Starting A Bunch of DDC pump testing, What would you want to see?

[Martinm210]

Just getting started in on a whole bunch of DDC pump top and reservoir top testing. Special thanks to Coolmeister, Petra, NaeKuh, XSPC, and Danger Den for the loaner and sample tops for testing. I think I have just about every top/res top there is including a few new ones.

I have a DDC3.2 fresh from Petras I bought a few weeks ago, as well as some old DDC2's I got in a trade.

I had planned to at least run the DDC 3.2 at 12.0V, but I was wondering. How many people are overvolting these pumps? Would you want a curve at 13.2V also?

Then there is the DDC2's, I think one is either about to die or a 10W version. It has a black impeller, so I'm thinking it may just be a 10W variety, then I have an orange impeller DDC2 that should hopefully be working ok.

I have about a dozen or more different top from several people stacked up here, so I'd rather just run maybe one 12V curve for the DDC 3.2 and one for the DDC2. But if there is enough interest, I'll consider other tests as well.

Here is my new pump test setup. I built a new test reservior with more elevation which does a nice job of almost instantly bleeding the system. I also built a new screw type gate valve that gives me much finer control over restriction/flow rate. I'm also zeroing out any static pressure in the digital manometer that's created by the reservoir and system up above my measured pressure point. This is one really nice feature of the manometer , so I'll be reading true pressure added in by the pump.



And this is what I was thinking for individual pump results. A standard Pressure vs Flow rate type chart. Blue line is the pressure or left axis, and red is the power consumption in watts. Then I'm following suit on my AP1510 test to plot an efficiency curve to get an idea of how the efficiency shifts with different tops.

Anything special you want to see?

Here is my first test of a plain stock DDC3.2 (Swiftech MCP355) at 12.0V:

[NaeKuh]

dual pump configs

[Martinm210]

Working that direction, but I wanted the stock pump baselines first.

I think I found my answer on the overvolting. My particular DDC 3.2 won't start at anything over 13.0V, some sort of protection feature. You can start at 12V and crank it up and beyond, but if it won't start beyond 13.0V, I don't see much use in plotting curves beyond that point. I think I'll just run one 13.0V curve with the stock pump to get a comparison and run all the rest at 12V.

[IanY]

Martin,

Appreciate if you could kindly confirm the following: in a setup with two pumps in series, the performance can be enhanced by placing a pump after the cpu block rather than two pumps back-to-back prior to the cpu block.

Thanks.

[rioja]

i'm interesting in DDC1-T 10W version tests.
Is it true they more quiet than 18W? How about heating issue between 10W and 18W?
Will be 10W enough for CPU DTek-Fuzion-GPU_fullcover loop?

Also how is EK 3/8 barbs top performs on 10W pump?

[Snyxxx]

How about the affect of the suction inlet size versus flow rate. I noticed EK now has G3/8 inlet as well as G1/4.

I remember drilling out the top of the stock top years ago following the directions of a Coolercases where the flow rate was substantially increased going from the top rather than the side inlet. http://www.xtremesystems.org/forums/...=140851&page=2

Maybe see what a 90° elbow does right at the suction inlet as well.

Thanks for all your efforts to this community.

[Martinm210]

Alright, I'll see what I can do.

Coolmeister did send me an EK G3/8 top, I just need to order some barbs for it. I'll see how it compares to the G1/4 EK top. Considering how small the volute chamber is, I'm going to guess bigger isn't really better on the inlet for pressure.

I also just tried the DDC3.2 minimum voltage startup and it actually fires up at around 5.8V, so you could also probably run a DDC 3.2 at 7V with the molex trick or a fan controller. The 13.0V maximum may actually be 12.99V. I'd be afraid to run one near 13.0V, it would only take a small increase in household current to throw it over the cutoff point, then you have no pump turning on.

And I can try some of those questions too, Series close together vs far apart, elbow inlet, etc.

I also now have a good sound level meter thanks to Kapt for the donation, so I'll gather some db results. Unfortunately I don't have a very quiet house, two young kids and TV blaring, but I'll try to gather what I can. I'd call it completely silent up to 9V, and quiet beyond that assuming the pump is isolated from vibrating against something.

With just the DDC 6" measurement from pump, if ambient noise is 53db, 9V = 53db, 11V=56db, 12V=58db, 13V=59db. Pump noise also depends alot on flow rate where low flow/high restriction is much quieter than low restriction/high flow.

[Polizei]

I would be very interested to see how an elbow on the inlet of Petra's top effects performance.

[SiGfever]

Alright, I'll see what I can do.

Coolmeister did send me an EK G3/8 top, I just need to order some barbs for it. I'll see how it compares to the G1/4 EK top. Considering how small the volute chamber is, I'm going to guess bigger isn't really better on the inlet for pressure.

I also just tried the DDC3.2 minimum voltage startup and it actually fires up at around 5.8V, so you could also probably run a DDC 3.2 at 7V with the molex trick or a fan controller. The 13.0V maximum may actually be 12.99V. I'd be afraid to run one near 13.0V, it would only take a small increase in household current to throw it over the cutoff point, then you have no pump turning on.

