EK-Quantum Vector RE RTX 3080/3090 Active Backplate D-RGB - Plexi

[VSG]

We begin the much-awaited RTX 3080 GPU block series from EKWB. The EK-Quantum Vector Active Backplate is the first of its kind, a second water block on the back of the GPU PCB. It complements the company's flagship EK-Quantum Vector GPU block and cools critical components on the back of the RTX 3090.

Show full review

 

[mouacyk]

Thanks for the review. I'm not sure I understand the VRAM temp readings. If I add the delta to ambient, the VRAM is only running in mid 40's? Even on 3080, I'm seeing mid 60's with a heatsink on backplate.

In any case, seeing just 1 degree lower with the active backplate, I'm glad I went with a $7 copper heatsink to attach on rather than spend the equivalent of another water block.

 

[claster17]

If I add the delta to ambient

Delta to coolant temperature.

I would have liked to see ETH mining added to the test as a worst case for VRAM temps. Time Spy Extreme is rather tame in comparison.

 

[Mussels]

I own this with the active backplate, as my first custom cooling hardware ever
My thoughts align with yours on a lot of points, from a beginners perspective:


1. The generic packet of screws was an issue, as using a too long screw could crack acrylic blocks (and i DID use too short ones at first, that let the block fall off!)
1a: I had the same issue as the backplate review with one or two screws not wanting to actually screw into place, and needing to do that tricky one first

2. Pre cut thermal pads would be nice, but the instructions were clear enough

3. Information on how 3090s top VRAM handle with no backplate at all is NOT mentioned anywhere on EKs information, despite being an optional purchase

4. Active backplate should be offered as part of a bundle, to cut back on excess product (the extra top fitting, screws, thermal pads etc)

5. The front plate has a mere 5 LED's, where the active backplate has 8? Why?? the front is 99% likely to be more visible...

6. The LED cables dont match where they leave the GPU, and they dont match in length. Thats a minor issue but the products are made to be paired together so why didnt this happen?

7. On a 3090 this makes an insane difference depending on your cooler, i went from 105C and throttling while mining at -400MHz on the VRAM, to 80C with a 1500MHz VRAM overclock. 30C+ drop!

 

[VSG]

Delta to coolant temperature.

I would have liked to see ETH mining added to the test as a worst case for VRAM temps. Time Spy Extreme is rather tame in comparison.

Let me see if I do something similar in the next set of testing.

 

[Emanulele]

Would have been great to see it tested with a 3090 which is probably the target GPU here, I wouldn't bother getting one with a 3080.

 

[VSG]

Would have been great to see it tested with a 3090 which is probably the target GPU here, I wouldn't bother getting one with a 3080.

Yup, but also I was lucky to even get a reference 3080 so can't help it unfortunately.

 

[Breit]

7. On a 3090 this makes an insane difference depending on your cooler, i went from 105C and throttling while mining at -400MHz on the VRAM, to 80C with a 1500MHz VRAM overclock. 30C+ drop!

I can second that. The 3090 definitely needs active cooling on the backside-VRAM while in my opinion the 3080 does not!
Hopefully that backplane also fits on the 3080Ti (reference) so it is not wasted developement time, but I doubt it.

 

[VSG]

I can second that. The 3090 definitely needs active cooling on the backside-VRAM while in my opinion the 3080 does not!
Hopefully that backplane also fits on the 3080Ti (reference) so it is not wasted developement time, but I doubt it.

Keep in mind that I don't know anymore than you here, but I would be shocked if the RTX 3080 Ti reference PCB is any different from the 3080 reference PCB.

 

[Breit]

Keep in mind that I don't know anymore than you here, but I would be shocked if the RTX 3080 Ti reference PCB is any different from the 3080 reference PCB.

It has to be different. Probably more like the 3090 reference with one memory chip space left blank.

 

[claster17]

We pretty much already know that the 3080Ti will come with 12GB and full 384bit. Coincidentally every 3080 PCB has 2 unpopulated spots for memory.

 

[Mussels]

It has to be different. Probably more like the 3090 reference with one memory chip space left blank.

My view is that its more or less gunna be a 3090, missing the back VRAM - with some small decrease to the core to keep the product stack distinct (some shaders gone, lower TDP, whatever)

 

[Emanulele]

My view is that its more or less gunna a 3090, missing the back VRAM - with some small decrease to the core to keep the product stack distinct (some shaders gone, lower TDP, whatever)

VRAM only on one side is a cooling win there.

 

[Mussels]

VRAM only on one side is a cooling win there.

I agree and thats why i think they're going down that road - its a perfect in-between of the 3080 with more ram vs a 3090 with less ram, and they can avoid the issue of the overheating 3090 top VRAM entirely by just not having it there

 

[Breit]

We pretty much already know that the 3080Ti will come with 12GB and full 384bit. Coincidentally every 3080 PCB has 2 unpopulated spots for memory.

Alright, must have missed that. Well, then all is good cooling wise. But you also don't need an actively cooled backplate then?!

 

[VSG]

Alright, must have missed that. Well, then all is good cooling wise. But you also don't need an actively cooled backplate then?!

Need... no. Want?

 

[ypsylon]

I have this block+backplate combo. Amazing drop in temperatures.

When I render no longer have to worry my 3090 will cook itself to death. Before adding block was running on air for about 2 weeks to test things. When I started rendering memory was 96-100C in 3 minutes. Stickers on the default backplate (just thin piece of steel plate) unstick from heat alone. Now it's more like 55-60 on the core (at full load when rendering, in gaming it's much lower) and memory keeps at the respectable 70C. Stock cooler was pure madness. In games yeah it did fine, but for work don't even think about air cooler on 3090.

