flow rate

[Isriam]

is there anyway to determine how flow rate affects temperature? i realize that it depends on a lot of things, but would it be possible to determine how flow rate could change possible loop setups?

say you have 6c delta with 1.96gpm, what if you lowered that to 1gpm, would it be possible to determine what your delta would be then?

[Martinm210]

is there anyway to determine how flow rate affects temperature? i realize that it depends on a lot of things, but would it be possible to determine how flow rate could change possible loop setups?

say you have 6c delta with 1.96gpm, what if you lowered that to 1gpm, would it be possible to determine what your delta would be then?

Sure, for a really rough guess, look up the c/w curve for your water block if there is one.

Estimate how many watts you are dumping into the system.
Here is one cpu calculator:
http://www.extreme.outervision.com/tools.jsp#cpuoc

Find the difference in C/W of your water block between the two flow rates and multiply by your watts in.

Take it a step further and do the same for your radiator and add the two.

It's not going to be much above 1GPM for todays blocks. It made a bigger difference in the older days when blocks were large open channels and C/W curves were much steeper above 1 GPM.

[The0men]

Can your Flow rate be too high? I have a pretty low restriction 1/2 system, and my radiator is a MC320QP, meaning its high flow low noise and restriction, But my Q6600 at stock seems to be about 50degrees.

I have done a nice job of lapping the Q6600 and the Fuzion block, and have tried reseating to see if it makes any difference, and even tried other pastes (using AS5 atm).

The flow is high enough to make the water at the back of the MCres push up to nearly the top, even though theres about 18mm to the top at the front where there is no flow effect.

[Martinm210]

Can your Flow rate be too high? I have a pretty low restriction 1/2 system, and my radiator is a MC320QP, meaning its high flow low noise and restriction, But my Q6600 at stock seems to be about 50degrees.

I have done a nice job of lapping the Q6600 and the Fuzion block, and have tried reseating to see if it makes any difference, and even tried other pastes (using AS5 atm).

The flow is high enough to make the water at the back of the MCres push up to nearly the top, even though theres about 18mm to the top at the front where there is no flow effect.

No, but you can have more heat dump than necessary from a pump that's too strong. Because the thermal curves flatten out so much with todays blocks, it's not impossible to gain performance by reducing something like a D5 from setting 5 down to setting 3 simply because the reduction in heat dump was more beneficial than the added flow rate. But this seems to be fairly rare, and I think more typical of smaller radiator setups and low restriction.

In your case with an MCR320, and a quad, I think it's just the quad.

You'll hear the term load in the forums and if you dig in the details that will mean about a dozen different things.

50C is not bad if you're overclocked on a quad and load means TAT 100% load. Orthos blend or small ftt is not as intensive as TAT. And many of the other stressing programs also don't load that hard.

I would check the coolant temp first compared to ambient, I'm guessing it's probably around 6-7 degrees above ambient depending on your fans and placement. If it's higher than that you might be drawing in hot air through the radiator or recycling hot air back through.

Next I would suggest at least the washer for the fuzion, and maybe a quad nozzle. Also if you don't mind the hassle, Coolaboratory Liquid Pro is good for a couple of degrees particularly on a lapped IHS, I got 2-3 degrees improvement just from going from AS5 to CLP.

Hope that helps some...

[Isriam]

awesome thanks.

the reason i'm asking is because i added a chipset block to my cpu only loop, and saw no temp loss on my quad. made me think that flow rate isn't as important as it sounds, even though we talk about powerful pumps and tops

[Xilikon]

The main reason for going with stronger pumps would be to have the headroom to add components or put more restrictive components. With a very weak pump, you are a bit limited with options and it's not a good thing being limited.

As long as you keep the flow above 1 GPM, you should be fine with possibly 1.5 GPM as a sweet spot.

[Martinm210]

Yeah, big strong pumps show their strength in really restrictive nozzle type blocks. Havn't tried myself yet, but I've heard more than one person say running dual D5 or and Iwaki RD30 at 21V or so was giving noticeable gains on blocks like the Ek Supreme.

I suspect the same is true for running a Fuzion with one of the smaller nozzles. Put the smallest nozzle and washer in a fuzion, it'll bring any pump to it's knees...:D

When stringing multiple block loops together, it's usually the added heat and stressed radiator that reduces performance moreso than flow rate. Not uncommon to gain 1 ro 2 degrees in water temperature simply by introducing another 30-40 watts from a chipset northbridge. Above 1 GPM I think the system is much more influenced by heat added, flow rate has minimal impact.

It's just easy to get wrapped up in flow rate, it is important, but strong mainstream pumps like the D5 and DDC's have made it harder to screw up now. 5 years ago everyone was running low head pond pumps and open channel blocks where the pumps just didn't have the pressure capabilities, and the blocks were much more sensetive to flow rate.

[FuriousSalesman]

Where can you buy CLP if you are in America?

[The0men]

Yer I have extremely little availabilty of parts being rural in Australia. Thank you very much for your advise Martin. However I am taking cold air from directly out of the case to the rad, and the case is no where near walls to get a recycled air effect. Of coarse it is summer here, and ambients are generally around 30, So my coolent must be getting windchill from radiator because my idle is also just over 30 on a 30 degree day. In retrospect the temps probably arnt so bad, is the idle to load delta that concernes me.

My load was 4 instances of prime 95 btw on small FFT's, I use this because under absolutely no other circumstance would I ever load the cores so hard, nor get so hot, so I figure thats pleanty of load.

[STEvil]

MartinM - it is important to point out that most people dont know the proper procedure for using TAT on 4 cores at once as the program by default can not do this and will not run multiple instances.

Here's how:

Open up as many command prompts as you have cores.

In each command prompt go to the directory you placed TAT in and type: "Run MeromMaxPowerVer0p3.exe 100 0" (maybe need to do /100 /0)

Next set the affinity of each running instance.


Linpack 64bit will create even more load than TAT.

[The0men]

MartinM - it is important to point out that most people dont know the proper procedure for using TAT on 4 cores at once as the program by default can not do this and will not run multiple instances.

Here's how:

Open up as many command prompts as you have cores.

In each command prompt go to the directory you placed TAT in and type: "Run MeromMaxPowerVer0p3.exe 100 0" (maybe need to do /100 /0)

Next set the affinity of each running instance.


Linpack 64bit will create even more load than TAT.


Sounds sweet, but prime is well good enough for me, Im not after a world record, just to check stability for gaming. And for that it does the job. Thanks for the info though.

[Martinm210]

MartinM - it is important to point out that most people dont know the proper procedure for using TAT on 4 cores at once as the program by default can not do this and will not run multiple instances.

Here's how:

Open up as many command prompts as you have cores.

In each command prompt go to the directory you placed TAT in and type: "Run MeromMaxPowerVer0p3.exe 100 0" (maybe need to do /100 /0)

Next set the affinity of each running instance.


Linpack 64bit will create even more load than TAT.

I didn't know that either...Thanks!

Still plugging along with my E6600...:D

[The0men]

I might turn my D5 down to 4, it will reduce the heat dump, and also it should mean the water stays in rad longer, perhaps I should also get some better fans, I'm only using the 3 default Thermaltake ones that came with my Armor case. I think turning it down won't effect the efficiency of the FuZioN, like the water literally swirls right up in the MicroRes.