit right before the RTX4000 series launch)
Picture taken from the video:  youtu.be/tYzJf71WUcM

Edited August 21, 2024 by LucShep

[kksnowbear]

LOL Well, we might agree to disagree about the 8pin PCIe connectors.  Yes, they're more robust, I wouldn't argue that.  But I myself have cursed the 'hose monster' that 3 8p connectors entails, along with routing/management/dressing, etc etc etc...and I'd bet anyone who's done this more than a few times knows what I mean.  Not to mention that per the PCIe specs of the time, those connectors were already at their limits (never mind they are capable of more, I'm referring to what the specs said).  I could be wrong, but I want to say that three 8pin connectors doesn't even technically meet spec under PCIe at the time - GPUS were limited to 300w, I seem to recall(?) And I actually own two eVGA 770 Classifieds, as well as a couple other models that have three 8p connectors.

A better solution was long overdue (Whether the 12VHPWR concept was that 'better solution' is another matter).

LOL The eVGA pic looks very familiar...I'm sure that's the tail end of the 3090Ti I had.  And yes, it's a damn shame about eVGA...but we also have to acknowledge, 'end' connectors like that don't work for everyone either (these cards are long enough as is).  I think it's likely that the industry settled on the side connector because (of the two admittedly tight places) there was less likelihood of space constraints with the side connector than the end type.  Lesser of two evils.

 

[LucShep]

2 minutes ago, kksnowbear said:

LOL The eVGA pic looks very familiar...I'm sure that's the tail end of the 3090Ti I had.  And yes, it's a damn shame about eVGA...but we also have to acknowledge, 'end' connectors like that don't work for everyone either (these cards are long enough as is).  I think it's likely that the industry settled on the side connector because (of the two admittedly tight places) there was less likelihood of space constraints with the side connector than the end type.  Lesser of two evils.

If the connector was pointing up/down (or had a "L" convertor like those currently sold by 3rd parties) that wouldn't been a problem....
Being far of the "hot zone" (and usually where the intake fans of most ATX cases are) would have prevented a pretty big part of the "heat" issues....

[kksnowbear]

Oh I'm not saying you're wrong about heat.

I'm saying there's obviously a reason it wound up where it did (perhaps even despite the heat)...they don't generally spend the kind of money that goes into designing these things by accident - and my thought is that it's because of space.

13 minutes ago, LucShep said:

usually where the intake fans of most ATX cases are) would have prevented a pretty big part of the "heat" issues....

I get what you're saying, but you can't assume that.  Plenty of cases don't have intake fans, obstructed by drive cages, etc.

If you assume the case has intakes there as part of your cooling design, that's a mistake IMHO.  In fact, it could actually be that the designers felt that residual airflow from the GPU fans was enough to cool the connector area, much the same as both Intel and AMD stock CPU cooler designs used residual airflow to cool motherboard VRMs, chipset, RAM etc.

Where you can't count on front intake fans, the GPU fans will always be there - and you can control their speed, too.

Edited August 21, 2024 by kksnowbear

[LucShep]

2 hours ago, kksnowbear said:

Oh I'm not saying you're wrong about heat.

I'm saying there's obviously a reason it wound up where it did (perhaps even despite the heat)...they don't generally spend the kind of money that goes into designing these things by accident - and my thought is that it's because of space.

I get what you're saying, but you can't assume that.  Plenty of cases don't have intake fans, obstructed by drive cages, etc.

If you assume the case has intakes there as part of your cooling design, that's a mistake IMHO.  In fact, it could actually be that the designers felt that residual airflow from the GPU fans was enough to cool the connector area, much the same as both Intel and AMD stock CPU cooler designs used residual airflow to cool motherboard VRMs, chipset, RAM etc.

Where you can't count on front intake fans, the GPU fans will always be there - and you can control their speed, too.

 

The "conspiracy theory" is how EVGA built their prototype is also how they adviced it to be, period, but it was more expensive to manufacturer.
They decided on a much bigger cooler to circumvent the 450W+ resultant heat, and that would get in the way of construction costs, when it's already a very expensive GPU.
So, silly monumental size won and "no no, no can do on connector in that place... eff off".

