Stupid Check

[orclev]

I've got an idea, but I've got a feeling it's a stupid idea so I want to run it past the forum first so you guys can all tell me how much of a noob I am. Ok, here's my problem, I've got limited space for a reservoir, and I only have room for one really big radiator, but there's spots I can put a couple small ones, and I'm making 2 loops. I don't think the small ones will be enough to cool my second loop on their own, and the big one is overkill for loop 1.

So, here's the idea, what if I put a Y connector in the line leading from the reservoir, and have it split going to each pump, then after it's run through all the blocks have another Y connector but this time bring the two loops back together to run through all the radiators and dump it back into the reservoir. This makes it 2 loops from the pumps until the radiators, and 1 loop for the radiators and the reservoir.

Ok, obvious problems. First, the split and join points are probably going to cause nasty restriction lowering the flow. Second, having all those radiators in series is going to hurt flow badly. Third, by joining the two loops together the thermal difference of the two loops will be averaged. It's unclear if the performance of all the radiators will be enough to make up for that.

Would it be better to just make it one giant loop and the pumps in series to overcome the added restriction, or should I just suck it up and figure some way to cram more radiators and reservoirs into the case?

[Martinm210]

It's better to make one giant loop and share several radiators with two pumps in series than to run parallel. Series is almost always going to do better than parallel because our loops are always on the more restrictive side.

[IanY]

Unless you want to run parallel RD30s, in which case them in series wouldn't be a good idea rofl

[orclev]

It would be dual MCP355 with petra tops. It is a quasi-dual loop, but at the same time a single loop which makes it hard to figure the performance on it, but I'm willing to take Martins word that the single would outperform the quasi-dual, after all he's certainly done enough testing to have a feel for these things. That's also why I asked on here, the thought occurred to me, but even to me it sounded kind of stupid.

Still not sure if I'm going to go with 1 giant loop, of if I'm going to try to figure out how to squeeze bigger radiators and a spare reservoir in.

FYI before someone asks, it's going to be cooling a overclocked Q9300, 780i everything, and a highly overclocked GeForce 9800 GTX, and possibly at some point in the future my RAM. For the radiators I was going to go with a Feser XChanger 360, and a pair of XChanger 120s. I want the CPU on the 360 as well as at least part of the MB, and the GPU and the remainder of the MB on the dual 120s. Problem is, I don't think the dual 120s give enough cooling for the GPU and NB and/or SB. I guess I could put the CPU on the 120s by itself and run everything else of the 360, but that still seems a little weak.

[fart_plume]

Check out Cyber Druid's buid on FlyingA2's cosmos he fit three rads in it................

[orclev]

Check out Cyber Druid's buid on FlyingA2's cosmos he fit three rads in it................
Trying to find that thread right now, but I also want to mention I'm trying to keep everything inside the case, I don't want some ugly radiator hanging off the back ruining the lines of the system.

[Martinm210]

You can always try it.

Series pumps are theoretically (I havn't actually tested series, just based on common pump engineering knowledge) going to stack the pressure side of the curve, so it's generally going to give you something like 15-30% more flow rate than a single pump, where parallel is usually going to net maybe half that unless you happen to have a very low restriction loop.

But regardless 15% change in flow if over 1GPM is going to be really hard to measure, likely less than a degree, so if it worked out for your particular setup and you just wanted to try parallel, go for it!

The only real downside is you lose the second real benefit of pumps in series and that's redundancy. In parallel if you loose one pump, it could be bad news where in series chances are you wouldn't even know unless you had some sort of other clues that one of the pumps was dead.

[Martinm210]

Martin...

If you run pumps in series, should you periodically unplug the pumps (one at a time) to make sure they are working?

I would probably just feel them for vibration, or monitor RPM if you have a DDC.

You could also do like Ira-K does on OCnet and run a pressure guage, that would be a good indicator that both pumps are working (Pressure would drop in half), and it would also indicate if a block is getting clogged(Pressure would increase if placed right after the pump)

[orclev]

If anyone would know this, I'm sure Martin would. How much is it going to hurt performance to stack a pair of XChanger 360 radiators on top of each other, perhaps with a extra pair of fans sandwiched between them? Thinking that since I only have 1 spot in my case big enough to hold a 360, but it has a good bit of depth to it, that I might be able to fit 2 radiators and 2 sets of fans, so that both loops can have a 360 in them. Would probably have to orient them opposite of each other though otherwise I think the fittings would be an issue. Then again, a set of 90 degree barbs might suffice.

[Martinm210]

Where would obtain said pressure guage?

Are they restrictive?

You can find them at McMaster, no more restrictive than a T-fitting, so no big deal really.


If anyone would know this, I'm sure Martin would. How much is it going to hurt performance to stack a pair of XChanger 360 radiators on top of each other, perhaps with a extra pair of fans sandwiched between them? Thinking that since I only have 1 spot in my case big enough to hold a 360, but it has a good bit of depth to it, that I might be able to fit 2 radiators and 2 sets of fans, so that both loops can have a 360 in them. Would probably have to orient them opposite of each other though otherwise I think the fittings would be an issue. Then again, a set of 90 degree barbs might suffice.

I wouldn't stack, the TFC was already using over 80% of the available cooling power the air can provide with the slower speed fans, so a second one using that same air probably wouldn't afford you more than 10%. They both need fresh cold air coming in.

Even the double thickness doesn't gain as much as you'd think. The slim rads use around 50-60% of the cooling power, so the extra double thickness is only providing another 20% or so.