dng25

cnez0528
damn mine 523x still not shipped.

What color did you order?

Gotspeed_2000

fragility_V1
Got my block last week and have it installed on the card.  Fairly straightforward process, but was missing a screw from what they provided.  Have a set of small metric, so was able to find a replacement.  Black, not silver, but I got it installed. 
 
Now trying to put the loop back together but am finding a slow leak under pressure testing.  Gonna try to isolate it and see if it's in the rads/tubing rather than the block itself. 
 
Two things of note thus far:

1. I have the Signature block the cerakote coating and the satin silver cerakote of the KPE block is quite a different look than the raw silver of the stock signature block.  I was under the impression they would be similar, but that's not so much the case. Well at at least they're both silver-ish.
2. I'm swapping my Z590 Dark for a Z690 Apex, and tried it on the Dark to see if it would even fit.  The block will not fit into Z590 Dark primary x16 PCIe slot horizontally.  I would assume it's the same for the Z690 Dark as well.  So you plan to use the Optimus KPE block with a Z590/690 Dark, you're going to have to go with a vertical GPU mount.

Please let us know if it's the block or your system when you find the leak.  I pressure test all my components individually first before I put them into systems to reduce the amount of possible failure points.  If it is your block, take a look at the cold plate and if it is the Optimus block please share where your leak occured.  Mines is fine after I took off the cold plate and re-installed it with even tension across the 4 cold plate mounting screws.  

fragility_V1
It's the block. Still chasing the leak. Tried reseating the backplate, the gpu block, and the terminal. No luck there. Just repeated the cold plate but still have to leak test. I feel like it's the terminal. I want to try removing the rubber gaskets, but don't want to tear them taking them out. All the bolts on the cold plate were pretty snug. But maybe seating helped.

fragility_V1
https://wccftech.com/nvid...ual-12-pin-connectors/

https://videocardz.com/ne...2-pin-power-connectors

Groan. If this is true would be nice the PCB layout didn't change significantly, but doesn't look to be the case.

Gotspeed_2000
I also recall seeing some content creators suggest that was going to be the move to get the VRMS off the backside of the gpu and move it to the front where the air radiators are for better VRM performance.  Moving the VRMs from the back I would think takes away most of the reason to have active back side cooling.  Most of the production costs associated with the Optimus block should have been covered by the cost of the block during the initial planning process when they designed the block and factored pricing.  The KP is a small market so to recover costs related to development I would think would require a key number of blocks sold to recoup development costs.  From what it sounds like in the article linked, there might be two models of the KP sold at the same time.  If so, it would provide Optimus a slowly growing market to continue to sell its block.  The current KP would benefit from the rear cooling, however those who are focused on overclocking might want that ability to pull more power and that small performance uplift provided by the rumored release.  
 
For the hardcore users, I would think they would want the option for additional power.  Not sure how much more power the current series of gpu cores can take without adding or changing the gpu die.  The amount of heat that will be generated by going to the change of power delivery if it was to actually draw that much power would I think require artificial cooling ie LN.  If you don't use LN, then I'm not sure how much real performance you'll get out of the refresh.  I would think it might be better to wait till the new 4090 comes out in a year or so unless cost was not an option.  
 
For casual overclockers and gamers who use the KP, the current KP is still more than enough for most people imo.  If you are 4k gaming, I'd wait it out till the next gen comes out before dropping another ton of cash.  Not sure where on the PCB Nvidia would move the VRMs to so until pictures come out for the pcb layout its hard to say how much that impacted the layout of the board and if by moving them does it create any other issues that can negatively affect performance.  In PCs, we often see that when manufacturer's try to resolve one issue, they create additional issues trying to fix something else.  Sometimes they are directly related, other times it may not but we wont know for sure how this pcb refresh is going to work until they release it and people put it through its paces in the real world.  Nvidia had time to test and create the current generation of gpus, so the VRMs on the backside and the heat issues related is something I would have thought they would have seen in their testing but they still released it anyway.  Not sure how much faith we should have that the new pcb design will work better than the current and if so by how much.  Even AIBs could have done what Optimus does with their passive rear plate and increased performance of the current gen gpus, but they all didn't for one reason or another.  That upgrade for AIBs wouldn't have costed too much more but would have impacted performance of standard AIB cards by reducing the temps of the VRMs and the associated heat soak of the pcb.  
 