The fact that it's a much bigger problem in the 4090s than it is in the 4080s (not even 10% versus of the RMA for that issue, according to my sources) also exhacerbates how badly it was planned and done.
IMO, it should never been more than 300W in that 12VHPWR connector (total power includes PCIe slot socket) yet they went ahead anyway with a single one even on models getting close to 600W - (IMO) should have been two connectors, not just one, but I guess even that would get in the way of lucrative returns...  lol

Edited August 21, 2024 by LucShep

[kksnowbear]

Well, as much as we may hate what the business needs dictate from time to time....

...we'd probably all hate it a lot more if Nvidia weren't around, due to crappy business decisions.

I admittedly haven't done the legwork myself, but it is my assumption that the manufacturers did the math on the current/wattage/heat capacity of the connector.  They don't generally have meetings where they all agree "Let's put this connector on here which isn't capable of handling the power, so we're sure to get sued..."

In my estimation, what happened was they may have 'cut it close' but can readily defend it's within tolerance for the application.  Perhaps not as 'idiot proof' as today's idiots are capable of.

Then what happened is (as I described earlier) in genuine human form...society built a 'better idiot'.

Edited August 21, 2024 by kksnowbear

[LucShep]

Thread derailed somewhat (again!) but hey

 

1 hour ago, kksnowbear said:

I admittedly haven't done the legwork myself, but it is my assumption that the manufacturers did the math on the current/wattage/heat capacity of the connector.  They don't generally have meetings where they all agree "Let's put this connector on here which isn't capable of handling the power, so we're sure to get sued..."

 

 

Not sure how accustomated you are with electronics on motorcycles, but it's very(!) often the case that they "budget" the voltage regulators and wire gage.
They crap out at about 20.000 kms or so, and that's in a very large number of motorcycles, higher end models inclusively.
And has always been so, even if you can buy aftermarket equivalent parts that are far better (and long lasting reliable), which should have been there right from the start. 

If multimillion budget capable and reknowned manufacturers do them for vehicles and still get away with it (and knowing it's like that, with decades of bad experiences), so do too PC hardware manufacturers - believe it.  It's the triumph of the bean counters....  

Edited August 21, 2024 by LucShep

[kksnowbear]

I absolutely don't dispute the "bean counter" part.  What I'm saying is that (if they're even halfway worth a darn) they don't plan that it results in their being sued.

What I mean is they do it, sure...but they do it so that they squeek by, even if barely.  Otherwise they'd fail as a company.

And it's not even a problem, until you couple that with the "better idiot" which society has produced.

In other words, the stuff isn't built such that it will withstand the lack of expertise that laypeople possess. It fails earlier because the tolerances no longer allow for the sort of things that experienced people are trained to avoid.

In (other) other words lol The more experienced and skilled the user is, the less likely the device is to have to problems. The less careful the user is, the more likely the devices will fail from said mishandling.  These 12VHPWR connectors remind me (a lot) of various connectors I worked with after the military.  Trained maintenance people learned to work with them, while so-called equipment "operators" broke them constantly.

[LucShep]

6 hours ago, kksnowbear said:

I absolutely don't dispute the "bean counter" part.  What I'm saying is that (if they're even halfway worth a darn) they don't plan that it results in their being sued.

What I mean is they do it, sure...but they do it so that they squeek by, even if barely.  Otherwise they'd fail as a company.

And it's not even a problem, until you couple that with the "better idiot" which society has produced.

In other words, the stuff isn't built such that it will withstand the lack of expertise that laypeople possess. It fails earlier because the tolerances no longer allow for the sort of things that experienced people are trained to avoid.

In (other) other words lol The more experienced and skilled the user is, the less likely the device is to have to problems. The less careful the user is, the more likely the devices will fail from said mishandling.  These 12VHPWR connectors remind me (a lot) of various connectors I worked with after the military.  Trained maintenance people learned to work with them, while so-called equipment "operators" broke them constantly.

