lrighty so finally split my card tonight. My Bykski Block arrived today so it was time.

I havn't fitted the block, i just wanted to split the card, clean it up and put it back together and run it hard for a few days (while i prep the rest of the build).

Now my Method of splitting the card was this and i feel there is absolutely zero was any "death" of the card can be contributed to PCB bend or anything like that.

1: I removed the backplate and sat it aside, leaving only the GPU Leaf spring and rear IO inplace holding the thing together.
2: Reinstalled the card and ran it under a heavy benchmark at full power, with a Ethereum miner running hard on the memory whilst also forcibly setting the fans to 50-60% while keeping an eye on temps to hold the memory around 98-100 Degree's and the core around 80 Degree's.

This had an effect of heating the card and heatsoaking the crap out of the heatsink (to retain heat during splitting). I did this for a few minutes, around 55% held those temps pretty stable
3: For the final 30 seconds or so i actually disabled the fans and let the heat spike then immediately shut down the system.
4: Removed the card and sat it on an anti-static surface. At this point the Heatsink was heatsoaked bad and was burning hot to the touch (being an auto sparky/mechanic by trade im used to touching hot things but yeah, this was damn hot).
5: Removed the Rear IO and the GPU leaf spring (doing this by releasing tension on each screw in a criss-cross pattern half a turn at a time so the heatsink evenly released pressure off the GPU die)
6: Pulled all cables/fan plugs/etc
7: Split the card by sitting the card flat on the work surface and just barely twisting the HSF in a clockwise/counter clockwise motion, but only slightly/enough to break contact but not hit any components on the board (not up and down, or side to side pulling away from the card, but a twisting motion still in contact with the card). Once it was easy to "twist", i just lifted the HSF straight off, zero card bend or anything.

The thermal putty was actually more like a sticky liquid goo from all the pre-heating, i feel it would have just slid straight off without "twisting".

From there, i "piled" up the putty to reuse it (i wasn't ready to clean it off just yet and replace with pads, though majority of it ended up on the bottom of the HSF rather than the VRM's so it was quite simple to reshape), cleaned all the dust off, cleaned out the fans, reseated all the thermal pads, repasted the GPU and put it all back together. (again, placing the HSF on to the card, then flipping it and screwing the lot together in the reverse of how it was taken apart).

Card is currently running under the same bench conditions as it was pre-disassembly without missing a beat. I'll let it run for a few days (might mine it too just to keep the load on).

Then its time to block it and fingers crossed, everything goes well! 

Heres a couple pics of my board right after i split it. You can see pretty much all the Pads just popped off and the putty is visibly "gooey"

Theres a couple close-ups of the BykSki block, incase anyone wants to compare it to the EK blocks when it comes to stand off size/positioning/etc. Btw i actually really like the simplicity of the Bykski block, it feels solid and heavy, pretty good machining all round and ive seen pictures, looks like they light up really well with the inbuild ARGB strip (lighting the blank acrylic up nicely too). Keen to see how it goes! At about $150 AUD cheaper than the EK & Backplate, its a steal if it performs well enough

  


 
   

cos702003
I know this is an old ticket, but I just installed my block on a 3090 FTW and had a similar, but more successful outcome, than others here.
The card (on air), posted without any issues, except to reboot the machine after first boot up to recognize the card).
I then swapped out to the EKWB Quantum vector board for 3080/3090, replacing the ram thermal pads with identical thickness variants from Grizzly. Upon booting the board, My bios stalled and restarted the machine 3 or 4 times before the card would output a video signal.
Eventually, it posted and got stuck on "AE" bios code, then later the "b2" code others here reported seeing.
My system started to hang, to the point where the mb would not even post, or if it did manage to post, would hang during windows boot up. It was so bad that the bios screen was loading chunk by 8-bit chunk to display the page in entirety, and took almosta full minute to do so.
My issue turned out to be the Riser card I installed at the same time. The card (By Fractal for the definer7) is PCI-E 3.0, while my mb and gpu are 4.0 compliant. After switching the PCI-E gpu slot to 3.0 native, from auto, all issues cleared right up and I got no more hangs trying to detect legacy drives, or on the "b2" code. Bios response time went back to normal and I have not had a single issue since.
 
I should note that the first time I managed to get the card to display video, it took a good 4 or 5 minutes of leaving it on running until the "ae" and "b2" codes cleared themselves. Have patience.
Hopefully this can save at least one card from an RMA. I know it saved mine.

