How much pump is enough? How much is too much?

[BillA]

I'll not critique an article I've not seen

some simplifications are apparent, which could be very significant depending on the system design
- in case
- in case, case air exhausted through the rad
- external cooling box
- actual flow rate

a biggish metal frame pump in a closed case will do more than just heat the water

[BillA]

hey pH / JoeK
can you change Cathar's title to "Cooling God" ?
sorry Stew, could not resist (better you than me, lol)

[Blackeagle]

In so far as the amount of power used, and heat added to water, I see what you are driving at Cathar.

But 2 X 50Z => $160.00 USD + 30$ for power adaptor........and if you need a separate PSU to run the dual pumps, or just a bit larger & higher quality PSU to be sure of all working well...$$ more. So while the 2 X 50Z @13.8v will perform, they also cost at the same level as the MD30, or depending on PSU extra cost.......even a bit more.

The above is getting expensive quick. At this point the MD20Z is looking better when costs of such set ups are factored in.

The estimate you used for the 4 X 12cm fan rad seems to sell the core short of what I would expect of it. But such are estimates. And I understand you used the 120.3 figure as it's a solid set of numbers for the comparison. But a higher performing rad would reduce the pump heat effect of the larger pumps more. I'll have one before much longer, then I'll have a chance to see how good it is.

Not really fair for anyone to expect Cathar to have exact numbers in a thread asking questions regarding pumps not originally part of his test, or rads he doesn't have & his article isn't done yet anyhow.....gets crazy sometimes.

Will look forward to the finished article Cathar.

[Blackeagle]

Originally Posted by unregistered

I'll not critique an article I've not seen

some simplifications are apparent, which could be very significant depending on the system design
- in case
- in case, case air exhausted through the rad
- external cooling box
- actual flow rate

a biggish metal frame pump in a closed case will do more than just heat the water

Any in case use of a Iwaki or panworld ......ect is going to heat up a system big time. But Niksub1 is about the only Iwaki user I've seen who had the pump in case. They are hard to fit to the point this isn't going to be a issure for many.

A properly designed rad box should, IMO, pull air through the rad before it passes over the pump cooling the pump. So the rad box should show better performance/temps.

Your last point is one very few have the equipment to messure Bill, but it's still important. Hope you'll be doing some pump tests soon.

This is a really good thread. G'night.

[Cathar]

Originally Posted by Blackeagle

In so far as the amount of power used, and heat added to water, I see what you are driving at Cathar.

But 2 X 50Z => $160.00 USD + 30$ for power adaptor........and if you need a separate PSU to run the dual pumps, or just a bit larger & higher quality PSU to be sure of all working well...$$ more. So while the 2 X 50Z @13.8v will perform, they also cost at the same level as the MD30, or depending on PSU extra cost.......even a bit more.

The above is getting expensive quick. At this point the MD20Z is looking better when costs of such set ups are factored in.

I would really only recommend the single AquaXtreme 50Z @ 13.8v, and not dual pumps. The differences between the two really is quite small (negligible), and certainly not worthy of getting a second pump.

The 20RZ is actually a lesser performing pump. I ripped those two posts at the start of this thread out of an evolving thread with little or no editing. The 20RZ I would not recommend over a single AquaXtreme 50Z @ 13.8v. I would only recommend the 20RZ in a DC powered format, which sadly doesn't exist, yet.

I apologise to all for the confusion. The statements in the second post of this thread are meant to over-ride the statements from the first post. I see that this is causing some unintended confusion.

The estimate you used for the 4 X 12cm fan rad seems to sell the core short of what I would expect of it. But such are estimates. And I understand you used the 120.3 figure as it's a solid set of numbers for the comparison. But a higher performing rad would reduce the pump heat effect of the larger pumps more. I'll have one before much longer, then I'll have a chance to see how good it is.

Sell it short? How so? I have and actively use a 12" x 10" radiator here. I use 2 x Panaflo L1A's on it at 12v and estimate that my radiator's C/W is around the 0.020 - 0.022 mark with ~33dBA noise (based on manufacturer fan noise ratings).

