Just because there is short causing a no power scenario does not mean taping up the block will do anything. 

Earlier in the thread there was concerns of caps being to close to one of the standoffs per EK. I have put it under the microscope and my caps in that area look fine and also test fine in diode mode.
Insulating the top of the block is not a clear solution to eliminating shorts if like other people have said caps are to close to standoffs and end up getting crushed. A few fets on my board are clearly shorted when testing with a meter. Even with the block off they remain shorted.
All im saying is its impossible to test for the specific shorted area when the block is on(besides the power connectors and pci connector) and taping is not 100% going to solve a short. You cant completely isolated the block from the board without modifying EKs product.
At this point my 3rd card arrived. I could mount the block again and multimeter the power connectors and pci slot to look for shorts before powering on. I am not going to do that. Its up to EK to make sure their blocks work with EVGAs product across the board.

Reading these posts makes me worry. Many people here have been installing waterblocks for years, including myself, and suddenly the cards are just breaking after installing the block? Something must be shorting the card. What else would it be? I highly doubt it that people are breaking the PCB because of some dried paste or whatever, this isn't anything new and many people are aware that cards behave this way because the paste is always garbage (so are the pads, on ANY card) so obviously you take out the PCB slowly. For a card to immediately die after installing a water block (after being fine on air for months) means something is shorting the card. You can't tell me all the people here have broken the PCB by taking the cooler off?? What is this, 2008? Have you guys checked everything after installing the waterblock, for example, placing the card under direct sunlight and checking it from the side to see if there's something touching a component somewhere? Maybe everyone is just being sloppy and breaking their cards? Who knows.
 
My block arrived (and so did the MORA 420) but now I don't even wanna risk it. There's just 0 cards anywhere and who knows how long the RMA will take. I really hope someone finds what's happening here.

The problem is that there seems to be a quite enormous variations in the factory for these new cards. IgorsLab has measured the bending of these cards for example and these are not flat at all. 
This issue can make a lot of trouble, even for experienced users. Without the right equipment you simply don't know how worse your card is and every single card is different in this regard. The cooler might perfectly fit on one card and what will make trouble on another card with the same sku. The manufacturers of the waterblocks can't adjust either as every card is so different and unique... 

garbage-79
The problem is that there seems to be a quite enormous variations in the factory for these new cards. IgorsLab has measured the bending of these cards for example and these are not flat at all. 
This issue can make a lot of trouble, even for experienced users. Without the right equipment you simply don't know how worse your card is and every single card is different in this regard. The cooler might perfectly fit on one card and what will make trouble on another card with the same sku. The manufacturers of the waterblocks can't adjust either as every card is so different and unique... 

The air cooler is able to not fry these 'unique' cards - so this is hardly as complex as you claim.

I've only murdered one card via waterblock installs. I killed it when I forgot to unplug the power cable that was attached to the videocard and did a loop refill which had me constantly turning on and off the system which fried the card. RMA got me another card quick

Here is my latest response from EK.
---------------
There seems that there was some misunderstanding, we did do extensive research on the topic. 
Please understand that we pulled back those "faulty" blocks and tested them and all worked perfectly on our GPUs. The same blocks were then returned and were used by users to mount them on the RMAed gpus and then it worked perfectly, they confirmed that. We do not have a single confirmed case that EK block "killed" a single FTW3 GPU. All the GPUs were RMAed by EVGA and are now working on the same blocks.  That is why we are confident to say that you can mount the same block to the RMAed GPU as we are certain there are no issues with the block or backplate. That was not a single confirmed case but many of them, and there was not one confirmed case that the issue was because of the EK block. 

EK is renowned to admit and fix its mistakes which is evident in the past if it comes to that. 
That is why we would like to offer you to send us your RMAed GPU(when you receive it) the block, and the backplate, and we will mount it for you, do the tests and send it back to you. 
For that, you will need to open an RMA ticket at this link: 
You would need to send clear pictures of the GPU from all sides and also do the same for the block and backplate to that RMA ticket. 
--------------------
 
I have already had 2 GPUs die to this block. They are still saying that its user mounting error. I guess im just not skilled enough to mount a EK block anymore. Time for the trash bin. Bought EK all my life. I guess its over.