And I can try some of those questions too, Series close together vs far apart, elbow inlet, etc.

I also now have a good sound level meter thanks to Kapt for the donation, so I'll gather some db results. Unfortunately I don't have a very quiet house, two young kids and TV blaring, but I'll try to gather what I can. I'd call it completely silent up to 9V, and quiet beyond that assuming the pump is isolated from vibrating against something.

With just the DDC 6" measurement from pump, if ambient noise is 53db, 9V = 53db, 11V=56db, 12V=58db, 13V=59db. Pump noise also depends alot on flow rate where low flow/high restriction is much quieter than low restriction/high flow.

Martin when you "undervolt" the pumps would you please take some amp readings so we can get a picture of pump volts vs current under load?

[Martinm210]

Martin when you "undervolt" the pumps would you please take some amp readings so we can get a picture of pump volts vs current under load?

Sure, I've been regularly reading all four item, flow/pressure/volts/amps, So I'll have the same table for the undervolted version too.

Going to go do that now, 12V done, 13V done, now 7V.

[systemviper]

Wow Martin, your just amazing, your dedication to all the little factors that help squeeze out the best performance possible. Just want to say thanks for all you do,it is invaluable! Great Job.


PS, i am just still in awe on that chiller you built, DAMN that's good!

[C'DaleRider]

As for your DDC pumps..... A real DDC2 pump has an orange impeller, a DDC has a black impeller. So, if one of your pumps is acting like it's a 10W pump (the one you suspect is dying) and has a black impeller, it is almost assuredly a 10W DDC pump, not a DDC2.

I've had two DDC and three DDC2 pumps.....and they all followed the above color scheme.....black=10W, orange=18W.


BTW.....which tops do you have for this testing session?

And, I think we all appreciate the hard work, effort, and time you've been putting into all your testing of components. The information has been invaluable! Thanks!

[Waterlogged]

Martin,

Appreciate if you could kindly confirm the following: in a setup with two pumps in series, the performance can be enhanced by placing a pump after the cpu block rather than two pumps back-to-back prior to the cpu block.

Thanks.

I too would like to see this test as I was one of the first (if not the first) to walk out on that branch. I don't hear it cracking but...I'd like to be sure.

I doubt there'll be much difference with just a CPU block, but now if it was done with a CPU and full GPU (and maybe a NB), there should be something measurable there.

+1 on all the comments about the work put in.

[Martinm210]

As for your DDC pumps..... A real DDC2 pump has an orange impeller, a DDC has a black impeller. So, if one of your pumps is acting like it's a 10W pump (the one you suspect is dying) and has a black impeller, it is almost assuredly a 10W DDC pump, not a DDC2.

I've had two DDC and three DDC2 pumps.....and they all followed the above color scheme.....black=10W, orange=18W.


BTW.....which tops do you have for this testing session?

Thanks!, probably jua DDC1 then..

I've got a bunch of them, Petras, EKs, Alphacools, XSPC, Koolance, and all of the reservoir tops. I wouldn't be able to do it without everyones contribution on the loaners, I'm just having fun with more toys to play with..

Going to take a while to get these all run, but it shouldn't take forever. Then I can do a couple of the odds and ends tests

[leo_bsb]

using elbows on the pump would be a nice test.
I don't think any voltages over 12V that usefull since the average guy can't do it at home.
I really want to know how the EK RES/Top combo performs.
Other than that a graphic showing how each pump stack against each other but measured on temperature performance gain

[NaeKuh]

Martin seriously a thanks goes out from me to you...

I really enjoy looking at your results, very professional, and very accurate.



seriously your a big value to the XS forums.

[Xilikon]

Martin seriously a thanks goes out from me to you...

I really enjoy looking at your results, very professional, and very accurate.



seriously your a big value to the XS forums.

I will chime what NaeKuh said, this is unvaluable knowledge to learn. I have no questions since everyone already said what I have in mind (like a elbow just before the inlet for space saving purpose).

I especially like the bell curve of efficiency vs flowrate.

[Martinm210]

Ok, I'm done with the stock DDC3.2 and voltage testing. It turns out it has it's highest efficiency at the highest voltage, unfortunately that's a very fine line between the pump starting and the pump failing to start!!!
12.9V = life is great, 13.0V mean you have NO FLOW!

Anyhow, here are the stock pump curves:

Overvolted to about 12.99V: It was actually run and held at 13.0V, but it would not start from about 12.95V on up. This .05V is probably just meter accuracy, looks to me Laing is protecting the pump at the 13.0V and higher mark.

It's unfortunate because it is a noteworthy performance improvement. At 1.5GPM i means about 24% more pressure. I would just be afraid to run here knowing a little surch in power could mean the pump won't start that day.



In the end, it's a good thing to slightly overvolt the pump, but very scary when you find out the pump will simply fail to start if you are at or over or even near 13.0V or higher!.

