Does loop order matter? (pressure)

I know the water temperature tends to be nearly uniform through a loop, so please don't tell me this.

Is the pressure through the loop also uniform? Since there is generally a reservoir involved I suspect it is not, and will be at a higher pressure closer to the outlet of the pump than partway through the loop. However it is quite possible that it is all the same pressure and I've just missed the point somewhere.

Reason for asking is that some waterblocks are pressure dependent, so putting my ek supreme immediately after a pump would be better than just minimising tubing length.

Cheers

(if interested, the actual motivation for the question is trying to work out where in a loop to put a second ddc. I'm inclined towards just before the cpu, suspended between an sli block and the processor. However it may make more sense to put it just after the other ddc)

Water pressure decreases round the loop with the most pressure at the pump outlet and least pressure at the pump inlet. Think of pressure as voltage in electronic terms and the pump as the battery and the other loop components as resistors. In a loop with a fixed number of blocks, it doesn't make any real difference what order the blocks are in as each block will cause the same pressure drop in the loop regardless of order, you'll still get the same flow rate (or current in electronics).

Some water blocks are impingement blocks which are deliberately restritctive to increase the local water flow velocity to help make the flow as turbulent as possible. This is because one of the features of smooth flow (laminar flow) is that the flow velocity drops to zero at the very surface, which if you're using the flow to cool means you have an additional layer for heat to pass through meaning you have a bigger temperature gradient.

In reality flow rate has very little impact on the effectiveness of a cooling loop, the main two things are, how good the waterblock is at removing heat from the CPU, and how good the radiator is at removing heat from the water.

It's good to know I was thinking sensibly after all.

Since I'll be using multiple non-impingement blocks and one which does deliberately do this, and I believe reacts positively to water pressure, does it then make sense to arrange the loop so as to have the restrictive block near the outlet?

I'm a bit unsure on your last paragraph, does it suggest that putting a block near the outlet will help but not significantly, or that adding a second pump is probably a waste of time?

Cheers Jokester

It doesn't make any difference what order the blocks are. Adding a second pump won't make much difference either. The best thing to do with impingement type blocks is to have them in their own loop, otherwise the reduced flow rate caused by the other blocks in a common loop will negate any benefit of the impingement design.

Ahh. Thank you, one loop with the ek on and one loop with everything else on looks like the way to go then. A common reservoir presumably works fine for this, I don't really want to thermally isolate the two.

I'm a little confused by why this is so. The reasoning behind a second pump in series was that it would roughly double head pressure, so assuming the rest of the loop has similar resistance to the ek supreme, the flow rate through the supreme would be the same as with them on seperate loops. Is the flaw in assuming the rest of the loop has comparable resistance to the supreme?

As you say it's a good idea to use a common reservoir as it allows an idle loop to cool the water from the other loop giving you better temperatures.

It's intuitive to think putting two pumps in series will give you better performance, and it might do in the case of an impingement block, but I've not seen anything to back it up. What good testing has been done on it shows that it makes next to know difference though for most cases. It would be interesting to see if it makes any difference with an impingement block though.

All sounds good from here then. I think I'm going to spend a long time testing various combinations, which is probably the best route to go down anyway.

I'm reassured that you also consider this intuitive, I'm starting to lose faith in being able to guess at how fluid dynamics behave. It'll be a while before any results turn up I'm afraid, still waiting on a new ek top at the moment. There are so few results available about high restriction loops, I'm hoping its because they're untried and not because they're invariably rubbish.

I believe that the best results will be achieved in a single loop, arranged as pump -> gpu block -> pump -> cpu block -> everything else, using a T line instead of a reservoir. It's hard to pin down why I want the gpu block between the two pumps, but it 'feels' right to have something between the two to break up the flow a bit.

Cheers man

Wouldn't having two pumps be counter productive. HAving the fuild pass through the rad too quickly not giving time for the rad to do its job?

That's one of the few things I'm not worried about. The radiator is always full of hot water, it'll cool just fine

elmuzzy66. said:

Wouldn't having two pumps be counter productive. HAving the fuild pass through the rad too quickly not giving time for the rad to do its job?

Click to expand...

If the water is going twice as fast through the loop it will spend the SAME amount of time in the rad... Think about it, its a closed system... twice as fast = visits the rad TWICE in the same time as the slower loop visits it once, hence same amount of time spent there....

Pressure in a liquid does not change anywhere in a system, with the exception of the resorvoir using an air gap.

Basic Physics 101

Liquids exert pressure on the sides of a containing vessel and on any body immersed in them, and pressure is transmitted through a liquid undiminished and in all directions

Click to expand...

