Calling all Gurus, I'm in desperate help here, please come in and help regarding GPU issues.

[TonyTurbo78]

Few more people having same issue.could well be that fact mem is hotting up.try putting a fan direct in front, over them, and see if it fixs it?

http://forums.guru3d.com/showthread.php?t=399708

Zia

tried that fan last night thinking exactly the same thing. Still got the black screens then, memory temps and VRM temps were pretty much below 50c so well below what I would call overheating.

Memory temps would be no different using dp, HDMI or dvi if running same fps.

 

[zia]

how about try upping the volts in afterburner while in game see if that helps?
logical speaking if all you've done is take stock cooler off and then put water block on the gpu should be exactly same as before.unless there is an issue with fitment of the block.mem heatsinks.
Got spare gpu you can try with the benq moniter?
You could try put stock cooler back on and see if it cures it? shame there is no one around you who could try the moniter just cancel it out of the equation.


Zia

 

[TonyTurbo78]

yep everything just went on lovely, no leaks, no slips, no issues, just simply putting on a water block, doing everything on anti static straps etc there should be no issue. The annoying thing is, for both cards to do exactly the same thing, I find that to peculiar.

I will speak to the guys here on Monday see what they advise. I need DP to be working though, that's the one thing I do need for 144hz and 1440 Freesync. I'm just totally wounded here right now. I just dont know what to do now regarding what path to take.

 

[muziqaz]

I don't really know how Freesync actually enables itself, but I assume you need to disable v-sync in setting in order to enable freesync? Am I right? Maybe try playing around with AMD CCC?

How about GPU Z rendering screen? Have you tried that one on iGPU and your cards?

 

[humbug]

Shot in the dark.... are the blocks over tightened?

 

[TonyTurbo78]

Only finger tight, not loose but not wrenched up so hard to strip the threads. Would that cause an issue with dp as apposed to dvi or HDMI just ny tightening to much

 

[humbug]

Only finger tight, not loose but not wrenched up so hard to strip the threads. Would that cause an issue with dp as apposed to dvi or HDMI just ny tightening to much

I don't know, its a shot in the dark, i have seen blocks over tightend and short with the micro cap thingies around the GPU.

 

[Ash0787]

Reading the description my first thought was that you might be overloading the power supply ( extra wattage from the pump or something ) but you said that you had made sure it was fine

I'm not really sure what to say in that case, the system looks well built so I assume its not your first time doing hardline watercooling, it can be quite a physical task to get everything connected and an amateur could easily cause damage to a card,
that would be my second guess, but as I said it looks well built

If the hardware hasnt changed and the monitor hasnt changed then I cant see what could be causing the problem

You say you dont get the 'driver crashed' message ? could try looking for other information using GPU-Z or the like, might give some clues as to what is happening
in those intervals when the screen is black

Try plugging the pc into a hdmi TV if you dont have another monitor

 

[MrMD]

Gonna throw a random thing out there,almost 100% it wont make a difference but at this point it cant hurt really.

If your mobo has the option in the bios set the PCI slots to gen 2.0 rather than 3.0? See what happens?

 

[TonyTurbo78]

Reading the description my first thought was that you might be overloading the power supply ( extra wattage from the pump or something ) but you said that you had made sure it was fine

I'm not really sure what to say in that case, the system looks well built so I assume its not your first time doing hardline watercooling, it can be quite a physical task to get everything connected and an amateur could easily cause damage to a card,
that would be my second guess, but as I said it looks well built

If the hardware hasnt changed and the monitor hasnt changed then I cant see what could be causing the problem

You say you dont get the 'driver crashed' message ? could try looking for other information using GPU-Z or the like, might give some clues as to what is happening
in those intervals when the screen is black

Try plugging the pc into a hdmi TV if you dont have another monitor

PSU is a 1300w Superflower Ash, got a meter on the plug and it's not even drawing 800w under load.

It is my first time building with hard tubing, but coming from a Precision Engineering background (13 years working in the aviation/optical industry working to thousands of an inch tolerances) I'm very particular with measurements, triple checking things and taking my time to be precise. The fits were perfect, no leaks, over effort needed to connect bits.

No driver crashed message, as far as I can tell whatever is running in the background continues to run because I can see that the clocks, usage graphs in MSI Afterburner do not dip at the times of the black screens.

And regarding the HDMI in the TV, have been doing that.

 

[docsonic]

Just to rule it out - does your monitor have a factory reset option? If so - worth a try?

 

[TonyTurbo78]

Gonna throw a random thing out there,almost 100% it wont make a difference but at this point it cant hurt really.

If your mobo has the option in the bios set the PCI slots to gen 2.0 rather than 3.0? See what happens?

ok will give that a go when I get home. Needed a break, been on it solid for 24 hours, albeit a few hours sleep in between and im not to far from throwing stuff out the window and calling it a £700 write off.

Just to rule it out - does your monitor have a factory reset option? If so - worth a try?

Possibly, will have a look later.