Estimating a ~0.02 C/W with tolerable fan noise levels is not selling a 25 x 25cm radiator short at all, unless you desire to wear ear-plugs while computing. I would estimate that accounting for the radiator's inefficiencies you would need to be pushing at least around 150CFM through a quad fan radiator to achieve the 0.015 C/W figures I atrributed to it with strong fans. I assure you that achieving 150CFM through even such a large radiator with tolerable noise levels is not as easy as it may seem. Can always achieve quite fantastic radiator performance so long as noise is not part of the equation.

[Blackeagle]

Originally Posted by Cathar

.

Estimating a ~0.02 C/W with tolerable fan noise levels is not selling a 25 x 25cm radiator short at all, unless you desire to wear ear-plugs while computing. I would estimate that accounting for the radiator's inefficiencies you would need to be pushing at least around 150CFM through a quad fan radiator to achieve the 0.015 C/W figures I atrributed to it with strong fans. I assure you that achieving 150CFM through even such a large radiator with tolerable noise levels is not as easy as it may seem. Can always achieve quite fantastic radiator performance so long as noise is not part of the equation.

Yes it does depend a lot on fan selection. 2 or perhaps 3 X Delta 120mm X 38mm 1212HE fans at 130cfm do however offer a lot more air flow/pressure than lower noise 25mm thick fans. Also more nosie, but then that is what my reobus is for. And even with them dialed down to 80cfm each @ approx 35dba, I expect they offer better pressure than a 25mm fan can.

I'll have th core soon, I figure it's the best one to use with the Storm + Fusio + MD20.

Also note, I'll use a rad/pump/fan case which is 1" thick walled wood + 1" sound foam throughout & is some areas 2" sound foam. So it's ending noise levels isn't bad unless I turn up the fans.

[greenman100]

geeze guys

the man goes out of his way to do a whole lot of work

and what's he get in return? silly questions and criticisms

let me say

good job

[DMOS]

Originally Posted by greenman100

the man goes out of his way to do a whole lot of work


good job

I concur. I look forward to the completed articles and graphs. By it's nature it's going to be more or less restricted to equipment on hand for Cathar, but I think I can extrapolate enough data from the graphs to make decisions for my own pump replacement options. Thanks for putting in the work!

[bobkoure]

Originally Posted by Cathar

Eheim 1046, adds around 1.5W of heat to the loop
Eheim 1048, adds 3W of heat
Eheim 1250, adds 9W of heat
MCP600, adds 8W of heat
MCP650, adds 15W of heat
Iwaki MD-15R @ 60Hz, adds 22W of heat
Iwaki MD-20RZ @ 60Hz, adds 31W of heat
Iwaki MD-30RZ @ 50Hz, adds 50W of heat

Any notion if these numbers are loop dependent or not? I'd guess not - but that's a guess...
Yes I realize you only have a couple of loops to test in, but if you get approximately the same numbers for the same pump in different loops that'd be a strong indication...

[ralf_c]

cathar, based on this info would it be fair to assume that two eheim 1046's in series pushing a moderately restrictive loop would be a viable pump solution for a silent non performance loop?

[friku]

What is the ID of the tubing inside of the thermochill 120.3? I ask this because I have 2 Evo 360 (from "Aquacomputer") and I´ve seen that in my radiators that the ID of the tubig is not very big...

Thanks,
friku

[Krum]

Hey guys, I was just wondering if there is a simple, cheap circuit I could build to convert the 12 volts from my PS up to the 14 or so that the MCP600 can take. If it can take up to 14.5, then heck, why not do it? If I could build a converter then I wouldn't need an extra wire into my box. How 'bout it? Anybody got experience building this kind of circuit?

[Belenar]

Originally Posted by Krum

Hey guys, I was just wondering if there is a simple, cheap circuit I could build to convert the 12 volts from my PS up to the 14 or so that the MCP600 can take. If it can take up to 14.5, then heck, why not do it? If I could build a converter then I wouldn't need an extra wire into my box. How 'bout it? Anybody got experience building this kind of circuit?