I understand, but if this business ALSO depends on hobbyists (and increasingly so), then this connector is a resounding failure.
Because it needs to be 100% safe AND idiot proof.  And it is neither.    ....not funny dealing with this after paying $1000 plus.

Perhaps I should leave that sort of opinion for someone who is (I believe to be) one of the few yutuberzz to often be 99,9% correct in whatever PC matters... 
If you're short on time or patience, skip to 13:27 time of video:
 

^^ ....I rest my case.

 

Edited August 21, 2024 by LucShep

[kksnowbear]

Ya sure...I think Der8aur is great as a source, too. Got one of his delidding tools. Used it many times to successfully delid many units.

But then, see, it's his business to find fault with the big guys.  Generates clicks.  Generates viewers. Generates money. He's the little man's champion against the big evil tech monsters...

Not saying he's wrong but, again...lots and lots of people who don't have problems. Can't just dismiss that, and if I'm Nvidia, that's among the first things I'm bringing up in court.

(EDIT: PS If I understood what he said, I also strongly disagree that it doesn't matter which GPU you use.  There are absolutely cards that are more and less likely to have problems with this connector, and they will also suffer more or less damage if something goes wrong.  There are a lot of factors involved, and we're already pretty far off course here, so I think I'll leave it at that.

Hey, I just noticed...did you know that (at 0:28) he actually says he doesn't think it's a good idea to use his WIreView device?  Crap...now I gotta watch the whole thing lol

Edited August 21, 2024 by kksnowbear

[SharpeXB]

1 hour ago, LucShep said:

Perhaps I should leave that sort of opinion for someone who is (I believe to be) one of the few yutuberzz to often be 99,9% correct in whatever PC matters... 
Conclusion in 14:43 time of video, for those with little time or patience:

Amazing what happens when he just puts a little bend on that cable that appears to be fully connected

Yeah it’s the freakin size of the 4090 too that just defies getting the cable connected properly. He’s an expert too, and it still happened to him. I’m going to speculate that’s exactly what happened when my machine was being worked on. Boom

[kksnowbear]

"He’s an expert too, and it still happened to him."

He's an expert who is creating a video that is explicitly intended to demonstrate failures. 

Generates clicks.  Generates views.  Generates money.

A failure like that is so easy to replicate it's not even funny (if I wanted to make a video).  Among other things, I've been re-pinning cables like that for a long time, and with the tools I use, I can make the connector do pretty much whatever I want.  (The wife and I both have military and civilian experience building aircraft cables which have to pass FAA certifications, too - so this is not an uninformed opinion).

He also states that "This is a Corsair (Type 4?) cable that I've been using for quite a while..."

So, the guy frequently does videos explicitly demonstrating failures...yeah, I'm gonna guess the hardware he's 'been using for quite a while' has been torqued on quite a bit, including intentional mishandling (as the video shows/to demonstrate failures) and so on.

Naturally, this isn't recommended.  I do plenty of mean and ugly sh*t to equipment I own to test things at times.  I don't recommend others do it, and equipment that's been subjected to "failure analysis" isn't made available to the clients.

I've had my 4090 in and out of three different platforms (probably more, kinda wasn't keeping count), a number of times each including both in and out of a case...no black screens, no funny business, no melting.

Again, I'm not saying the guy's wrong.  I'm saying perspective and context are critical, and not everything is what an (uninformed) viewing might lead someone to believe.

Edited August 21, 2024 by kksnowbear

[SharpeXB]

4 hours ago, kksnowbear said:

He's an expert who is creating a video that is explicitly intended to demonstrate failures.

If you watch the video he describes how it happened to him during a review and even he took a long time to figure it out. He said that was part of why he made this video. And all he needs to do to demonstrate the problem is just touch the cable. I think the new spec on that 12V-2x6 design mentions a 45N (10lb) locked connector retention. That’s not 10lb demonstrated there. 