Mordgier

eg0911
Here is my latest response from EK.
---------------
There seems that there was some misunderstanding, we did do extensive research on the topic. 
Please understand that we pulled back those "faulty" blocks and tested them and all worked perfectly on our GPUs. The same blocks were then returned and were used by users to mount them on the RMAed gpus and then it worked perfectly, they confirmed that. We do not have a single confirmed case that EK block "killed" a single FTW3 GPU. All the GPUs were RMAed by EVGA and are now working on the same blocks.  That is why we are confident to say that you can mount the same block to the RMAed GPU as we are certain there are no issues with the block or backplate. That was not a single confirmed case but many of them, and there was not one confirmed case that the issue was because of the EK block. 

EK is renowned to admit and fix its mistakes which is evident in the past if it comes to that. 
That is why we would like to offer you to send us your RMAed GPU(when you receive it) the block, and the backplate, and we will mount it for you, do the tests and send it back to you. 
For that, you will need to open an RMA ticket at this link: 
You would need to send clear pictures of the GPU from all sides and also do the same for the block and backplate to that RMA ticket. 
--------------------
 
I have already had 2 GPUs die to this block. They are still saying that its user mounting error. I guess im just not skilled enough to mount a EK block anymore. Time for the trash bin. Bought EK all my life. I guess its over.

 
I would absolutely take up EK on mounting the block for me if they can guarantee that they won't fry it. Hell, I'd pay for that service.
 
What bothers me is that my card did work initially with the block - PC was rebooted fine several times and had video - then dead. 
 
So EK could mount the block, turn it on, get video, send it to you, and then it'll die on you within hours.
 
 

Mine ran for almost a month the 1st time. Its just not worth it for me until these gpus are on shelves.

eg0911

Mordgier

eg0911
Here is my latest response from EK.
---------------
There seems that there was some misunderstanding, we did do extensive research on the topic. 
Please understand that we pulled back those "faulty" blocks and tested them and all worked perfectly on our GPUs. The same blocks were then returned and were used by users to mount them on the RMAed gpus and then it worked perfectly, they confirmed that. We do not have a single confirmed case that EK block "killed" a single FTW3 GPU. All the GPUs were RMAed by EVGA and are now working on the same blocks.  That is why we are confident to say that you can mount the same block to the RMAed GPU as we are certain there are no issues with the block or backplate. That was not a single confirmed case but many of them, and there was not one confirmed case that the issue was because of the EK block. 

EK is renowned to admit and fix its mistakes which is evident in the past if it comes to that. 
That is why we would like to offer you to send us your RMAed GPU(when you receive it) the block, and the backplate, and we will mount it for you, do the tests and send it back to you. 
For that, you will need to open an RMA ticket at this link: 
You would need to send clear pictures of the GPU from all sides and also do the same for the block and backplate to that RMA ticket. 
--------------------
 
I have already had 2 GPUs die to this block. They are still saying that its user mounting error. I guess im just not skilled enough to mount a EK block anymore. Time for the trash bin. Bought EK all my life. I guess its over.

 
I would absolutely take up EK on mounting the block for me if they can guarantee that they won't fry it. Hell, I'd pay for that service.
 
What bothers me is that my card did work initially with the block - PC was rebooted fine several times and had video - then dead. 
 
So EK could mount the block, turn it on, get video, send it to you, and then it'll die on you within hours.
 
 

Mine ran for almost a month the 1st time. Its just not worth it for me until these gpus are on shelves.

At a month though it's pretty hard to say if it really is WB related. After all it wouldn't be the first GPU to drop dead out of the blue https://forums.evga.com/3090-Died-Again-This-is-ridiculous-m3307211.aspx   - and plenty of people are fine - https://forums.evga.com/30803090-FTW3-waterblock-deaths-m3235633.aspx  Yes I know your RMA card died as well - and yup that's certainly not just a DOA card  - but I have no reason to suspect that the failure rate is not the minority overall.
 
As always, the failure rate is over reported vs success rate.
 
If my card didn't die - I wouldn't be here posting about it working fine. So all we see here are people whose cards fail. Those for whom it worked fine are too busy gaming to come to the forum to laugh at our misery.