Sure, a simple transfo could do that for you, but i'm not sure is 12V -> 14V is a commonly used kind of transfo. And then again: Dual coil transfo's add alot of heat to your system and are quite inefficient.

[bigben2k]

Transfos convert AC, not DC.

Some step up circuit can do it, but it'd be custom at this point: cheaper to find another PSU, like a variable PSU from eBay.

[Cathar]

This thing would be perfect.

Costs ~$25 Australian, or around $19 US.

15v PSU. Has trimmer pot to go from 13.5v -> 16.5v.

http://www.computronics.com.au/meanwell/s-25/

Specifically the S-25-15

Will happily fit inside any spare CD-ROM bay.

Definitely enough to power a single 12v pump to an over-volted level, and cheap too.

These PSU's can run in either 115v or 230v countries.

[lolito_fr]

Cathar, I have just been checking my calcs for the graphs in the nexxos thread. I have a problem regarding heat dumped into the loop by MCP650, which you give as 15w.
It seems this pump only draws 15-20w?


Courtesy of Roscal

[9mmCensor]

Transformers do not convert AC. They step it up, down let it remain the same and issolate it. Bridge rectifiers "convert" ac to dc.

[9mmCensor]

Originally Posted by Cathar

This thing would be perfect.

Costs ~$25 Australian, or around $19 US.

15v PSU. Has trimmer pot to go from 13.5v -> 16.5v.

http://www.computronics.com.au/meanwell/s-25/

Specifically the S-25-15

Will happily fit inside any spare CD-ROM bay.

Definitely enough to power a single 12v pump to an over-volted level, and cheap too.

These PSU's can run in either 115v or 230v countries.

why not the S-25-12 which runs 12volts?

[Cathar]

Originally Posted by lolito_fr

Cathar, I have just been checking my calcs for the graphs in the nexxos thread. I have a problem regarding heat dumped into the loop by MCP650, which you give as 15w.
It seems this pump only draws 15-20w?

What's the problem?

Typical real-world operating range is 250-500lph, or 17-18W as per Roscal's graph.

The pump transfers very close to 90% of its input power draw to the water as heat, and the other 10% exits via the plastic casing. 15W is a lower-end estimate for the heat dumped by the pump. In typical use it's more like 15-16W. I personally measured 16W with a Laing D4, and 15.5W with the MCP650, with a Cascade block in the loop as resistance.

If you're surprised at just what a large percentage of input power gets converted as heat, then don't be. Look at the pump's design. Metallic wetted body with eletromagnets attached to the metal that water flows over constantly. The electro-magnets produce most of the heat, and the water cools them almost directly. A small portion of the heat gets convected by air into the plastic pump body, as well as the small micro-controller generating some of its own heat - again - being dumped into the sealed plastic body - for which the heat can exit by the plastic housing, or through the wetted metallic end.

[Cathar]

Originally Posted by 9mmCensor

why not the S-25-12 which runs 12volts?

The question was basically "How do we achieve the ~13-15v that you talk about for over-volting the pumps".

A 12v PSU isn't going to do it.

[9mmCensor]

Originally Posted by Cathar

The question was basically "How do we achieve the ~13-15v that you talk about for over-volting the pumps".

A 12v PSU isn't going to do it.

oh ok. sorry.
yeah the 15 would only get you to 13.2v

[lolito_fr]

Thanks for the explanation Cathar. Just wanted confirmation that you were certain it really was putting such a high fraction of the input power into the water.

[DeadEye]

Does anyoneone have figures for the Hydor L20 wattage put into the water?

[|kbn|]

Cathar nice article.

Im confused why the laing D4 and hydor pumps were not included though? Apart from the niose, I have always considered the D4 to be the perfrect pump for watercooling pc's.
Currently I have mine running @ 9v.

I am Considering changing it for a silent pump, like a hydor L20, as they are very cheap, I could get some money back from the D4...

The hydor pumps also give better pressure than the eqivalent eheim, iirc.. so I dont see why anyone would recommend an ehiem over the hydor, becasue they cost twice as much and are bigger.

[Ruiner]

the D4 is the same product as the mcp650