Swiftech MCR320 testing results

[Martinm210]

Here is another one done, this radiator testing is quite a chore and it's getting too hot to do a whole bunch more of this. I think I'm going to switch back over to CPU block testing after one more of these...

This was just one of my old MCR320's I pulled from my system I bought over a year ago from Sidewinders. A little beaten and chipped, but still performing well.

Anyhow, here is what I found.

Full review:
http://www.martinsliquidlab.com/Swif...20-Review.html

Pressure drop was better than expected, I guess the old NPT barb versions are quite a bit more free flowing. I also had 5/8" barbs on mine:






Thermal testing it did better than I expected:


Some conversions to a few assumed heat loads for more of a real world numbers look:


Cheers
Martin

[Vapor]

You kickass.

So does the radiator (for the $$).

Awesome as always Martin

[Hidetaka]

Great job Martin!

$132 vs. $57

My vote goes for Swiftech

[eXa]

i have the mcr320 in my old machine, and i have the pa120.3 ready for my new build.

Oh man, i really hope u can test the feser rads. im so curious how the pressure drop is in them and if they can beat the PA series at all.

[MomijiTMO]

Oh they do preform rather nicely. Thanks for testing.

[eXa]

Ye the mcr is way closer to the PA than what i thought. Still, would have liked to see the graph with yates at 6-700rpm too as thats where the PA shines... kinda. U can see that from the graph. The delta between the PA and the MCR is higher on 1000rmp than 1350rpm(on all the heat loads)

[Movieman]

MCR320 on the dual clover since last August and not a hiccup.
Also 25C cooler than air.
great radiators..

[disruptfam]

for the price swiftech rads are awesome... i have several

nice work martin thanks again for testing.

i would like to see what feser has to offer

[lowfat]

They do better than I expected I can handle 1-2C different for 1/3 the price of the PA120.3.

[Martinm210]

Yeah, it was closer than I expected, but I'm still learning about this testing and tempted to retest the PA as a check.

[Serpentarius]

ah.. the memories ... around 2005 or 2006 .. there's been massive tests just like this in this forum .. that was mentally challenging and rewarding
keep up the good work

[leo_bsb]

thanks for your work and time Martin! As always a great review.

[Jimmer411]

Very nice results. Its good to finally have a reliable source to compare all watercooling hardware so that many myths can be disposed.


OT: Martin, in your sig it says you have a D5 EK top? Pics pics pics!

[voigts]

I have been waiting on this as I just switched from a PA120.3 in my last build to 2 MCR320 rads in this one. Your work is exemplary and is the only current source of reliable testing that anyone seems to be doing on so many things.

You know you are good when companies that make parts start voluntarily sending them to you for testing, and they modify their products based on your work.

It would be interesting to see a retest of a PA120.3 on the same testbed, but I doubt there would be much difference in the results over what is already published.

[Me1tDown]

Very nice work Martin!
Now I have some more justification for recommending the Swiftech rads to people. I have one of the older design with the NPTs as well as two newer ones with G-1/4, but no test setup for comparing them.

[Martinm210]

I have been waiting on this as I just switched from a PA120.3 in my last build to 2 MCR320 rads in this one. Your work is exemplary and is the only current source of reliable testing that anyone seems to be doing on so many things.

You know you are good when companies that make parts start voluntarily sending them to you for testing, and they modify their products based on your work.

It would be interesting to see a retest of a PA120.3 on the same testbed, but I doubt there would be much difference in the results over what is already published.

Yeah, I'm still learning on this rad testing. Wish I had an environmental chamber or some way of controlling ambient, but I don't. I've retested a few oddball point here and there and generally come up with the same result within a couple of tenths of a degree so far, but there's still a few things I don't quite understand.

Fan RPM is a tricky one, you'd be really suprised just how much the RPM wants to change as the fan warms up running exactly the same voltage. Also voltage in a house fluctuates alot more than you would think and messes with things particularly my heat load jumping around 5 watts or more.

Then all these air sensors are extremely sensetive. A small thing like opening the door in the room will create an obvious blip in the results that I can't use.

Somehow the wife didn't take to kindly to my insisting she keeps the door closed for the sake of science!!..

And I havn't even asked her to wait on firing up the dryer yet....doesn't see know that the household voltage droop is causing problems!!...

[Pellepel]

Somehow the wife didn't take to kindly to my insisting she keeps the door closed for the sake of science!!..

And I havn't even asked her to wait on firing up the dryer yet....doesn't see know that the household voltage droop is causing problems!!...

thanks Martin good work!

[sirheck]

HA. It seems size doesnt always matter

[evil-98]

wow thats awesome! glad i invested in a few myself.

thanks for posting it up, very easy to understand.

[septim]

another nice testing done. keep up the good job martin... will be waiting for the full comparisons between PA, MCR, RS, TFC, GTX when you're done compiling them. Thanks a lot for this...

[voigts]

Somehow the wife didn't take to kindly to my insisting she keeps the door closed for the sake of science!!..

And I havn't even asked her to wait on firing up the dryer yet....doesn't see know that the household voltage droop is causing problems!!...

If your wife endures your hobby as she obviously does, she is a keeper for sure. Of course my wife thinks I'm nuts (and she is probably right) for tinkering with making my own case, etc., but she readily entertains my insanity and actually gets involved in the process. She fortunately however likes our quiet, fast, unique computer.

You could always find a reason to switch to a gas dryer, although currently finding a reason to switch to gas might be challenging.

[Chickenfeed]

The difference is actually quite small and it surprised me. Still think that < 1000RPM the Thermochills are the way to go though. I had a PA 120.2 and now have a Feser x360 and I have to say the Feser is extremely similar to the 120.3s as far as performance goes. I'd like to see how Noctua's do on the MCR320s myself as they have strong static pressure and should be quieter than those tested here ( played with some DSL12Ms and they are louder ). Obviously this only matters if your allergic to noise like me Subjectively I don't think you can really cool more than 300 watts or so before noise because an issue with a single loop. Seeing the MCR320 manage 570 watts is quite impressive though even if its with 3000RPM fans Didn't think a single loop could handle that.

[Tulatin]

The difference is actually quite small and it surprised me. Still think that < 1000RPM the Thermochills are the way to go though. I had a PA 120.2 and now have a Feser x360 and I have to say the Feser is extremely similar to the 120.3s as far as performance goes. I'd like to see how Noctua's do on the MCR320s myself as they have strong static pressure and should be quieter than those tested here ( played with some DSL12Ms and they are louder ). Obviously this only matters if your allergic to noise like me Subjectively I don't think you can really cool more than 300 watts or so before noise because an issue with a single loop. Seeing the MCR320 manage 570 watts is quite impressive though even if its with 3000RPM fans Didn't think a single loop could handle that.

You could also use the Kaze Ultra fans, and just undervolt them if you want quiet. If I recall, they're not "Silent" at the lowest rpm, though.

[ownage]

Martin, what about a comparison between Swiftech, PA and XSPC?

MCR320


XSPC RS360


The RS360 is 10 euro cheaper in Europe then the MCR320 and looks to be better on paper.

[Martinm210]

Martin, what about a comparison between Swiftech, PA and XSPC?

MCR320


XSPC RS360


The RS360 is 10 euro cheaper in Europe then the MCR320 and looks to be better on paper.

Thermally it would be PA>MCR>RS in that order, but fairly close results within 10-15% gaps between each other.

Lowest restriction is PA>RS>MCR in that order, but again not any real significant differences, they are all low in restriction compared to most CPU blocks.