Will this work? (contains an image, no 8000bps)

ubersonic

ubersonic

ubersonic

Soldato
Joined
26 May 2009
Posts
22,177
Yeah so basically im thinking of changing my loop to this:

loopx.png


For reference the two cooling towers are Zalman Reserators that have had the pumps removed, basically what im thinking is if I link them like this (running off an external pump in the PC) it shouldn't be possible for the pump to pull more/less water out of Res A than its dumping into Res B so the levels should remain equalized.

The thing im not sure about however is heat, if the hot return is being dumped into the first Res will the temperatures equalize too or will the first one just take all the heat and go into meltdown? I would *assume* that it will heat up faster but as the water in it heats up and therefore the water equalizing to Res B heats up they should plateau together or with Res A slightly hotter, but im not exactly sure about this, I know other people on this forums have linked their Res's like this so any information would be appreciated, Thank you.

And as I know somebody's bound to ask, the CPU is a Xeon E3-1245V2 (basically an i7 3770 with a faster GPU a cooler name and a smaller price) on an EK Supremacy (Nickel/plexi) block, the GPU is a HD9750 on a EK FC7950 (nickel plexi) block and the pump is a Laing D5 with the new EK top (acetal).
 
It's going to be a pain to bleed, and ideally you'll want the pump following the radiators (as they're your reservoir).

The water will equilibrate in height only if the rads are level, the temperatures round your loop should remain fairly constant.

I'd be very careful of getting a hydraulic lock in the reservoirs as the air will expand as it heats up, personally I'd try and keep the system as air free as possible.
 
Yeah thats the order it will be in, the loop I currently have is basically the same but with only one Reserator and no GPU block, bleeding isn't an issue as I have a suction pump I use to vac out the Res and then the lines.

Sorry for sounding dense but, hydraulic lock? if its what I assume it to be the cooling towers have a tiny air hole in the top which I believe is there to prevent this.
 
hmm, not sure I like the sound of a hole in the top, if the website I am reading is correct, these zalman's store 2.5 litres, if theres flow restriction, the hole at the top might just start leaking. Trying to think how the internals work, inlet at the bottom and outlet at the top which returns to the bottom port. So long as the holes are level and theres no flow restriction, they would need to be the highest part in the loop, you might get away with it.

I would be tempted to just run cpu and gpu on different loops and pumps.
 
Have done this with one reserator with a 3 block loop and a large rad, had no issues if anything it was easier to fill bleed and drain.

Edit: install a snorkle tube on the first res on the return line, should curb the heat build up.
 
Last edited:
andybird123 said:
could you put a Y in before and after each of the reserators so that it feeds to and pulls from each one equally?
Click to expand...

Like so?

loope.png


Im now planning on doing this as although the original plan works and keeps the CPU/GPU at low 50's under full load I don't like the way tower A gets a lot hotter than tower B.

do the tubes from the "Y"'s to the towers have to be even length or is the a bit of leeway available? never ran anything in parallel before.
 
any length of tubing will have a resistance, so keeping them as close as possible to the same length will be advantageous, but we're not talking about a lot. It's gonna depend on the diameter of the tubing, larger the better, but think you could get away with a few feet before it became an issue you have to factor in.
 
If the flow paths through the parallel radiators offer different resistance, you'll get different flow rates through them. You need significantly different resistance to get significantly different flow and the resistance in either route is going to be dominated by the radiator. I wouldn't worry about getting them bang on.

That the first radiator gets so much hotter than the second suggests the second isn't doing a great deal, it'll be interesting to see whether changing to parallel lowers temperatures or only balances temperature drop across the two towers. Thanks for making this thread.
 
Gonna be setting this up tonight as ive managed to get a hold of a couple of Y connectors for 3/8 tubing.

Just thought I would ask once again though if the tubes from the Y connectors to the towers have to be equal length? as the last thing I want is for a tower to overflow. Atm I'm planning on having one side of the Y ~10cm longer as that way I don't need to rearrange the towers. That is of course unless somebody says otherwise.
 
10cm is absolutely fine, I would have thought you'd need certainly more than 30cm more like 1m before it became an issue - straight tubing is very low resistance compared with fittings / radiator etc.
 
Last edited:
ubersonic said:
Just thought I would ask once again though if the tubes from the Y connectors to the towers have to be equal length? as the last thing I want is for a tower to overflow. Atm I'm planning on having one side of the Y ~10cm longer as that way I don't need to rearrange the towers. That is of course unless somebody says otherwise.
Click to expand...

It might be worth trying it out away from the pc first. There would be a difference, but it would only be by the amount the extra bit of tube hold. Essentially the towers will start where the Y splits, so the longer tubed tower will hold less water(as some of the water is in the extra bit of tube) (minimal amount) than the other tower. Each split will be equal. I hope I made sense.
 
I see, I think :P my worry was that instead of the water returning to the towers splitting evenly it would end up with more water taking the shorter hose and thus resulting in the water level in that tower rising.
 
Ok update as promised, first a couple of pics:

img0871z.jpg


img0863ol.jpg


img0869tz.jpg


And the results, well temps now seem to be distributed more evenly between the towers which is what I was aiming for, temps on the components in the loop haven't changed but I wasn't expecting them too as the loop was already dissipating the heat fine the radiators are just doing it more evenly now :P
 
You must log in or register to reply here.
Back
Top Bottom