For me, I'm going to use my long awaited KP with the Optimus block until the new 4090 comes out.  I am not a heavy overclock guy, so I think it will be fine for my use.  If you are or thinking of using LN, then it is really up to you as by the time you get a new KP, how long will it be before the 4090 comes out and would the small performance gains be worth it for the increased cost.  On this to each their own and their budgets.  

Nereus

fragility_V1
https://wccftech.com/nvid...ual-12-pin-connectors/

https://videocardz.com/ne...2-pin-power-connectors

Groan. If this is true would be nice the PCB layout didn't change significantly, but doesn't look to be the case.

link and link  - Whaaaaat???? 1250W EVGA KingPin 3090Ti???? I find that hard to believe, at least not without a restricted custom bios/firmware, and definitely only for LN2 setups. Can solder even survive that kind of power and resulting heat? Hell, with a heavily overclocked CPU, even a 1600W PSU may not be enough to run that in a system. Bring on the mysterious 2000T2 PSU!!! I'm guessing LN2 bench setups only with a separate PSU for the GPU. Oh, and and 2x12 pin power connectors, or maybe a new 16 pin connector? Hope they are supplied with the GPU, or good luck finding stock.
 
I can't see Optimus putting out a block for this before 2023 (if ever) based on their existing efforts.
 
On a positive note, maybe I'll get lucky and finally be able to buy a 30 series KingPin! lol.
 

Maverickyao
This is a good warning to Optimus. They need to know that there will be consequence of breaking promise and missing dues.

Gotspeed_2000The only saving grace might be if EVGA designs the new TI KP pcb around a standard 3090 FTW3  TI pcb which would then allow both TI models to share the use of a single block design.  

Gotspeed_2000
Just finished up the block conversion.  The optimus instructions are okay, but be sure to read ahead and plan.  The alignment of the two plates and the top port can be tricky so read ahead and plan for pages 3/4 of their instructions.  I have to leak test this tomorrow, but I'm pretty sure it should be fine.  The use of instructions from the EVGA hydrocopper block conversion is really good for taking apart the Hybrid.  Read that a few times and follow the disassembly instructions and you should be golden.  
 
My block is Silver Cerakote with Nickel cold plate.  The silver cerakote is nice, but there are imperfections that kinda look like over-spray in certain areas.  Nothing too noticeable, but not perfect either.  
 
The finished block is really heavy.  I would not feel comfortable mounting this horizontal without major support brackets to help spread the weight load.  If I was to mount this horizontal into the PCI slot, I would use the Lian Li GPU anti-sag bracket, and then also a gpu post support to help hold up the front.  This thing is that heavy.  Vertical mount should be fine as you are not loading weight onto the PCI slot.  I bought the Lian Li O11 XL vertical mount bracket with the PCI-E 4.0 riser cable.  I am already running a PCI-E 4.0 cable in my vertical mount, but this vertical mount kit will allow me to adjust the gpu position either further/closer to the mobo because the vertical mount has multiple locations to mount the PCI-E riser cable.  It will still be a while before I'm able to get this into my pc, but atleast the conversion is done.  
 
Merry Christmas to everyone.  Be safe cause Covid does suck.  My son has it so we are all stuck in not being able to go out.  Thankfully his symptoms so far seem to be mild overall.  He is vaccinated so I am assuming he has the Omicron.  Take care of yourself and your families.  

bavor

Nereus
link and link  - Whaaaaat???? 1250W EVGA KingPin 3090Ti???? I find that hard to believe, at least not without a restricted custom bios/firmware, and definitely only for LN2 setups. Can solder even survive that kind of power and resulting heat? Hell, with a heavily overclocked CPU, even a 1600W PSU may not be enough to run that in a system. Bring on the mysterious 2000T2 PSU!!! I'm guessing LN2 bench setups only with a separate PSU for the GPU. Oh, and and 2x12 pin power connectors, or maybe a new 16 pin connector? Hope they are supplied with the GPU, or good luck finding stock.

Most people running one or more high wattage GPUs run two or more PSUs and not a single PSU.  So a 2000+ watt PSU isn't needed.  In the past, when I was going for my highest scores, I ran two and/or three PSUs on two different circuits in the house.  I used a 850 watt 80+ gold PSU for the motherboard, CPU, and cooling.  I used a 1500 watt BeQuiet and a 1600 watt EVGA PSU for the graphics cards.  Each card drew a peak of 700-800 watts and sustained 600-700 watts depending on the benchmark.  That was without going below zero.