First glimpse at the EK-FC7970

[Martinm210]

The EK-FC7950 water block's hydraulic curve resembles much the one of EK-FC7970. The design once again features extremely low flow resistance, again even lower than the CPU waterblock EK-Supreme HF. As a result any medium performance water pump will run four of these water blocks yet still offer enough flow for optimal system operation. Picture below shows the hydraulic curve:




The EK-FC7950 water block - which will come in all four versions (Copper Plexi, Copper Acetal, EN Nickel Plexi, EN Nickel Acetal) - is already listed in the EK webshop and will be available for purchase on March 13th 2012. The matching FC Backplate is also in the works and will be available in the following weeks.


Niko Tivadar - tiborrr - for EK Team
February 29th 2012

Glad to see you guys doing your own pressure drop testing now, very good!

One of these days I'll put my 570 under water again..VGA is still on air....

[tiborrr]

Thanks Martin!

Can we use just the back plates on reference 7970? I'm not water cooling but I want a back plate.

Well, you can, but you have to figure out your own mounting mechanism, we do not include one.

[Rattle]

Does it not just use 4 screws from the existing bolt pattern?

[tiborrr]

No, the block uses it's own M3x0.5 threaded mounting mechanism.

[bill_d]

any word on the EK-FC7970 DCII

i got two ASUS HD7970-DC2T-3GD5 shipping from newegg and will need blocks

[tiborrr]

Soon, my friend! Keep them warm for a bit

[Bestofbritish96]

Hi,

I have three XFX Core Edition 7970 cards with installed EK blocks and back-plate. My middle card is idling at around 25-28C and 50C under load the other two idle around 58C and go straight to 90C under-load and I have to quit Battlefield 3 quickly!

I am going to re-mount the two blocks and back-plates tonight, are these types of problems normal and I just need to experiment? Is this something that could be caused by excessive Artic Silver or is it more likely to be contact problem e.g over-tight screws?

Also, I read there are problems with capacitor changes to the reference cards, how do I check I have a compatible reference card?

[tiborrr]

Best of british: What does your contact between the GPU and water block surface looks like? It appear you either have:
- air bubble or no water in both blocks (perhaps you're using parallel flow with unadequate pump?)
- no contact on both blocks

Overtightening the screws will not help to solve this issue.

Please avoid using thermal greases such as Arctic Silver which conduct electricity. There are a lot of good, electrically non-conductive thermal greases on the market.

Regarding cap issue: http://www.ekwaterblocks.com/index.p...t01returnid=17
The higher capacitor should not affect the temperature on the GPU but would result in warping of the graphics card voltage regulation module. Effectively you would crush the capacitor and/or nearby Driver/MOSFETs.

[Elmy]

Hi,

I have three XFX Core Edition 7970 cards with installed EK blocks and back-plate. My middle card is idling at around 25-28C and 50C under load the other two idle around 58C and go straight to 90C under-load and I have to quit Battlefield 3 quickly!

I am going to re-mount the two blocks and back-plates tonight, are these types of problems normal and I just need to experiment? Is this something that could be caused by excessive Artic Silver or is it more likely to be contact problem e.g over-tight screws?

Also, I read there are problems with capacitor changes to the reference cards, how do I check I have a compatible reference card?

Do you have enough rad for all three cards? should have at least a high quality 480 just for those 3 cards....

[Bestofbritish96]

Thanks for the advice.

My thermal paste is Artic Cooling MX-2, apologies for confusion.

I have an XSPC 120mm RX480 radiator and a single black-ice stealth 120mm radiator.

I have a lower overall system flow rate of 60L/PH using dual pumps, I am in a serial EK bridge configuration.

I will take the blocks off and check my contact points as suggested.

[tiborrr]

60L/h is a very, very low flow. I'm sure your blocks still haven't been properly bleed of air, hence the spike to 90°C.
Still, it wouldn't hurt to check the contact with the GPU dies, yes.

What other components do you run in the system? CPU block, pumps? The EK-FC7970 is one of the best flow full-cover water blocks on market to date, if not best.

[Bestofbritish96]

Tiborrr,

I am running dual MCP35x pumps with EK top, Supreme HF CPU block, EK Rampage 4 blocks, EK Dominator RAM blocks and an Aquaero block for my controller.

I am going to re-route my loop and report back.

[tijgert]

Hi Tiborr, I'm also desperately awaiting the Asus 7970 DCII block... I'd hate to find out too late and having to wait in line... Any chance for pre-ordering one of those beauties??

[Bestofbritish96]

I made some minor adjustments to my loop and am now running 120Ltr/H (thanks to Martinlabs page for the calibration number). The loop has been running for about 3 hours, the middle GPU is still registering 23C at idle clocks and the other two cards at 48-50C idle clocks.

How long should I wait for possible air bubbles to disappear from the EK 7970 blocks and bridge before I rip the two cards out and re-mount ?

[tiborrr]

This is too big of a discrepancy, there should be no more than a couple of degrees difference between the cards. Try pulsating the flow to get the air bubbles out, also try tilting the case. Power up only the water cooling
pump, do not do this with your computer turned on.