JustinMcLauchlin
So best course of action at this point, is to do what? I still plan on watercooling since my build was based around it, so I don't really have a choice to revert to air. Besides heating up the GPU under a bench or two before removing the stock air cooler, what is recommended? Any new thermal pads somewhere? Any help on experience with it would be great. 

Side note: my RMA gpu that I will be getting back has the 2114 serial number, and the original one that died had the 2014 one. Not sure if that matters any tbh

I'm getting my replacement GPU today - and I'm going to be taking another go at it. 
 
I have reached out to EK for the smaller standoff but they have so far just stalled with stupid questions like "What do you mean the GPU is dead? Do you have any video out?" 
 
All I wanted was for them to send me the smaller standoff and I offered to pay shipping yet they are basically stalling.
 
In short EVGA was able to get a $2000 GPU to me faster than EK a 10 cent standoff.
 
Either way, I see no path forward besides retrying with the new GPU.

MeleeSparks
I just received both my RMA FTW3 3090's.  I'm pretty f*ing annoyed at this point.  One of the 3090's seem fine, however the second 3090 has the "void" sticker on the backplate removed, and I can obviously see a ton of thermal putty/paste missing and nothing making contact to the surface -- just from looking at the side of the card.  I threw both 3090's into my rig (still air cooled) and the one that's obviously missing thermal putty is about 15 degrees hotter during idle.  Not only that - the only LED working on the video card is the "EVGA" logo on the backplate.
 
Enough to send this thing back or what?

No LEDs over the PCIE plugs? That'd be enough for me to send it right back.


On the other hand, if you can clearly see that the card works and the air cooler is basically detached - well - seems like a candidate for easy removal of the air cooler and putting a block on.

JustinMcLauchlin
So best course of action at this point, is to do what? I still plan on watercooling since my build was based around it, so I don't really have a choice to revert to air. Besides heating up the GPU under a bench or two before removing the stock air cooler, what is recommended? Any new thermal pads somewhere? Any help on experience with it would be great. 

Side note: my RMA gpu that I will be getting back has the 2114 serial number, and the original one that died had the 2014 one. Not sure if that matters any tbh

Sure you can.  It might not be "optimized" for air, but it'll work just fine.  My build was built for water, but outside of the crappy 2nd RMA card I received, it's currently running fine on 100% load during my workflow.
 
before

 
after receiving RMA's and just running on air (looks like **** because the diverted loop path from CPU to bottom rad)

jnayna
Working to see if I can return the waterblocks and support now saying the smaller standoffs are not for performance but will be changed for all future production. Support is still willing to send me the smaller standoffs if I want them.
 
Takes 2 days per 1 response so might not be this week, but getting feeling that as 'nothing is wrong with the EK product' won't be able to return the block.
 
Ended up upgrading to larger size rad to go with Lian Li O11D-XL case and also fighting over fact that I had installed fans on the smaller rad. (Not Mint)
 
Wish EK was as easy to work with as EVGA....

Yup - EK has been an absolute pain to work with. I explicitly stated that I wanted the smaller standoffs right away - and they're just...dragging their feet. 
 
I settled on EK because Heatkiller or Aquacomputer blocks aren't available and Optimus blocks have a wait list that is unappealing.
 
I think next time, I'll wait for higher quality blocks. Even if EK isn't the cause, dealing with the has been a poor experience. 

I'm wondering if they are purposely putting them back together kinda slack-like just cause they know people will put a block on them  or so, i hope

Mordgier
So I got my RMA 3090. Clearly a refurb with some of the backplate screws heads stripped - and actually I saw that as good news.
 
If the replacement card had clearly been disassembled before and still works - that's good right? Looks like whoever had it before took it apart, squirted on enough thermal paste to drown an elephant on the GPU and slapped it back together. Either way the air cooler and backplate came off with zero effort. In fact - they sort of popped off once unscrewed. The card clearly has curvature and was being flexed by the oem aircooler - not great.
 
I tested the card on air, tested it with the air cooler removed  - and finally yolo'd and slapped the block on.
 
So far - it works. Original card also worked at first so I'm not counting my chickens yet. I'll let it run overnight and report back I suppose.

Which thermal pads are you applying?