How low is too low flow?

[SparksNelec]

I recently purchased this flow meter by Innovatek:
http://www.frozencpu.com/ex-flo-03.html?id=33x8r9rF

My setup consists of all Innovatek products: a i-Cool Rev4 CPU block, a Graph-o-matic Rev2 (vid card), Graph-o-matic Rev3 straight (NB), a single 120mm radiator, and I believe the first generation Ehiem 12V pump all connected with 8mm tubing. The easiest config for my case ends up like this:
pump->NB->CPU->Vid Card->Rad->Flow Meter->res->pump

The motherboard reports a speed of 1080RPM, which according to the formula on that page give ~0.4l/m. Seems kinda low. Maybe I have too many blocks for one pump Maybe someday I'll try measuring one block at a time but not now since I just got everything back together.

Should I reverse the flow so the output of the pump goes to the flow meter first?
Would it be worthwhile getting a stronger DC pump? I don't want to go AC.
Temps are OK but nothing too great. Keeps everything under 50C loaded and right around 40C idle.

Athlon XP 2400+ Mobile 35W
Abit AN7
9800Pro

 

[Black-Falcon]

I'm not too sure, but if your happy who cares.

 

[J-Pepper]

I'm not too sure, but if your happy who cares.

I agree...

 

[Pooky]

If the pump is an HPPS (the first ones were simply 12v Eheim 1046's the HPPS has more head pressure), you should be able to overvolt it (effectively turning it into an Eheim 1048).

 

[Top Nurse]

There is no such thing as too low of a flow providing it does what it is supposed to. Your figure is a little low, but not entirely unheard of. If you didn't have the flowmeter you wouldn't even know so the big question is whether otherwise you are happy or not. All a flowmeter does is give you an indication if what you do to modify a system is going in the right direction or not.

 

[Cathar]

Anything less than around 0.02 LPM is too low, period. Can still cool quite effectively with 0.04LPM, given a waterblock that's designed to work properly with these flow rates, and no, that DOES NOT include anything on the market today from any country.

As far as most typical water-cooling setups go, generally anything less than 0.08LPM is too low.

 

[nealio]

I think the determining factor is the temp. If you happy with that temp then who cares? If you want to lower it, or make your system more effective, then I would say either get another pump or go dual loop.

I think that is a lot to put on one 120 rad, but I'm not [H] gawd.

$.02

 

[Top Nurse]

Anything less than around 0.02 LPM is too low, period. Can still cool quite effectively with 0.04LPM, given a waterblock that's designed to work properly with these flow rates, and no, that DOES NOT include anything on the market today from any country.

As far as most typical water-cooling setups go, generally anything less than 0.08LPM is too low.

I didn't even know that there were cooling systems out there that even ran at 0.08 LPM. Even the old Exos ran at about 1 LPM. So are you saying you might have something up your sleeve that will cool at really super low flow rates?

 

[SparksNelec]

Thanks for all the great replies!

If the pump is an HPPS (the first ones were simply 12v Eheim 1046's the HPPS has more head pressure)

I always wondered if my pump was HPPS or not since I saw them appear shortly after I bought mine. I'm sure I have the first one then. I know the HPPS uses a different circuit board. Anyone know if I can slap that in my pump and basically make it into a HPPS or will I have to buy a new pump altogether?

I think that is a lot to put on one 120 rad, but I'm not [H] gawd.

I agree, but I'm not a [H] gawd either. I like my case and I'm not ready to cut a big hole in it to fit a dual. I'd love an external passive rad but they're just too spendy ATM.

There is no such thing as too low of a flow providing it does what it is supposed to. Your figure is a little low, but not entirely unheard of.

Thanks TopNurse. I think I'll keep an eye on it for awhile and see how it does over time. Maybe next time I change out the water I'll try something different; perhaps with a new pump.

 

[Cathar]

I didn't even know that there were cooling systems out there that even ran at 0.08 LPM. Even the old Exos ran at about 1 LPM. So are you saying you might have something up your sleeve that will cool at really super low flow rates?

No, I was just calculating the theoretical limits with respect to the thermal mass of water, and assuming that around 80C was a maximum workable CPU temperature, and that our presumed waterblock was 100% efficient in transferring the heat energy into the water flowing through it.

There are research papers that revolve around this sort of thing, and they do this sort of stuff using either true micro-channels (as in 25-50 microns, or 0.001 - 0.002", channel widths), or true micro-jet impingement (similar sorts of jet diameters).

Seems to be a whole field of research out there trying to achieve the best possible with the lowest possible water-flows. I don't see too much point to it though - once you get THAT micro, things start clogging as soon as you look at it.

As a rule I generally try to stay in the "mini" realm, or a factor of 10 larger (250-1000 microns, or 0.25-1.0mm), just for those reasons.

 

[mota]

Huh? My true microchannel block (lytron cp-20) never had problems with clogging. I didn't even use any sort of filter.

Seems to be a whole field of research out there trying to achieve the best possible with the lowest possible water-flows. I don't see too much point to it though - once you get THAT micro, things start clogging as soon as you look at it.

 

[Cathar]

Huh? My true microchannel block (lytron cp-20) never had problems with clogging. I didn't even use any sort of filter.

That's not a true micro-channel though. The Lytron CP-20 still has ~10x larger channels than what I'm talking about with these research papers that talk about channel sizes measures in small number of microns.

Also, you probably took the adequate precautions to not clog it up. Some people have no idea....