Pump question

[AllRoy]

Here is a question for you, I have seen a lot of O/C say the bigger the pump the better but it seems to me you must balance flow (in terms of actually staying for long enough in the channel to transfer heat) with heat transfer. Am I an idiot?

[redleader]

The inside of many radiators (mainly heatercores) is much wider then then the rest of the system, 5/8 at least. So the coolant slows down as it passes through there anyway.

But you are right. Above a certain limit more flow does nothing at all. I think the idea is to make sure you go a little over that limit rather then be a little under it.

[AllRoy]

what size do you all recommend?

[dw]

Well i had a eheim 1048 pump.. that was
158 U.S. gal./h.

it broke.. and i went back to the shop.. switched it to a 1250 pump wich has
316 U.S. gal./h.

all this with 3/8" tubing.. didnt notice any difference at all.. so about 150 Gallons / H should be enough i suppose..

and a question from me to the more pro users on this forum :P should i switch to 1/2" tubing instead.. and get a maze2 with 1/2" fittings? would that improve cooling?

[Butcher]

You never get less performance from a too big pump so the bigger the better, this crap about no time to pick up heat is rubbish, it's a case of moving a little heat fast or a lot of heat slow, in the end above a certain point it doesn't make much difference (i.e. move twice the heat in twice the time). The main disadvantages with a bigger pump are noise, heat generated by the pump and of course size.

[AllRoy]

So when do the benefits outweigh the size of the pump??

[redleader]

As a rule of thumb some people say no less then 2 GPH per w produced by the CPU. I find that thats a little too low.

For a 3/8 inch system w/ TBird I'd say ~200-250GPH would be fine. Use 1/2 inch tubeing and more is always better. Use a cooler CPU and less will do fine.

In my case I went with the 317GPH 1250 because it was $9 more and I wanted to leave GPU/NB blocks as an option later on. I'd be worried aout using 3 blocks on a 1048.

[AllRoy]

Anybody ever use the Supreme Mag Drive Pumps by Danner?

[GuyBFF]

The Danner Mag-Drive Supreme Utility is probably the most commonly used pump available. I have a 1/2" "upgrade" Danger Den Maze 2, 1/2" tubing and a Chevy Chevette heater core with 1/2" barbs. I originally used a Mag-Drive 250, but when my friend became interested in it for his fish tank, I switched to a 350. There was no change in temperature.

In my opinion, a 250 is more than enough for a well flowing system (like the newer Danger Den products, or spiral mazes.. with a heater core style rad), but if your more restricted a larger pump may help push things along. One thing to note however is I have heard from others that in well flowing systems without pelts or additional blocks, they found temps lower with the smaller pump. Ultimately due to the lower wattage transfered as heat from the pump.

[jastrckl]

If you guys ever did any testing on your own or looked at someone else's, you'd see that there are not only diminishing returns after a certain point is reached in GPH, but performance _decreases_ after a certain point in most systems. The going theory is that when you pump the water too quickly, the radiator doesn't have time to dissipate the heat, and you end up recirculating hot water. I know it goes against reason, but I've seen real world numbers back this up. It's hard to argue against a graph plotted with flow rates and temperatures that shows a curve of performance that looks like a hill as opposed to a slanted line. If you want, I can dig up some links again to show you what I mean.

[AllRoy]

I dont know if this is the correct forum to ask but where is the best place to get them?

[Butcher]

how mugh of a degradation, you're likely to get some from a larger pump putting out more heat (this may be more significant than it seems at first glance), especially if it was a submerged pump.
I've seen data that used the same pump with a restriction in the pipe to simulate lower flows, they showed no degradation at higher rates, but above a certain point no real improvement - much as you would expect.
It goes both ways, less heat dropped in the rad at higher flow, but also less heat picked up in the block, so the net temp shouldn't be higher, lower if anything as the water is in the block less time so can't get so warm.
Additionally, if more flow is bad why do big fans work? Surely if you use a too big fan the air doesn't have enough time to pick up heat from the heatsink by your theory, it just doesn't add up when you think about it.

[redleader]

The going theory is that when you pump the water too quickly, the radiator doesn't have time to dissipate the heat, and you end up recirculating hot water.

Actually its because 40-50w pumps heat up the water nearly as much as the CPU they are trying to cool. Just as an SMP water setup will be hotter, a pump and CPU combination pushing nearly as much heat will also run hotter.

More flow at a given heat load will NEVER decrease performance.

[LiquidCool]

Remember air cooling and watercooling are the EXACT same thing, just using a different medium to cool. Here is a good comparison...

You have a regular heatsink with, as an example, a 20cfm fan. This can remove a certain amount of heat to keep your processor at a given temp.

Now you take this same heatsink and put a 100CFM fan on it. What happens? You get cooler temps as more air is being pushed over the heatsinks fins.

Now take this same heatsink and put a HUGE 500CFM fan on it? What happens? Nothing. Why? Because there is not enough surface area for a fan this large to affect it.

Now take the 100cfm fan and put it on a larger heatsink. Temps are lower than when using the smaller heatsink. Why? More surface area!

What does all this mean? Yes you will get a point of diminishing returns. BUT, if a system is designed right (with more surface area) then more flow will have a larger impact on performance. What we need is waterblocks with a larger surface area and also radiators with a larger internal surface area.