Edited August 21, 2024 by SharpeXB

[nephilimborn]

15 hours ago, kksnowbear said:

Hi again @nephilimborn I am grateful you shared the pics and are also willing to invest the time in actually understanding what could well be going on.

Yeah, that bend seems a little tight for me; maybe not terrible but as you said, not within the guidance (thanks BTW for the correct figures, lol I was too lazy to go find them).  I am very relieved you actually understand the 'root cause' factor at issue here.  It always makes things so much easier to work with someone who's actually capable in that regard.

I was concerned about it because it is among the issues that cause the connector failures.  I do sincerely appreciate your acknowledgement in this respect.  I am flatly amazed at the number of people who will file lawsuits etc, rather than take responsibility for their own/their builder's work.  As I've said (and as Nvidia will surely say, if pressed in a lawsuit) there would have to be a reason that so many people have used the exact same connector and *don't* have problems (and if I understood correctly, we're talking into 6 figures).

Like I said I wouldn't really have any reason to blame my mechanic or Toyota because my wife rides the brakes (and believe me, those maintenance guys appreciate that I understand what's taking place! LOL)

Someone reading this might ask why the distinction matters so much to people like me.  If I can take the liberty, the reason is that if one assumes it's just a 'bad cable' and simply replaces the cable in the same fashion...then given the same heat, time and mechanical stress...(as I bet you already know) the replacement cable will also fail, possibly quicker than the one that was replaced.  The 'repair' didn't really correct the problem, because the user failed to recognize the cable issue was a symptom of the problem, rather than the problem itself.

That's why the 'root cause' matters so much (for the benefit of the reader who actually wants to learn) - and I can't tell you how glad I am that you seem to actually understand this.

BTW the example I just made actually assume the cable is even failed...the problem could easily be that mechanical strain is causing poor mating of the contacts - meaning the cable isn't even bad at all, just needs to have the bends remedied.  I'd honestly prefer to think that would be the case, as opposed to that nice MSI PSU actually having a "bad" cable.  Again - actually understanding 'root cause' matters.   A lot.

Hopefully the stuff you've ordered will help overcome the root cause and thus correct your issues once and for all.  I'm glad if my input helped.

Now if you'll excuse me, my mechanic is calling...says the wife's car needs brakes (again? lol)

 

Received the thermal grizzly last night & installed it while also rerouting the pcie5 cable (relieving torsional tension, etc.). @kksnowbear I also  am hoping that my previous obliviousness did not permanently damage the PCIE cable.

@LucShep I rewatched the first video you recommended and realised that i was an idiot when i first attempted the undervolt with GPU Tweak III. I have since invested the time this deserved and was able to successfully tune the VF curve to .970volts @ 2700mhz (+150). This keep the max (timespy extreme benchmark) TDP to ~415watts (a BIG reduction in both power and heat).

Last night i ran 15 different benchmarks with two tools without a black screen event. I may eat my words later, but sofar so good!!!!

Going to put in some heavy gaming session tonight in DCS and the UEVR injected and see what happens.

Thank you BOTH!!!!

 

Edited August 22, 2024 by nephilimborn

[LucShep]

7 hours ago, nephilimborn said:

Received the thermal grizzly last night & installed it while also rerouting the pcie5 cable (relieving torsional tension, etc.). @kksnowbear I also  am hoping that my previous obliviousness did not permanently damage the PCIE cable.

@LucShep I rewatched the first video you recommended and realised that i was an idiot when i first attempted the undervolt with GPU Tweak III. I have since invested the time this deserved and was able to successfully tune the VF curve to .970volts @ 2700mhz (+150). This keep the max (timespy extreme benchmark) TDP to ~415watts (a BIG reduction in both power and heat).

Last night i ran 15 different benchmarks with two tools without a black screen event. I may eat my words later, but sofar so good!!!!

Going to put in some heavy gaming session tonight in DCS and the UEVR injected and see what happens.

Thank you BOTH!!!!