Think of it as a hose, once the hose has expanded to accomodate the pressure inside, the pressure at the nozzle is identical to the pressure at the tap.

It does not follow from this that pressure is uniform through a loop,

Conservation of mass combined with velocity known to vary throughout the system implies pressure variations. Pressure is going to vary with distance from the ground for that matter. By all means teach me something different, but that quote is not sufficient

so tell me, in a closed loop, with a pump drawing exactly what it is pushing, where exactly does the acceleration and deceleration happen. even if you you are pumping high as long as the loop is closed, the downside will have syphonic action and maintain an equilibrium of pressure throughout the loop. The force of gravity on the upside is equal to the force of gravity on the downside.

But like I said, if you introduce a resorvoir with an air gap, that can add a random factor dependant on its position within the loop and its seal.

elmuzzy66. said:

Wouldn't having two pumps be counter productive. HAving the fuild pass through the rad too quickly not giving time for the rad to do its job?

Click to expand...

With a constant heat load the greater the mass flow rate, the smaller the difference between inlet and outlet temperatures. With a caveat or two, the extra flow increases friction raising fluid temps and the second pump probably dumps more heat in the coolant than any tiny improvement from flow.

Dr Who said:

so tell me, in a closed loop, with a pump drawing exactly what it is pushing, where exactly does the acceleration and deceleration happen

Click to expand...

There is a pressure drop across every millimeter of tubing, every bend, every block. They all add up to the head the pump has to overcome. Bernoulli applies with the heights, the total pressure is equal to the static pressure plus the dynamic pressure. The impeller of a centrifugal pump creates pressure by acceleration, fluid accelerates flowing radially outward over the blades. At the outlet there is a net gain in potential energy (pressure) when velocity is reduced.

Thanks fornowagain, that's a rather better explanation than I could have offered.

18W maximum isn't so much heat to add to a loop to find out if it cools better though. Martin's flowrate spreadsheet predicts a pretty dire outcome for me with only one pump, though I am certainly going to try with just one before I run out and buy another

It really does work remarkably like DC electric circuits.

fornowagain said:

With a constant heat load the greater the mass flow rate, the smaller the difference between inlet and outlet temperatures. With a caveat or two, the extra flow increases friction raising fluid temps and the second pump probably dumps more heat in the coolant than any tiny improvement from flow.
There is a pressure drop across every millimeter of tubing, every bend, every block. They all add up to the head the pump has to overcome. Bernoulli applies with the heights, the total pressure is equal to the static pressure plus the dynamic pressure. The impeller of a centrifugal pump creates pressure by acceleration, fluid accelerates flowing radially outward over the blades. At the outlet there is a net gain in potential energy (pressure) when velocity is reduced.

Click to expand...

I don't know where people get these ideas, bends increase the speed of the water. Not that speed matters much anyway, i've seen a few people running naturally convective loops.

Really? Please explain how the water gathers speed? Does it somehow stretch? Or does it leave a gap behind it?

Only if the diameter of the tubing narrows at the bend would it increase speed. Any restriction produces friction which removes energy from the water slowing it down.

Careful super, looks a bit like you're contradicting someone who's fairly obviously on the ball here. I'll give explaining the source of 'these ideas' a go. The following is not rigorous.

Consider the following.

Tube with water flowing down it at constant veloctiy, fed from an infinite reservoir and going to an infinite drain. It has momentum associated with it proportional to mass of water and velocity.

It then goes through 90 degrees. Momentum is conserved, but now it's all going perpendicularly to the original. Therefore a force acted to change the direction of the water flow.

This force is fairly obviously linked to the bend in the tubing, the tube walls exerting a force on the water which causes the change in direction.

Since the force inevitably carries with it frictional losses, the bend warms up and the kinetic energy of the water decreases. So right there you have energy moving out of the water flowing. These frictional losses are why you don't want 90 degree bends in your loop.

The force to turn the water through 45 degrees is obviously rather less, so frictional losses from a 45 degree bend are less than through a 90 degree bend.

If that isn't sufficient now might be a good time to work out the loss in flow rate for a given volume of water moving through a 90 degree bend, it was part of last terms course but I can't remember the derivation now.


If this is too difficult to grasp, have a play with this zipped excel file and trust that the author knows exactly what he is doing.

Out of curiousity, do you have a water cooled system or is this conjecture on your part?

Well beaten to it by Mike, serves me right for writing too much



-----------------------------------------------------------------

How terrible is inadequate flow for a loop? Martins sheet had a guess at around 1/3 gph for me with one pump iirc

my brain hurts