If all this fails, what would you say is the best thing to do. Drain the loop, take out the cards, remove the blocks and passive heatsinks then put the air blowers back on and re-try? I really want to rule out the monitor once and for all first before I do any of that. Only way I can see me doing it now, is to go and buy a 1440 freesync screen, take it home and test then returning it to the shop with some sort of excuse.

 

[ketma]

I think I have a similar issue as you....black/blank screen occurring at random intervals using the DP cable. I've ruled out the OS as it happens at the BIOS level, tried 2x DP cables. Screen never loses sync just comes back as is nothing happened....tested this over HDMI too, same thing occurs.

 

[1kiwi]

If all this fails, what would you say is the best thing to do. Drain the loop, take out the cards, remove the blocks and passive heatsinks then put the air blowers back on and re-try? I really want to rule out the monitor once and for all first before I do any of that. Only way I can see me doing it now, is to go and buy a 1440 freesync screen, take it home and test then returning it to the shop with some sort of excuse.

Unless you have damaged the card during the H2O setup it wouldn't make sense that the issue would go away after putting the air cooler on. Despite the H2O setup is there anything else you have changed?

Have you tried another monitor?

 

[TonyTurbo78]

I dont kiwi, would be so much easier If I had one at hand to rule out the monitor before I go tearing things apart. I've already drained the loop this morning and messed around, I don't want to do it again for the sake of it. I don't have any pc gaming friends down here with high end rigs. If I did then would be easy to rule out the screen.

There's no way in hell I've damaged both cards exactly the same. That's just not happening. It's not like I'm some clueless bull in a China shop when it comes to upgrading and doing pc upgrades.

I'm just really lost at the moment. I need another freesync screen to try dp on with things how they are first, that's the tricky bit.

Only other thing I've changed is the psu. But I tried my old one this morning just to power 1 gpu whilst running the rest of the system off the new psu, still got the black screens.

 

[HazardO]

Tony, a few thoughts..

You say you have 2 DP cables you've tried. Are they the ones which came with the GPUs or is one of them the one which came with the monitor? If you haven't tried the one which came with the monitor, I would recommend trying.

Is the PSU a single rail job or multiple rail? If it's multiple rail, you could try moving one GPU to different output ports on the PSU. It's a powerful PSU, but if it's multiple rail you may be overloading one rail if everything is stacked on that one rail.

My last thought (a long shot) is that it's a grounding issue maybe caused by static build up as the water flows through the blocks, or the blocks are shorting a component of the GPUs power delivery which is only an issue when the card tries to ramp up to full power.

 

[TonyTurbo78]

ok guys big update.

Drained the loop this morning, removed the waterblock from one card and put the original msi 2 fan cooler shroud back on, ran some tests, games, no black screens.

So monitor is fine, card is fine, must be something to do with the blocks right?

Obviously the heatsink chokes sit on the VRM's and memory chips with the original cooler, could passive cooling not be up to the trick when I have the block on? I did look at VRM temps when it was on water, they were only 60c, funnily enough the same temps now I have the cooler back on so I can't see that being an issue, certainly not after a few seconds. chips are fine up to about 100c right??

 

[Gregster]

Sorry to hear TT and a nightmare bud. I was having some similar issues with the 290X and the Fury X with DP as well. If I set 60Hz on the screen, it was fine but then when I fired up a game, it would black screen pretty much right away, 2 new leads and a no show but switching to Nvidia and 144Hz was fine, so I assumed it was just an AMD thing and my 144Hz monitor but being determined not to give up, I bought another lead from Ugreen and like magic, no more problems (very occasional flicker or black screen).

In short, I am pretty sure it will be your lead.

Edit:

Scrap that, as I see you have it sorted with the cooler back on.

Having water cooled many cards, I can't see anything you would have done wrong for the issue to only happen when gaming. Does the block not cool the VRMs as well?

 

[TonyTurbo78]

no greg, there isnt a full waterblock available for the 390msi. Even passively cooled, the VRM's and memory chips were still very cool, exactly the same as now that I've tried with the cooler back on. Last night I did stick a huge honeywell fan in there to see if the temps were causing it, still blacked screen then. I highly doubt looking at the temps that this is the cause.

I'm going to remove the blocks again, reapply the TIM, double check everything and see what happens. Few things from this, at least now I'm sure it's not cable, monitor or GPU borked.

 

[Gregster]

no greg, there isnt a full waterblock available for the 390msi. Even passively cooled, the VRM's and memory chips were still very cool, exactly the same as now that I've tried with the cooler back on. Last night I did stick a huge honeywell fan in there to see if the temps were causing it, still blacked screen then. I highly doubt looking at the temps that this is the cause.

I'm going to remove the blocks again, reapply the TIM, double check everything and see what happens. Few things from this, at least now I'm sure it's not cable, monitor or GPU borked.

Ahhh got ya. You should have a fan blowing over the VRMs but they shouldn't need much to keep cool and I would be surprised if even without any cooling on them that would cause the blackscreen problems. I am sure the VRMs are rated for 120c (or close to that), so that shouldn't be the problem and memory isn't really something that gets hot but maybe being 512 bit, it could need that fan.