 

 

 Nice numbers there!
  "Lets goooooo, get to da chopaaa" 

[nephilimborn]

Three 4 hour DCS sessions over the last 3 days and no black screen events 

came across this vid re horizontal mount PCB stress; no idea how much of an issue this is over time ... i know a year ago people were having weird stability issues with PCIE riser solutions

; 

[SharpeXB]

2 hours ago, nephilimborn said:

Three 4 hour DCS sessions over the last 3 days and no black screen events 

came across this vid re horizontal mount PCB stress; no idea how much of an issue this is over time ... i know a year ago people were having weird stability issues with PCIE riser solutions

; 

Surprising. I wonder if people are turning up the tension on the support too much. It seems like it should only need enough to keep it in place, not so much that it bends the card upwards. 

Edited August 25, 2024 by SharpeXB

[LucShep]

 

4 hours ago, nephilimborn said:

Three 4 hour DCS sessions over the last 3 days and no black screen events 

came across this vid re horizontal mount PCB stress; no idea how much of an issue this is over time ... i know a year ago people were having weird stability issues with PCIE riser solutions

;

 Good to know that the problems seem to have been surpassed. 

Yes, with bigger and heavier GPUs, that is an increasing problem.

The vertical mount with a good respective kit of bracket and ribbon can be a solution, instead of having it in the traditional horizontal position (directly on the motherboard). 
No more sag causing pressure and bowing on the GPU's PCB, or pressure on its PCIe connector.
That is, so long that it is one placed with distance from the side cover of the PC case, to keep good ventilation. 
Some vertical mounting solutions put the GPU too close to the side cover, and then it overheats because the GPU fans don't have enough room to feed air in properly.

The Phanteks one he recommends is good, but not sure I'd recommend it for "fat" GPUs like the new 3-slot(+) high-end models (RTX4080/4090, RX7900XT/XTX, etc). 
I really like the Coolermaster kit V3:  https://www.coolermaster.com/en-global/products/vertical-gpu-holder-kit-v3/
LianLi one is also good, and there are others. But it's really a matter of selecting the right kit for each case.

But... the vertical mount of GPU is not always possible or prefered - it depends on invididual case!
That's why GPU brace supports, or sag holder brackets, are still being used so much. With bigger and heavier GPUs, also due to misuse, bowing still becomes a problem.

If it's for systems with the traditional horizontal position of GPU, I use the simplest telescopic GPU supports available (pretty much same as these).
And also use it in my own personal gaming system (it's been years now, no issues - "good for you" like he says! LOL).

 There is a catch/method though, precisely to avoid bowing - that should be placed at the middle of the GPU, not at the very end of the GPU like you'll see most often done.
Also, its height adjustment (with the screw/unscrew for up/down) should be made as to have the GPU exactly level (to the PCIe connector & slot), never overdone, to not create any contrary effect and avoid that bowing he described. 

 

Edited August 25, 2024 by LucShep

[nephilimborn]

A retrospective to close out this thread...and a big thanks to both @kksnowbear & @LucShep for your experienced help & insight in resolving an issue that was plaguing me for months!

It has now been 1.5months and no black screen, 100% fans events. The resolution included the following actions

• Root cause: Rerouting PCIE 12VHPWR cable, eliminating the 15mm 180degree bend with the introduction of the thermal grizzly wireview
• Mitigating factor: High power draw exacerbated the above cable issue, undervolting advice & guidance mitigated the occurrence however has long term value for me in terms of TDP & PSU capacity (at load i am near 80% with a 1000w PSU). This is esp useful for me now as i have recently upgraded from a Q3 to PCL and the GPU demand is significantly higher for significantly longer
• Other: Eliminating GPU slot sag with additional supports

This forum thread helped me see the forest thru the trees; i was too deep into the issue and clouded by frustration and pc hardware inexperience. 

[XBlackstoneX]

I suspect the PSU.

[BitMaster]

FWIW, since I had terrible sagging and temp issues with my former 980GTX I have moved to vertical mounting and never had such problems again. OK, it's ~75€ expense for the kit but well worth it, trouble free.

 

...and it looks great