There're 2 Cons which are not listed.

1. If you want to remove stupid RGB you have to literally dismantle whole new product to remove the strip. That's the most retarded thing ever.

2. Packaging is pure joke. Backplatee flops around the box even if cardboard box is 50% bigger than the actual block is shipping. My packaging was completely trashed inside. Box was perfectly fine outside, but when I've unboxed I was in state of shock, more so because I've prepared card in advance and waited only for backplate delivery to go straight to the card. EK cheap-out on packaging for a long time. No foam, no clamps. Zero. Compared to companies like Phanteks they are just amateurs.

 

[Leonid]

Thank you for your review, super helpful! Any chance you can review the new Optimus Strix waterblock for 3090?

They claim that the backplate “matches or exceeds the real-world performance of liquid-cooled or heatpipe backplates. In testing, the t-junction will max at 72c during the most intense mining”, and it would be interesting how it compares to say EK active backplate in a real-life ETH test.

That sounds like a very strong claim, curious if that is actually true for an average-airflow watercooled case.

 

[Mussels]

I have this block+backplate combo. Amazing drop in temperatures.

When I render no longer have to worry my 3090 will cook itself to death. Before adding block was running on air for about 2 weeks to test things. When I started rendering memory was 96-100C in 3 minutes. Stickers on the default backplate (just thin piece of steel plate) unstick from heat alone. Now it's more like 55-60 on the core (at full load when rendering, in gaming it's much lower) and memory keeps at the respectable 70C. Stock cooler was pure madness. In games yeah it did fine, but for work don't even think about air cooler on 3090.

There're 2 Cons which are not listed.

1. If you want to remove stupid RGB you have to literally dismantle whole new product to remove the strip. That's the most retarded thing ever.

2. Packaging is pure joke. Backplatee flops around the box even if cardboard box is 50% bigger than the actual block is shipping. My packaging was completely trashed inside. Box was perfectly fine outside, but when I've unboxed I was in state of shock, more so because I've prepared card in advance and waited only for backplate delivery to go straight to the card. EK cheap-out on packaging for a long time. No foam, no clamps. Zero. Compared to companies like Phanteks they are just amateurs.

I mean, they have non-RGB versions of the product so point 1. is redundant

 

[VSG]

Thank you for your review, super helpful! Any chance you can review the new Optimus Strix waterblock for 3090?

They claim that the backplate “matches or exceeds the real-world performance of liquid-cooled or heatpipe backplates. In testing, the t-junction will max at 72c during the most intense mining”, and it would be interesting how it compares to say EK active backplate in a real-life ETH test.

That sounds like a very strong claim, curious if that is actually true for an average-airflow watercooled case.

I don't have that GPU, and also Optimus can barely make anything right now. I asked them about a reference PCB GPU block and I don't think it's going to happen anytime soon.

 

[timb320]

Just to add as a frame of reference, I have a strix OC 3090 with the EK GPU waterblock and backplate, coupled with the MP5Works (https://www.mp5works.com/) backplate waterblock to solve my VRAM temp issue, it does reduce temps (idle 40C with ek GPU waterblock), 80c tops while gaming in COD War zone, while nicehash mining in a recent test it gets hot to around 98c but NVIDIA latest drivers quickly de-rated mining. (Note: I don't mine anymore - was a experiment to see if I could pay for the card, but in the end not worth the wear and tear, network bandwidth and not profitable enough = LOL)...

So it appears this active backplate could yield me better numbers, and I'm tempted but already spent money on what I currently have so may opt out for now as long as I can keep my vram temps from going higher then 80c while gaming, only really intensive thing I do on this box other then coding.

 

[Mussels]

Just to add as a frame of reference, I have a strix OC 3090 with the EK GPU waterblock and backplate, coupled with the MP5Works (https://www.mp5works.com/) backplate waterblock to solve my VRAM temp issue, it does reduce temps (idle 40C with ek GPU waterblock), 80c tops while gaming in COD War zone, while nicehash mining in a recent test it gets hot to around 98c but NVIDIA latest drivers quickly de-rated mining. (Note: I don't mine anymore - was a experiment to see if I could pay for the card, but in the end not worth the wear and tear, network bandwidth and not profitable enough = LOL)...

So it appears this active backplate could yield me better numbers, and I'm tempted but already spent money on what I currently have so may opt out for now as long as I can keep my vram temps from going higher then 80c while gaming, only really intensive thing I do on this box other then coding.

I've got the latest Nv drivers, zero change to mining performance. 3090s have no limiters, you're being throttled elsewhere. (and ouch 100C even with that cooling!)

 

[timb320]

I've got the latest Nv drivers, zero change to mining performance. 3090s have no limiters, you're being throttled elsewhere. (and ouch 100C even with that cooling!)

Yeah, I decided to take the plunge and ordered the EK active backplate, based on linus testing it actually will lower GPU temps a little in addition to VRAM temps..

 

[Mussels]

Yeah, I decided to take the plunge and ordered the EK active backplate, based on linus testing it actually will lower GPU temps a little in addition to VRAM temps..

Yes, it can do that on a 3080 - on the 3090 it cools the VRAM and doesnt alter GPU temps that much

Then again, they're looooooow
(Mining load)

 

[timb320]

Tum, doesn't go above 60c on vram now. I'm happy I did it. GPU temps seem the same if not slightly cooler.