Modding EK Coolstream SE

Hermit2001

Trash Compacter
Original poster
Apr 26, 2016
43
31
So everyone slams the EK Coolstream SE for generally poor cooling performance...
http://thermalbench.com/2015/08/04/ek-coolstream-se-480-radiator/
http://www.xtremerigs.net/2015/07/03/ek-coolstream-se-360mm-radiator-review/
http://www.xtremerigs.net/2015/02/11/radiator-round-2015/7/
EK have compensated for slim design by increasing the fin density to try to maintain heat-transfer surface area. Unfortunately this has the effect of requiring high pressure/thick fans to take advantage. Yet many people report better temps (by ~few degrees) with lower FPI rads, such as the Alphacool ST30 or Darkside LP240).

So why don't we try and mod the EK SE!
So, I will be testing the effect of reducing the FPI in order to increase the air-flow. To do this, I'll be pushing every two fins together. This will reduce the surface area so it will interesting to see if we can get temps down by a few degrees.

This is going to fit into the bottom of an NCase M1 so I want to test it with thin fans: Scythe Slip Stream 120 Slim.
 
Last edited:

Brokoii

Caliper Novice
Jan 11, 2017
31
30
So everyone slams the EK Coolstream SE for generally poor cooling performance...
http://thermalbench.com/2015/08/04/ek-coolstream-se-480-radiator/
http://www.xtremerigs.net/2015/07/03/ek-coolstream-se-360mm-radiator-review/
...and that partly stems from the airflow resistance, not withstanding the fact that it is a slimmer rad. EK have compensated for the reduced heat-transfer surface area by increasing the fin density. Unfortunately this has the effect of requiring high pressure/thick fans to take advantage.

I have a SE 120 that I'm modding to fit into the bottom of an NCase M1 along with the thinnest 120mm slim fans: Scythe Slip Stream 120 Slim.

I had hoped I could get a push-pull arrangement (height: 26+12+12=50mm) as some people have managed an SE with a 25mm fan (26+25=51mm) but mine doesn't fit. Maybe in the future with more modding to the fans and rad shroud to reduce stack height but for the time being I will be doing a different experiment.

So to improve compatibility with Slip Stream 120s, I will be testing the effect of pushing every two fins together; the flow area should increase, allowing more airflow for the same RPMs. This will reduce the surface area so it will be an interesting trade-off to see what the limiting factor on cooling performance is.
What gpu are you running? I managed to fit a rx480 with a 25mm fani the front, so the IO is still working. I just had to take the rad apart and cut the plate away which reads coolstream se.
 
  • Like
Reactions: ricochet

Hermit2001

Trash Compacter
Original poster
Apr 26, 2016
43
31
What gpu are you running? I managed to fit a rx480 with a 25mm fani the front, so the IO is still working. I just had to take the rad apart and cut the plate away which reads coolstream se.
Titan Z. May have to modify the GPU water block as fittings are still clashing. I've already milled out some of the acetal cover to access the GPU fan header.
 
Last edited:

Brokoii

Caliper Novice
Jan 11, 2017
31
30
Titan Z. I too have ditched that part of the rad case. May have to modify the GPU water block as fittings are still clashing. I've already milled out some of the acetal cover to access the GPU fan header.
I had problems getting my 480 in the slot, but when I tilted the card a little bit upwards it popped right in.
 

Hermit2001

Trash Compacter
Original poster
Apr 26, 2016
43
31
Update #1:

So, having spent some time playing with the fins and some wooden toothpics, the mod now is complete. Not sure if we will see a huge difference in temperatures but worth a shot. Above is mid-mod and you can hopefully see some variation in fin spacing in the middle few rows.

Certain areas were easier than others as there was a bit of slack in the fins. Other areas and the fins hardly budged at all. You have to be careful not to force fins too hard or you will break the fin away from the tube, which is bad.

A better method would just slice out every-other fin, but I don't have a water-jet cutter or EDM machine to hand. Still, we will see how the brute-force method works.
 
Last edited:
  • Like
Reactions: Phuncz

Hermit2001

Trash Compacter
Original poster
Apr 26, 2016
43
31
Update #2:
Loop is finally ready! Just needed a new PSU, some new fittings and bits...

I'll be testing a standard EK 120SE / modified-120SE radiators in the following fan configurations:
single (pull) slim fan / double (push-pull) slim fan / single thick fan.

The slim fan is the Scythe Slipstream Slim 2000RPM SY1212SL12H, the thick: EK Furious Vardar FF5-120.

If anyone is interested, the pump is a Delphi 18W DDC-1VC (volume compensator) unit- the leads have been soldered to a sata power connector. Its friggin loud, but cheap; fine for testing purposes only.

I will probably leave the fan profile as a flat, low (quiet) setting to ensure that any small difference in cooling ability of the radiator mod comes to the fore (higher temp difference).

I will record the peak GPU core temps (with EVGA Precision), ambient air temp and fan RPM on a benchmark after the loop has reached steady-state.

For stress testing, I have decided on Realbench.
 
Last edited:

Hermit2001

Trash Compacter
Original poster
Apr 26, 2016
43
31
So initial testing done. Turns out fans set to 'quiet' is clearly not enough; temps were heading for 90+ after 10mins without much sign of slowing down. Definitely transient and not steady state behaviour.

Having recorded ambient and GPU temps, I've produced the following data with 2x Scythe Slipstream Slim 2000 (push/pull). Here dT is the temp over ambient.

Rad: Standard, fans at 100% (~1910 RPM)
dT(gpu1)@45
dT(gpu2)@38

Rad: Modified, fans at 100% (~1940 RPM)
dT(gpu1)@40
dT(gpu2)@35

Seems promising. Quite impressive that a 120mm rad can dissipate 450W (if the GPU is actually running at 20% over its max 375W rating) at a reasonable temperature.

Next up is a Vardar Furious FF5-120 at 100%
Rad: Standard, fan at 100% (~2820 RPM)
dT(gpu1)@38
dT(gpu2)@30

Rad: Modified, fan at 100% (~2820 RPM)
dT(gpu1)@31
dT(gpu2)@26

I'm sure there is a 1-2C error in these but the trend does seem to suggest a small improvement in cooling ability due to increased airflow.

Is anyone interested in a different fan configuration or RPM level?
 
Last edited: