Pump heat dissipation

[BadNeurons]

Anyone have information on just how much heat is added to your system by the pump?

Seems like even if the pump was working at 100% efficiency (electric-wise), you'd still have heat dumped into the water by accelerating all of those little water molecules.

The PondMaster 350 specs say that it uses 35W, and I'm curious to see if anyone has done any non-theoretical measurements.

[redleader]

35 watts would be added to the water minus the small amount radiated to air (maybe a few watts). All energy becomes heat eventually, your pump is no different.

[JasonC]

I think most people counter this by running their setup from the pump to the radiator and then straight to the water block so that the radiator will disperse as much heat from the pump as possible. Along with all the heat it got from the CPU.

I'm probably stating the obvious as I'm a newbie to water cooling. Just what I gathered.

Jason

[imagex]

from now on i will only use inline pumps like a magdrive, i have gone through 2 pumps in 6 months a 1000 gph & a 500 gph " sump type". the thing is that u can't tell if the pumps shaft seal is leaking!!mine didn't show untill the pump failed!! thank god i notcied both times but it could of wiped my system out.I tryed to save money and found it pays to get a magdrive or atleast a inline /air cooled pump "u can see if it starts to lead on a in line, just make sure it's at the bottom of ur case!"

[GuyBFF]

Mag-Drive's don't have shaft seals!

The only seal is the head unit (this is what can leak in the case). The rotor is magnetically driven and therefore no shaft between an electric motor and the rotor to wear out.

In-Line Shaft Drive units will have a head seal and a shaft seal (as they are shaft driven). When the shaft seal wears they consume liquid and burn out.

Definetly get a mag-drive as they are supposed to be twice as efficient (which means less heat in the water), and when the rotors begin to wear they get louder. They really can't fail until the rotor falls apart!

[Nordic]

I have an Eheim 1048 mag-drive pump in my in-line system. The first thing I did when I assembled the system was to run the pump for 48 hours to check for leaks. I had no leaks luckily. I checked the watertemp afterwards, it was 25 degrees in the loop and 20 degrees ambient temp (Celcius). I did not run any fans on the radiator during this time, the rad is the smallest model that DangerDen sells - fits my quiet Papst 92 mm just fine, mounted with a cardboard shroud in a sucking configuration.

Just as proof of how efficient this system is I tested to run the system with the pump turned off until the temp rose to 50 degrees and not even 5 seconds after I turned the pump on the temp probe on the waterblock read 30 degrees. After 10 seconds it was 25 degrees.
This is on Celeron-1 500@563.5 MHz(2.0V) under full load (folding@home).

[imagex]

i konw that mag drives dont have shaft seals i was talking about air cooled shaft drive"inline" pumps.

[gmat]

Good mag-drives like Eheim's have kinda small 'fins' on their side where the engine is located. Believe me a lot of heat goes through there.
Since there's no contact between the propeller/wet part and the engine part, very little or no heat goes into your water.

[deeznuts]

Originally posted by SureFoot:
Good mag-drives like Eheim's have kinda small 'fins' on their side where the engine is located. Believe me a lot of heat goes through there.
Since there's no contact between the propeller/wet part and the engine part, very little or no heat goes into your water.

i dont' know about that. i know a couple posters on other forums, who ran their setup without the fan on, outside the case to check for leaks, they said their water heated up really fast.

deez

[gmat]

Yeah did they say "without the CPU running" ?I doubt. I've got a rather powerful pump (about 35W) and i ran it WITHOUT the cpu and withouta any fan for hours, checking for leaks. All i had was like 1°C more heat. So long for the "really hot"...
Speaking of "really hot" i couldn't touch the pump motor casing without burning myself.

[redleader]

Deja Vu . . .

Water has a very high specific heat, 4.18 J/g, IIRC. Couple this with the effects of the radiator and its going to take a long time for the temperature to finally peak. On top of that, we are only talking about a difference of a few C before the radiator's passive cooling ability equalizes the pumps emission- which is less then the margin of error on typical probes. With that in mind I wouldn't draw any conclusions based on indivdual's tests.

All i had was like 1°C more heat. So long for the "really hot"...

My entire setup (50w Duron, 15w GeForce, 30w pump) only heats the coolant about 3C above ambient temps under full load. 33% is nothing to laugh at.

[Brad]

my 660gph's are rated at 40w. They get warm, but certainly not enough to actually worry about.

if you mount a pump inline, make sure it isn't going to leak, or overheat

if you mount a pump in a res, make sure it's heat can be removed in the rad before the water gets to the cpu

[gmat]

Oh yeah, of course, i forgot: mount the pump inline if you don't want its heat to go into your water.
About my 30W+ pump I think a large part of its heat goes into air, thats why it barely changes anything to water temp.
Now, i've no peltier, and i've noticed that ambient temperature will greatly affect the core CPU temp. Just a 2°C drop in ambient can cause a 4°C drop of your CPU temp... Well, thats in my case, where i got an oversized rad.