SOLVED praying it aint so......had a leak and now pci-e issues it seems

Hey guys and girls, I need some input from anyone who has had similar issues and greatly appreciate it. Bare with me please.
Heres some background- A cpl months ago noticed my gpu temps getting way up there 95+C, so thinking they are a cpl years old I'd swap the TIM and do some cooling mods to bring the temps down, which I added Kraken G-10's and H55's and alphacool M04 backplates to my r9 280x's. My cpu has an H80i in push pull and had a res and tube mod added which was done a cpl years ago.
So, after adding the kraken and h55 setups I get my rig back together, fire it up and when I added the second card it BSOD with a thread_stuck_in_device driver when I was trying to enable crossfire in CCC.
So being as simple as I am I figure its a driver issue, I do the DDU, then install the same drivers I was using before I removed the cards,15.7, set up the crossfire and no issues........life is good, temps on my gpu's never go above 60C and and I'm thinking I'm a happy man.
Well fast forward 2 weeks, and have been gaming away every night with no issues when I go to load WOT and as soon as the battle starts- BSOD- same thread_stuck_in Device. Figure huh thats weird, try the same process I used before to get them running and nope not this time- every time it tried to boot-BSOD with same warning. Investigating, I remove the lower card and pc boots and runs fine with just the top card in. So now I'm thinking its related to that card. Freaking out and paranoid I did something to the card when installing the cooling (the pcb had quite a curve in it from hanging in my case for 2 years and didnt like to be straightened out and bolted to the backplate was my thought) I remove the Kraken and H55 and put back to stock, plug it back in and boots fine....CCC sees the card and I set up crossfire, no issues, figure I should test and run furmark for an hour no issue.....thinking it was something I did and its working its time to start killing again...fire up my games and I'm a killing machine for like 6 hrs..no issues, played BF4, titanfall, WOT, life is good again....then it happens...BSOD...same error. I'm like WTF!!

So by now your wondering why did I post here? well to my horror when I removed the lower card this time to investigate my eye catches a glimmer behind the card. Grabbing a flashlight, I see a small trickle of coolant running down my board and right to the end of the #3pci-e slot which is where my lower card was.
So tracing the water up I can see where its leaking out the back of the fitting from the pump, runs down, flows around the lower cpu cooler bracket, then down the board to the #3 slot, then off the end of that down the board again. So, remove the H80i and leaking pump, both GPU's and did the hairdryer thing for awhile and dried it and left it open in my case for a few days. couldnt have been a bad leak as it barely changed my water level in the res and I know how stupid that sounds as any leak is a bad leak.
So heres where I'm at now, it boots and runs fine with only the top card in, as soon as i install in slot 3-BSOD with same error, I take the lower card and switch it with the top card and boots and runs fine. Whenever I try to install a gpu in slot 3 it BSOD's either on boot or when I try to load a game. So wanting to make sure its not a card I borrow my friends r9 280x and try it in top slot, boots fine runs fine, move the card to slot 3 and it BSOD's every time when trying to boot but this time gives me a system_service_exemption_ Atikmdag. error.
I'm thinking its def the board but hoping its not anyone have any other ideas? Thanks a ton to those that reply.

Most likely it's the board, I think. You could try cleaning the slot with contact cleaner or similar get an old card if you have it and put it in and out a few times . Just try and eliminate that it might be a bad spot on one of the fingers . May possibly clean it and get better contact. Since it was going on and off for a while it could just be corrosion/build up. You don't see anything burnt in the vicinity do you?

Johan45 said:

Most likely it's the board, I think. You could try cleaning the slot with contact cleaner or similar get an old card if you have it and put it in and out a few times . Just try and eliminate that it might be a bad spot on one of the fingers . May possibly clean it and get better contact. Since it was going on and off for a while it could just be corrosion/build up. You don't see anything burnt in the vicinity do you?

Click to expand...

Thanks Johan, no I didnt see any visable burns but with the board being black its hard to be certain, I did recieve a voltage warning before it went BSOD sunday, tried cleaning the connection, and have a card in and out quite a few times since then trying different things.
I was using catipillar anti corrosive anti freeze pre mix.....was what I had handy back when I cut the H80I apart. I'm thinking it was leaking since the time I did the Kraken install so for like 2-3weeks, and the board may have soaked up the coolant although I dont see any signs of it being wet looking or feeling and not getting anything blotting.Cleaned the board real well with alchohol, to remove any of the residue left by the coolant.

Hoping I didnt make a square frisbee

I wouls go so far as to try some like 2000grit paper. If nothing is burnt outside it could be burnt in the connector or eve burnt the finger off. Can you get some light in there with a magnifying glass and have a look.

K....didnt have 2000 but did have 1500, wrapped it over a thin file, cleaned all the contacts, looked at with my scope as best I could and nothing seems burnt in the connector itself, they all look decent not bent or burned and shiny.....booted with one card fine...shut down and did boot with second card but then BSOD with thread_stuck_in_device_driver error when I tried to enable crossfire in CCC...
Heres a pic, I see i missed a lil of the residue near the end of the #3 pci-e slot.

OK so the PC starts with both cards in,correct? It also recognizes both cards since you were able to try to enable CF. This could just be a messed up driver issue because of the crash. Have you tried, since cleaning everything to remove the old drivers , clean with ddu AND REINSTALL. i WOULD DO IT WITH ONLY ONE CARD THEN ADD THE SECOND AFTER THE it has restarted and initialized the driver. Reboot install second card.

