Martinm210

Having had alot of people asking me how much is "enough", also seeing people recommend pump upgrades over radiator additions, I thought it might be helpful to put some scale to flow rate effects.

This isn't any good for anything other than getter a rough understanding the differences in thermal response relative to flow rate. It will be a little different with every block and radiator, but for general feel..I thought it might help. The TIM number is completely assumed to give a ballpark result, but since it's irrelevant to the differences I was after, I thought it was ok. So with that... I took the fuzion stock curve published along with the published PA120.3 and assumed a .07 TIM c/w just to get a number and finally figured out what the delta was between .5, 1.0, etc GPM flow rates.

Fuzion C/W Curve:



PA120.3 C/W Curve (Chose the nexus fans at 12V0:



As I've been recommending, above 1.0 GPM the gains are fairly small, and I would say smaller than gains you would get from doubling your radiator capacity almost every time. Also the difference is very scaled by heat load, so I did both a run with 110watts which would simulate a C2D overclock and 200watts for an overclocked quad.

Not sure it helps, and if there is something out of place, let me know. I thought it was helpful to put some scale to the differences.

__________________
Flow Rate Estimator | Radiator Sizing Estimators

ira-k

Looks good Martin.....Stick your smallest jet in it and see what it does...I'm still going to say to stick at 1.5gpm or so for flow quality, to make sure it doesn't go laminar on you,and I have to stay well over 2gpm or i would have nightmares...

Like you said though the blocks and rads you use will have a lot of effect on it...And if you don't have enough rad thats the first thing you would want to upgrade for sure....

Thanks for the testing....

__________________
**Water Cooling Essential Threads** Overclock.net's Downloads Section**TEC Cooling Essential Threads

**E8600 @ 5.3Ghz 10x **E7200 @ 4.5Ghz 9x **5000BE @ 3.5Ghz 10x**150 Opty @ 3.2Ghz 12x**

Martinm210

Yeah, especially the stinger v8. With the open channel design, it probably gains alot more with higher flow rates than something like the fuzion.

The nozzles are tricky though, because you trade velocities overall for increased velocities only at the nozzle. I bet a single core gains the most from a tight small nozzle.

__________________
Flow Rate Estimator | Radiator Sizing Estimators

ira-k

Yeah the jet barb in them likes high head and high flow....Really the Fuzion is optimized for low flow, it will still cool better with more flow but its not needed to perform well...Most of the newer blocks are to, but I've been seeing a return to high resistance blocks like the EK Supreme or the jet impingement Fuzion GPU...Now that EK would be a good one to play with modding with that jet mid-plate plate in it.......


Have you tried a X or a cross jet in your Fuzion yet? I bet that would give pretty good coverage and depending on the width of the cut out a lot or moderate turbulence...

__________________
**Water Cooling Essential Threads** Overclock.net's Downloads Section**TEC Cooling Essential Threads

**E8600 @ 5.3Ghz 10x **E7200 @ 4.5Ghz 9x **5000BE @ 3.5Ghz 10x**150 Opty @ 3.2Ghz 12x**

Martinm210

Not yet, but I'm thinking the MC-TDX has even more potential with a nice large array of smaller nozzle holes. It has a fairly large recessed area for a nozzle plate, and I think to really take advantage of nozzles it needs to be an array that covers the entire core area. The fuzion is a bit limited with the 3/8" ID unless you bored out the midplate to accept a larger nozzle.

I just didn't like the .9GPM or better rule so I wanted to put some numbers on the c/w curves. You do still gain almost .5C from 1 to 1.5GPM.

I've heard the EKsupreme still can gain about 1-2C from using a really strong pump, so I'm thinking the straight microchannels don't cause as much turbulence as the pin matrixes.

I'll have to do some more block tinkering after I get the pump and radiator stuff done. Lots to do there still.

__________________
Flow Rate Estimator | Radiator Sizing Estimators

TonyL222

This info - coupled with your Flow Rate Estimator - is very useful in assessing the impact of added components and plumbing to a loop. As long as the resulting flow rate is above 1GPM (though like Ira-K, I'd just feel better with 1.5GPM) you should be okay.

ira-k

The new TDX is a sweet looking block, I think they have another winner there...It wasn't long ago that the Maze and TDX were king of the hill for a long while...

That EK kind of surprised me that they went back to a fin base on...Seems to be working OK, I think I'd have to pin it up a little right in the center of the chip to see what it would do...Have you ever seen the internals on a Aquaextreme MP-05? Its wild with 3 or 4 hundred little micro-pins and a jet plate over ...They are Storm restrictive but used to be a great block to, they sold mid-plates for them also like from a 1-5 restriction...

Hey on those Pana's Martin any that have a BX at the end of the designation have a rpm wire...I love those 115's, I have 4 on my rad now...

__________________
**Water Cooling Essential Threads** Overclock.net's Downloads Section**TEC Cooling Essential Threads

**E8600 @ 5.3Ghz 10x **E7200 @ 4.5Ghz 9x **5000BE @ 3.5Ghz 10x**150 Opty @ 3.2Ghz 12x**

Martinm210

Yeah, I've got a bunch of CPU thermal testing sitting here too. I just don't like tearing into my own system so much, so I either need to get a test bench going or that die sim.

Ok, I'll probably pick up a few of those 115cfm panos for the high speed fan testing. Yate loon D12sl12's will cover the low speed, then I just need to grab some mediums. Seems the Delta's are common for the tests, just not sure anyone is using them. Maybe I'll just go with the high speed yates, they are probably pretty common..

__________________
Flow Rate Estimator | Radiator Sizing Estimators

ira-k

What about those 86cfm Pana you have, they don't have a rpm wire do they? That would be a good medium cfm fan, a lot of guys use them...Shipping kills on 120x38mm fans though, sux...

Yeah I wouldn't tear down my main rig for all that and I love tearing one down......Thats just to much though...Get your test bed set up nice and how you like it first...I always have to slow myself down is why I say that, I'm not a patient person...Have fun with it, thats what its for...

__________________
**Water Cooling Essential Threads** Overclock.net's Downloads Section**TEC Cooling Essential Threads

**E8600 @ 5.3Ghz 10x **E7200 @ 4.5Ghz 9x **5000BE @ 3.5Ghz 10x**150 Opty @ 3.2Ghz 12x**

Martinm210

Yeah, those were good fans, I still have a few of them, but I gave most of them away to a guy on XS that was just getting started and needed some help

I'm still thinking die simulator though, alot of people don't like them, but that would be more fun for me..

__________________
Flow Rate Estimator | Radiator Sizing Estimators