Edited: I need your help PROs (running 2x MCP350 in parallel)
 

hello guys. I will try to make this as short as possible.

I have been watercooling my 3.2E along with 6800 ultra for a couple of months already using the damn noisy MCP650, well I just got so sick and went ahead and ordered 2x MCP350 to run instead of that noisy thing but my problem is my blocks and rads (2x dual heatercores) are 1/2" ID so I did a little draw for you guys of what I had in mind to make these pump work with my system without converting the whole system into 3/8" ID, let me know if that will work please. oh I will be running them in parallel just for the fact that these pumps' heads are great but flow is a bit weak for me, so running them in parallel will up the flow a little bit, am I right? oh and why 2, well let's say I don't think one will be enough to perform well in a 2 blocks/2 dual heatercore/res/12 feet tubing system (rads/res/pump are mounted externally)

http://img139.exs.cx/img139/6838/untitled8dq.jpg



Edited:

just to clear things up a bit, here is how the flow goes.

RES>Pumps>Rad1>Rad2>CPU block>GPU block>Res.

flow VS tubing

Res---------3/8"--------pumps(parallel with each pump getting watter from a 3/8" barb from the res)-----------3/8"------Y connector------1/2"-------Rads------1/2"-------Cpu block--------1/2"-------Gpu block----1/2"-----Res again

thanks everyone

I'm no expert, but I think you would want to rearrange the system so that your res is the highest point, and have that feed the pumps rather than trying to push the water through the res. Also, the Y connector from 1/2 to 3/8 for the pumps seems like it would choke the inlet on the pumps. The last thing I'd suggest is to put the CPU block before the GPU so that your CPU gets the lowest temps. There's my $.02 based on what I've seen around here.

Temp difference in water between blocks is like 0.2c. Route them whatever way gives you less bends and less tubing (less resistance), order does not matter.

As for the Y choking? The Y he listed is ideal. Barbs on the MCP350s are 3/8" You dont have much choice. There was mention of how to mount them though, IE ontop of each other or something. I didnt really read it but I think it made a difference in either ease of usability or actual performance. Stacking them on top of each other and having that Y as close to the inlets as possible makes sense though =)

Ok so running the 2 pumps with the Y connector for the outlet of the pumps is ok then, right? and for the inlets I was going to use another Y connector but figured it would be better for the pumps to be sucking water from the res each on its own is better than having the 2 pumps sharing water from a second Y connector, Am I right ?

the order of the parts is just fine guys,I mean I have been running like that for a while and my temps are good as well as my OC (3.2E @ 4.2 Max and 3.9 24/7) (6800 ultra @ 460/1.19 24/7 and 480/1.25 Max)

so try to keep the Y as close as possible to the pumps? same goes for the Res huh? (keeping the res as close as possible to the pumps too is what i mean)


thanks for the input guys

Having them each suck from the res is probably a good idea. I thought the Y was your inlet. Chances are having that Y close to the outlet or close to the rad wont make much difference at all. Might as well try both though =)
I think youve already seen the couple of people over at ocforums that actually have them in their hands right now. So Id ask them.

A flow indicator on the diagram would have been rather helpful. I guess I made the same assumption about the rads being the inlet as well. With the arrangement that way it makes much more sense.

series is better.

look into the Eheim 1250, i just got mine and the thing is dead quiet.

Fixed, now it should be clear.



this might be true for most pumps, but not these pumps. these pumps have great head but their flow is just weak, so running them in parallel will not add much head to the loop, but will almost double the flow. where if I were to run them in series, the head in the loop will be a lot higher than running a single pump, the flow will not go up much tho.


I already bought the pumps, plus I don't think Eheim will outperform theses little monsters in a big loop :)


I guess I will go with the setup I drew earlier, unless someone tells me of a better way.

thanks everyone

You sure?
Look at the P/Q curves. Look at the resistance of your components. When i look at your components and the MCP350 P/Q curves, i don't see parallel giving your better flow...

if you feel you must go paralel...
consider running it completly that way...
resevoir -> pump->heater core -> CPU block -> resevor..
ues 2 inlet and 2 outlets on the resevoir for each loop...
you seem to have 2 of everything already...

my 2¢

If you check out the P/Q curves, there is a crossover point in restrictiveness where series will perform better, and it looks like it is around 2.5 mmh20.

Since we don't know what blocks he is using (hasnt said, AFAICS) I am not positive, hence me saying:

However, I had not seen those test yet - thanks for the leak. The "knee" for series > parallel may be furthur out than i thought.

Of course, in the later tests in that thread (multi-block), series and parallel are basically giving the same temps.

I am using Swiftech MCW-50 and Swiftech MCW6002-P blocks. like I said b4, I have been using MCP650 which happens to have less head than the MCP350 and my temps are great as well as my OC, and since we all know MCP350 does not have a great flow, I thought parallel would be the way to go (head would be the same as running one MCP350 but flow would be close to doubled) am I right?

don't hope doubling your flow ! Resistance will increase faster f(Q²) than flow... As you can see on the link you posted, 2 mcp350 will gain only about 1.5°C on 1 mcp350...

Only you can say if 1.5°C is enough to spend about 80$, but don't expect more gain.

Sorry Althornin, I didnt read very closely. Just saw that XTATDSM was talking about stuff directly from that thread Im pretty sure, just glanced over both.

I guess I just dont think like you guys do and dont see the hype over this thing. If you need to get two to get the performance of a mcp600 or greater you are spending 160$+... Id rather get an Iwaki. Guess thats just me as I dont mind making an external box.

If I was trying to cram a multblock setup in a mATX case or something else really small then sure I dont see many other choices, single 1048 too weak(maybe), csp750s will just break, etc.

Ok so I finally got the pumps. first I hooked them up in parallel (like in the pic) and got the same temps as I did with my MCP650. then I hooked them up in series, not much different here either :rolleyes:


maybe the different is not enough to be shown with the crappy CPU sensor, anyway, should I keep them in series or connect them in parallel?

I must say, if your measurements show them to be the same, then does it matter?
Just do whichever one is neatest, or strikes your fancy.

Coould you try with only 1 to see the influence of the 2nd pump in your system ?

I just unpluged one pump and leaving the other one working just to see if the second pump is doing me any good.


why is it so hard to answer a simple question, never mind what my tests show, or why get 2 pumps instead of one. please.

let me repeate my question again.

in a 1/2" ID system with MCW6002 block + MCW50 block + 2x dual heatercores + Res + 12 feet of tubbing, and 2 MCP350 pumps, which is a better way, to run them in parallel or series?????? and if you are nice enough, you could go ahead and tell me why :)


thanks people

1 pump
simple

ahaaaaaa so what you are saying is that running one MCP350 with my system is better than running two (parallel or series) ummm do you care to explain how and why? it might be simple answer for you, but I just don't see how. thanks

Look, what we are trying to say is the diference is so small that you shouldn't even be thinking on getting a second pump. Too much expense for the gain.
However, you have already bought the pumps, so the question you are asking relates to running parallel or series.

Several fellows here could come and answer you with a bunch of calculations of restrictions and such. Meaning, you have to check your pump's pressure curve, and your block's total pressure curve, and check wich situation would be better for you.

You have to get the highest flow possible. Try series and parallel. The one who gives you more flow is the best.

satanicoo, thanks a lot for clearing that up for me bro.

We have a theorical answer for difference between 1 and 2 pumps, but it's theorical, so if you could tell us the difference you measure with your system with only one mcp350, we could know if reality is close to what we are in mind.