If buying a 3090, I'd choose to purchase the Link-up PCI-E 4.0 riser rather than hamstring the card to 3.0. I have the Link-Up and have had no issues since receiving my RMA card (though on Optimus Block now).
 
But for people testing, confirm riser works while still on air. Air/Watercooled should not effect if riser will work or not.

Anyone tried using a good thermal putty over trying to get the thermal pad thickness correct? 
Something like this stuff (which seems to be recommended a lot)

Really pretty equal to buying decent pads (here in australia anyway), save a lot of mucking about trying to get the correct pad sizing for the Bykski block too (and could probably fill the 5-6mm gap on the chokes to stop coil whine/heat build up)

I read through mostly half of this thread. Wondering when the updated standoffs were released? I bought a new waterblock from ek last May (25th). I measured the diameters of the specific standoffs and they where 5.6-5.9mm. The new standoffs are 4mm?

Ineedgfx

cbattung1016
I read through mostly half of this thread. Wondering when the updated standoffs were released? I bought a new waterblock from ek last May (25th). I measured the diameters of the specific standoffs and they where 5.6-5.9mm. The new standoffs are 4mm?

Around June it seems and the easiest way to know is the new standoffs are stepped in height vs the original ones are not.

Looks like mines are not stepped. I emailed EK and requested the standoffs. 
 
Are there any success stories with using the updated standoffs?

Looks like the FTW cards bricking is the fan controller frying.
Igor's lab published this subsequent article about their New World findings about the FTW 3090 paper weight.
Excessive heat apparently fries some fan controllers.

https://www.igorslab.de/e...se-of-design-failures/

Find fault in your card as soon as to can and RMA.

Think it easier to have the computer shop install it.

Just got my 3090 ftw3 from EVGA today. The two holes (the ones closer to each other) where the two (out of three) new standoffs are suppose to go, does not have anything screwed on there. I’m thinking about removing those standoffs and just be done with it. The 3rd one (farthest one towards the fan pins/sockets) does have a screw through it. 


https://imgur.com/a/wa1d1Io

cbattung1016

sdmf74
Hey guys I've got a brand new EVGA RTX3080 FTW3 arriving today along with an EK waterblock and backplate.
I just stumbled upon this thread and so far have only read the first 30 pages.
Since I plan on installing my new waterblock on the GPU today I was hoping you could brief me on the current issue as I have not yet had time to read the last 40 pages of this thread (although I will).

What is the deal with the standoffs? I just briefly saw it mentioned on the Last 2 pages. Are you referring to the standoffs on the EK waterblock? Please elaborate, and shou weld I be concerned or is there anything I should know before going forward with the install of my 3080FTW3 and EK waterblock?

Btw has EVGA or EK found what the problem is or has either one made an official statement, other than blaming the other party or the customer? Thanks!!!

Three standoffs on the water black are too big (diameter 5-6mm) that are potentially making contact with the electronics on the pcb. EK sent out 3 replacement standoffs (4.5mm) that fixed the issue. I have spoken c EVGA since the RMAd my 3090 after it got bricked by the block. Rep informed me that they are aware of the issue and will continue to RMA the affected GPUs as long as there are no physical damage (including water damage). No official statement released from either companies. Seems like EVGA is doing there part in regards to warranty. While EK sent out a fix which is somewhat telling that they know the issue with the block but not really putting a statement out. BTW, I just got my 3090 FTW3 RMA replacement in the mail today. Initially messaged EVGA on Saturday. Then got a response/acknowledgement of the issue on Sunday. Monday, it was approved to be sent back for RMA. Úú up arrived at EVGA. Wednesday, received by EVGA in their system and RMA approved, then shipped out. 
 
Pretty quick process and pretty much hassle free. I had more questions than they did to me. Just gonna say that I will be a long-time customer just based on this experience. 
 
You could check with EK if your block came in with the updated standoffs. I believe blocks from June-present have them. I got mines early-mid May. 

Thanks for the reply! I just ordered my EK block from Performance PCs like 3 days ago.
Has anyone posted a pic in this thread showing which standoffs have been changed from 5-6mm to 4.5mm?
If not can someone take a pic of the affected standoffs if you have one of these waterblocks laying around please?
I have a digital caliper and I should be able to measure too see if I have a new or old waterblock.