Undervolting has potential to make a DDC3.2 close to DDC3.1 performance with around 10Volts.

Enough of this stock top, time for some aftermarket tops..

[joecop120]

Very nice! Could you do a temperature analysis between the three pumps, with and without Petra's gel/neoprene foam?
Thanks for the hard work already!

[Martinm210]

Very nice! Could you do a temperature analysis between the three pumps, with and without Petra's gel/neoprene foam?
Thanks for the hard work already!

Sure, I had planned to do some sort of heat dump check on the pumps at the end. It might be worthwhile to do something like that. Maybe sitting on foam vs in the air with a fan on the pump to see if I can measure any difference in heat dump between the two.

[Kayin]

Should I send in my cast pump top?

It may suck, but I made it all by myself... I'm at least proud of that.

[C'DaleRider]

I'd NEVER put a gel pad under these pumps.....and this is from experience.

I ran a Petra's gel pad under the very first pump, a DDC, I owned. Then, when the Chicken Little "The sky is falling!" reports of DDC/DDC2 pump failures began, I revisited my pump and its gel pad.

While it's true the gel pad isolates the pump from the case vis-a-vis noise/vibration transmission, the gel pad is also a great insulator of heat created by the pump......that is, the gel keeps the heat the pump creates in the pump. My pad was extremely warm, almost to the point of getting to a semi-liquid state.

I was horrified....the pump bottom was very hot to touch, too.

I immediately tossed the gel pad in the garbage and have since moved to using 1/2" square by 1/4" tall foam blocks applied to the outermost corners of the pump. The foam does not insulate the pump and have it retain its created heat; instead they lift the pump off its flat surface letting air flow underneath the pump. The foam blocks also do a great job at isolating any noise/vibrations the pump may try to transmit to the case.

If you want to test the gel pad, be my guest. But, in my experience, it'd be just a waste of time.....unless you really want to shorten your pump's life with heat.

[DarK Nerevar]

Martinm210 could you test this pump and compare it to a mcp655?
The pump is: SANSO PDH 054 12 V ?
I've read a little review here.
Sorry it's italian

[MrToad]

You've saved me the legwork as usual Martin, thanks

I was going to over volt the pumps, the Meanwell PSU I'm using lets me adjust +-10%, but in the light of your results I may as well leave it alone.

Much looking forward to your results. In a leap of faith I bought the EK 1/4 tops for the pumps without any hard evidence but EK's reputation to back up the decision. Now we'll see if I did put my money on the right horse

[Martinm210]

Should I send in my cast pump top?

It may suck, but I made it all by myself... I'm at least proud of that.

Sure, always happy to test for fellow members. You never know until you test it

I'd NEVER put a gel pad under these pumps.....and this is from experience.

I ran a Petra's gel pad under the very first pump, a DDC, I owned. Then, when the Chicken Little "The sky is falling!" reports of DDC/DDC2 pump failures began, I revisited my pump and its gel pad.

While it's true the gel pad isolates the pump from the case vis-a-vis noise/vibration transmission, the gel pad is also a great insulator of heat created by the pump......that is, the gel keeps the heat the pump creates in the pump. My pad was extremely warm, almost to the point of getting to a semi-liquid state.

I was horrified....the pump bottom was very hot to touch, too.

I immediately tossed the gel pad in the garbage and have since moved to using 1/2" square by 1/4" tall foam blocks applied to the outermost corners of the pump. The foam does not insulate the pump and have it retain its created heat; instead they lift the pump off its flat surface letting air flow underneath the pump. The foam blocks also do a great job at isolating any noise/vibrations the pump may try to transmit to the case.

If you want to test the gel pad, be my guest. But, in my experience, it'd be just a waste of time.....unless you really want to shorten your pump's life with heat.

I agree, the bottom of the DDC's seems to get alot hotter than the casing on the D5's, and any heat you can get rid of externally means less heat in the loop. I like someo of the case builds I've seen where they elevate the pump on rubber mounts and cool the pump with air. Keeping it cool can't hurt lifespan as well.

Martinm210 could you test this pump and compare it to a mcp655?
The pump is: SANSO PDH 054 12 V ?
I've read a little review here.
Sorry it's italian

If I could find a place that's willing to send me a sample. I'm not sure where you can even buy that pump.

You've saved me the legwork as usual Martin, thanks

I was going to over volt the pumps, the Meanwell PSU I'm using lets me adjust +-10%, but in the light of your results I may as well leave it alone.

Much looking forward to your results. In a leap of faith I bought the EK 1/4 tops for the pumps without any hard evidence but EK's reputation to back up the decision. Now we'll see if I did put my money on the right horse

You could probably overvolt safely to about 12.5V, but I'd be too afraid of pushing more than that against the "no start 13V fenceline".

They are all good tops and a significant benefit over stock performance, but there are a couple of leaders developing. I got a bunch of them done on the DDC 3.2 today with some suprising results. I'll start a new thread with it all when it's done...a few more to go, then the DDC2 curves.