Johan45 said:

OK so the PC starts with both cards in,correct? It also recognizes both cards since you were able to try to enable CF. This could just be a messed up driver issue because of the crash. Have you tried, since cleaning everything to remove the old drivers , clean with ddu AND REINSTALL. i WOULD DO IT WITH ONLY ONE CARD THEN ADD THE SECOND AFTER THE it has restarted and initialized the driver. Reboot install second card.

Click to expand...

alas...yep have tried a few times.....even changing drivers to see if it made a difference....no change....and it did boot that time with both cards but thats random...sometimes does and sometimes doesnt....also...when I try my friends 280x it will boot with it fine in slot 1 but when moved to slot 3 BSOD everytime on boot but with the system_service_exemption_ Atikmdag. error

with my 280x I have only gotton the system stuck in driver error and then when moved to slot 1 both my secondary 280xBE and my friends 280x boot and work fine.
My first thought was drivers, but let me ask, when I upgraded to win 10 I didnt do a clean install yet....would that be complicating or have anything to do with it?

When I searched that error win8/10 came up a lot. Usually just a driver error. If you have a spare drive it may be worth downloading and installing it fresh and see if it makes a difference. Do you have a Win10 friendly BIOS?

thought the same.....I did alot of searching before I came here and asked...Ive also installed the amd drivers not using CCC and it does the same thing......I have made sure I updated my bios and chipset drivers hoping it made a difference but no...thats why I'm looking at the Pci-e as the culprit or something near it thats messing my card......sorta weird I get 2 different errors with the 2 cards in slot 3 and then either card in slot 1 works fine..no errors or boot issues everything seems related to slot 3...

Edit- also knocked all my clocks back to stock to make sure that didnt have anything contributing

Have you tried the card in the last slot? I think it's x4 IIRC but it may help determine if it's a driver thing or not. Or have you tried in only the 3rd slot?

Johan45 said:

Have you tried the card in the last slot? I think it's x4 IIRC but it may help determine if it's a driver thing or not. Or have you tried in only the 3rd slot?

Click to expand...

dude you read my mind....after i got looking a lil bit ago I thought about that and tried when I cleaned the connector as suggested theres too many connections...usb, audio, ect and I cant really seat the card all the way in with the cables there in the way, it bends them way over and pinches against them.
In device manager I dont see any issues even after a failed boot or BSOD if that makes any difference

so I have tried it in 1st and 3rd slots.....either of my 280xbe's work in slot 1 but both do the same BSOD in slot 3 and both same thread_stuck_in Device error.
my friends 280x boots and works in slot 1 but BSOD everytime in slot 3 with that system_service_exemption_ Atikmdag. error

Edit- with either of my GPUs in slot 3 the BSOD is random, mostly BSOD while booting but sometimes will boot through and recognize the GPU but when i try to enable it BSOD's or if I can get through the crossfire, it only lasts a short time and when i launch a game will BSOD.

You have a bad slot.

Mr.Scott said:

You have a bad slot.

Click to expand...

Damn Scotty.....just kick me in the b***s bro, I was afraid of that, but appreciate your advice, I was going to do a clean install win 10 and see if it made any difference. Think its worth my while or just bite the bullet and swap the mobo then clean install after?
Let me ask if I swap my mobo out like for like will I have to install win 10 again too?

Don't know enough about W10 to tell you for sure, but I would guess a like for like swap shouldn't be an issue.
RMA that board?

If like for like swaps do = reinstalls of windows, or at least reactivations.....MS is going to have the largest flood of complaints they have had since Vista.

Mr.Scott said:

Don't know enough about W10 to tell you for sure, but I would guess a like for like swap shouldn't be an issue.
RMA that board?

Click to expand...

The board I got second hand and cheap and was awhile ago so its been a good run either way if it was the mobo I'd just grab another I''ve been so happy with this one ....as an update- I did a clean install of win 10 last night to see if it made any difference. so far no bsods but only back up for a few hours now.
I actually think it still may have been damp when I was messing with it......seems the longer it sets the better it gets...so maybe it was just a water related thing and it didn't bang that pci-e i''m hoping. I'm gonna run some stress tests on the cards ad see what happens.

Good news oldiron, personally i'd lean more toward win10 upgrade issues. From what I have read it doesn't always go well and strange crash issues afterward.

Johan45 said:

Good news oldiron, personally i'd lean more toward win10 upgrade issues. From what I have read it doesn't always go well and strange crash issues afterward.

Click to expand...

Oh man I was hoping....but just had my first bsod when running furmark . Slightly different though now, with it will let me enable and got through 17 mins so at least it wasn't right away.
Gonna grab my multi and check voltage on those leads for shets and giggles.
Upside, the clean install fixed some other odd things and I have an ISO of my system now so if it continues and I swap it will be easy.

Don't use furmark for testing oldiron. It's really not good for the cards. Try running unigine heaven DX11 benchmark, it's plenty hard enough on the GPU to test for stability.

Johan45 said:

Don't use furmark for testing oldiron. It's really not good for the cards. Try running unigine heaven DX11 benchmark, it's plenty hard enough on the GPU to test for stability.

Click to expand...

Had no idea , thanks man