Regards,
Niko

[Bestofbritish96]

Niko,

I have changed the configuration of my loop. I have a flow rate of 120L/ph. I have remounted the blocks on the two "problem" cards, I also put the known "good" card into the end slot (slot 3 on the bridge)

What I am seeing is the same problem, the first (slot 1) and last GPU (slot 3) having high-idle temperatures (45C) and will not go under load. This card (slot 3) was previously in the middle slot and causing no issues.

The middle card position (slot 2) is fine, I see 27C idle temps and 40-50C under load.

Is this air trapped in the GPU blocks or, inside the bridge itself?

[tiborrr]

Hi!

I believe you have air trapped inside your water blocks and or bridge.

The only other explanation is that you are not using the correcti in/out connections on your bridge. Do you have serial or parallel bridge? A picture of the setup would prove very useful.

If you have a serial bridge this could be why only your middle card has normal temperatures. Perhaps you are flowing the water past the first and third block.

http://www.ekwaterblocks.com/shop/EK...1109855096.pdf

[Bestofbritish96]

Here is a picture of the setup (novice build!). It is a Serial link block. Water-in is on the left (upper port) Water-out is on the right (lower port).

Edit: looks I have made a simple mistake and got my loop incorrectly plugged into the wrong ports on the bridge, this explains why the middle card is ok - water is passing through it correctly and not the other two?

[tiborrr]

Yes, this is precisely what you have done wrong. Just use the opposite ports and everything will work as it should. No need for a teardown of cards!

[Bestofbritish96]

Many thanks for your help Niko. I have adjusted the loop and all my GPU's are now at 24c idle.

[tiborrr]

Hehe, great! Keep on clocking!

[tiborrr]

We have completed internal stress testing of our latest EK-FC7970 DCII water block, engineered specifically for ASUS design Radeon HD 7970 DirectCU II series graphics cards.





The test has been conducted using the following water cooling components and measuring equipment:
- EK-FC7970 DCII prototype (RTM) water block
- EK-RAD CoolStream XT 360
- 3x Yate Loon D12SH-12 120mm 2200rpm fans
- 2m 12mm ID tubing (1/2" ID)
- D5 Vario water pump + EK-D5 X-TOP V2
- EK-RES X2 Multioption 250 Advanced
- EK-12mm High Flow fittings

Measuring equipment:
- Greisinger GTH 175/Pt J-type thermometer, calibrated
- Voltcraft IR-800 20D IR thermometer
- Digmessa digital flow meter FKHU series

Since the DC2(T) series graphics cards are built with overclocking in mind we decided to push our sample in order to stress test our water block's cooling engine. Furmark torture benchmark which stresses the graphics card to the maximum was ran for more than one hour. Overclocking the graphics card from factory set 1000/1400MHz to 1300/1600MHz (GPU/RAM respectively) with Vgpu voltage set at 1.25V (1.22V after Vdroop) yielded the following result:

Tgpumax = 44°C
Th2omax = 33.1°C
---
dTmax = 10.9°C = ~ 11°C



Where Tgpumax is the highest temperature of the GPU core, logged by software; Th2omax is the highest temperature of the coolant measured by thermometer; dTmax is the highest delta (difference) between the GPU core temperature and the coolant temperature - lower is better.

Please note that the smallest resolution for Tgpu probe is 1°C therefore a rounding error is present. Considering this fact we can see that dT does not increase more than a 1°C even at overclocked conditions. The VRM (voltage regulation module) also remains cool, the hottest measured temperature was ~ 54°C (behind the MOSFETs, on the back side of the PCB; 61°C in overclocked mode). Note that ASUS software probe reports VRM temperatures higher VRM temperatures - 56°C in 2D idle mode (non-overclocked state) and 86°C while under full load (overclocked state).

Mounting mechanism also support installation of ASUS original backplate but user can install water block without one as well. EK will also include a 2-slot I/O bracket which will free the adjacent PCI/PCIe expansion slot and ease the installation of multiple graphics cards. Due to the shear size of the printed circuit board the EK-FC7970 DCII water block uses G1/4 ports on acetal top only. The interconnectivity of multiple EK-FC7970 DCII water blocks is ensured via EK-FC Bridge & Link system.

The EK-FC7970 DCII water block again excells in hydraulic performance chart. The design once again features extremely low flow resistance, somewhere in the range of EK-FC7970. As a result any medium performance water pump will run four of these water blocks yet still offer enough flow for optimal system operation. Picture below shows the hydraulic curve:



The EK-FC7970 DCII water block - which will come in two versions (Copper Acetal, EN Nickel Acetal) - is already listed in the EK webshop and will be available for purchase in upcoming weeks. The matching FC Backplate is also in the works and will be available in the following weeks.


Niko Tivadar - tiborrr - for EK Team
March 21st 2012

[junkun13]

no inlet/outlet on this side? weird

[Roadhog]

no inlet/outlet on this side? weird

It's so you have to spend more money to buy their bridge.

[tiborrr]

Haha, Roadhog. It's because the card is already wide enough. Making ports from above in this case means sacrificing chassis compatibility.