Also since this is such a serious issue don't you think EK would have recalled all the defective waterblocks from retail stores and replaced them with functional ones? Ie ones with smaller standoffs

Has anyone ordered an EK waterblock from PPCs recently and recieved a new one with the updated standoffs?

sdmf74
Has anyone posted a pic in this thread showing which standoffs have been changed from 5-6mm to 4.5mm???
If not can someone take a pic of the affected standoffs if you have one of these waterblocks laying around please?
I have a digital caliper and I should be able to measure too see if I have a new or old waterblock.

Also since this is such a serious issue don't you think EK would have recalled all the defective waterblocks from retail stores and replaced them with functional ones? Ie ones with smaller standoffs

Has anyone ordered an EK waterblock from PPCs recently and recieved a new one with the updated standoffs?

Post #640 has the best info, straight from EK.

Ineedgfx
So my revised standoffs arrived yesterday so I figured I would notate the changes.
 
The original standoffs are .236"/6mm in diameter and the revised standoffs are .177"/4.5mm in diameter.
 
 

Thanks for providing the info but im slightly confused. I just opened my new waterblock from PPC's & measured the standoffs. First of all you guys mentioned "stepped" standoffs
and mine appear to be stepped with the bottom of the step being square & the top is round. Here are the measurements of the 3 lower standoffs on the waterblock:
 
Bottom Step Square - .215" / 5.47mm
Top Step Round - .178" / 4.52mm

And that is the issue, What part of the step are guys measuring when you say that the new standoffs are 4.5mm??? It doesnt seem to make much sense measuring the top of the step cause the widest part
of the standoff (the bottom) is what is gonna interfere or damage a part on the pcb. My measurements are both close to the size of the old standoffs & almost identical size of the new standoffs
depending on which part of the step you measure lol.
 
EDIT: Here we are 7 months from the start of this debacle and STILL no "Official" statement from EK stating which lot-S/N: are affected & instructions on which & how many standoffs we need to replace
so we dont destroy a $1k or $2k piece of hardware that is in VERY VERY short supply (I F5 refreshed night and day since October 4th 2020 & just finally got 1 notify for purchase!!!) ? What is going on here! Whoever
the rep is is NOT doing his job, this is something that should not be swept under the rug, Did covid just 100% kill professionalism in society? (probably its most dangerous attribute).

sdmf74
Thanks for providing the info but im slightly confused. I just opened my new waterblock from PPC's & measured the standoffs. First of all you guys mentioned "stepped" standoffs
and mine appear to be stepped with the bottom of the step being square & the top is round. Here are the measurements of the 3 lower standoffs on the waterblock:
 
Bottom Step Square - .215" / 5.47mm
Top Step Round - .178" / 4.52mm

And that is the issue, What part of the step are guys measuring when you say that the new standoffs are 4.5mm??? It doesnt seem to make much sense measuring the top of the step cause the widest part
of the standoff (the bottom) is what is gonna interfere or damage a part on the pcb. My measurements are both close to the size of the old standoffs & almost identical size of the new standoffs
depending on which part of the step you measure lol.

Hi, the measured points are at the 'top' of the stand offs where the metal standoff itself comes in contact with the GFX PCB and that's the only dimension that matters in this discussion because being too big was presumably causing a short and the smaller or 'thinner' circle prevents the metal from touching something on the cards and causing a problem
 
Also since your standoffs have to different diameters you for sure have the updated stand offs because the original design were of a single dimension
 
See these posts to show the differences:
https://forums.evga.com/FindPost/3424064
 
https://forums.evga.com/FindPost/3424069
 
p.s. the base or bottom of my revised (shipped) standoffs were hexagonal vs your reporting of square

Subzero30k
Hello i had a question i am about to remove the Stock cooler from my evga 3090 ftw3 Ultra and add the ek quantum vector waterblock

If the card dies is there Sure not a Problem whit RMA the card? And what must i say to evga then?

I was completely honest when I started my RMA process. Told them I installed the EKWB water block and the gpu won’t boot. RMA was ok’d and 4 days later, I receive my GPU in the mail. Rep also told me that if it happens again even with the samr block (with or without the updated standoffs), they will RMA it. They are aware of the issue and will honor the warranty.

Decided to take out the standoffs instead of replacing them with the modified/updated standoffs from EK. Works great! No issues!

Brick two 3090 from original standoffs. Decided to updated standoffs from EK. Going strong for two weeks!!!