More Petra'sTech DDC Pump Testing (curves, power, new pumps/tops, etc.)

[Sparky]

OMGWTFBBQ!!!!11111ONEONEELEVEN can you try the RD-30 impeller in the DDC-2?

Unfortunately, that test didn't go so well and I haven't a clue as to why...




Yes, we're going to carry both the RD-20 and the RD-30. At the moment, we're just waiting for the Iwaki rep to clear our credit application and then begins the two week wait for the pumps to get here.

...in the meantime, however, I'll be putting my Iwaki WMD-20RLT on the site for sale... anyone interested?

[[XC] DragonOrta]

...in the meantime, however, I'll be putting my Iwaki WMD-20RLT on the site for sale... anyone interested?

I love the listing. For some reason, it reminds me of some kind of collectible item you'd get in an RPG. lol

[Walternowi]

Yes, we're going to carry both the RD-20 and the RD-30. At the moment, we're just waiting for the Iwaki rep to clear our credit application and then begins the two week wait for the pumps to get here.

Oh my!

I MUST resist, I have too many water cooling gear already!
I MUST resist, I have too many water cooling gear already!
I MUST resist, I have too many water cooling gear already!

[kuhla]

http://site.petrastechshop.com/images/RDDC-30.jpg

When will it be in stock?

[nikhsub1]

http://www.xtremesystems.org/forums/...ad.php?t=48533

^^ wow I've had my RD a long time... since '04

[123bob]

I'm flow testing my loop before putting it in the machine. Figured I’d add the data here since it involves Petra’s top. I’m strictly looking a flow, with full power to the pump. I used the actual parts of my loop to get a feel for what the pump does on it’s own vs what load the loop puts on it.

The full loop is;
MCR320 Rad>MicroRes>DDC2 pump w/ Petra top>Fuzion CPU block>MCW60 GPU block>DD Maze4 SPP (aka NB) block>Rad.


Test setup pic is below



Here are the results.




Edit - test #2 was done with only the Petra top.

Note that the Petra top kicks butt compared to the stock top. []
Flow is very good across the board with the Petra top. I'm a believer....[]

The current machine that this loop is going into is as follows;

MB - EVGA nForce 680i - Rev D - P26 Bios - back to defaults, pending better quad cooling!
Proc - QX6700 10X multiplier for 2.66gig
Mem - 2g Corsair TWIN2X2048-8500C5 800 Mhz
HD - 2x 500g Maxtor 7H500F0 in Raid 0
Graphic - 1x BFG 8800GTX
CPU Fan - Zalman CNPS9700 LED
PSU - Thermaltake 850w Toughpower
Case - Antec 900 Revision 1.1
DVD - 2x Sony DRU120C DVD-RW
Floppy - generic cheapo

Regards,
Bob

[the highlander]

Is the MCP655 a better option for a pump?

[MaxxxRacer]

Alex,
correct me if im wrong, but 123bob 's numbers look a bit high to me.

Scott,
All I have to say to you is DEEH DEEHH DEEHH!

[Kayin]

How would I go about generating numbers like this for my own top? I'm not exactly used to data gathering like this, mostly to synthetic benchmarking and playing lots of games...

[123bob]

Alex,
correct me if im wrong, but 123bob 's numbers look a bit high to me.

I thought so too. I must say I was surprised that the full loop performed that well. That's why I did the setup carefully and with three time trials. The container I'm filling is a 2.5 gallon distilled water jug. Note that I'm running full bore on the power supply pictured. I was most pleased by the full loop numbers. I'm on generic 1/2 inch tubing on the test, and the only other thing is that I don't have the tube bends I will have in the case. The full loop test is repeatable, with consistent numbers. I've run this case before. I think the fuzion block is a wonderful thing....

Note that I'm sucking the water out of the bucket in the sink, therefore it's experiencing atmospheric pressure only, not city water pressure as if I connected straight to the faucet.

That being said, if you look at Alex's opening flow numbers on page 1, my pump-only number is actually a tad lower than Alex's result.
Regards,
Bob

[serialk11r]

I thought so too. I must say I was surprised that the full loop performed that well. That's why I did the setup carefully and with three time trials. The container I'm filling is a 2.5 gallon distilled water jug. Note that I'm running full bore on the power supply pictured. I was most pleased by the full loop numbers. I'm on generic 1/2 inch tubing on the test, and the only other thing is that I don't have the tube bends I will have in the case. The full loop test is repeatable, with consistent numbers. I've run this case before. I think the fuzion block is a wonderful thing....

Note that I'm sucking the water out of the bucket in the sink, therefore it's experiencing atmospheric pressure only, not city water pressure as if I connected straight to the faucet.

That being said, if you look at Alex's opening flow numbers on page 1, my pump-only number is actually a tad lower than Alex's result.
Regards,
Bob

Wait a minute, the pump is sucking out of the sink yet is getting that much flow???

[NaeKuh]

Wait a minute, the pump is sucking out of the sink yet is getting that much flow???

no one ever said you needed to PUSH the water into the blocks


My loop has both push and sucking on the cpu. Its worked awesome like that for about 4-5 months.

Hence why i keep tellign everyone NO, im not going on dual loops.

Just a quick question tho, hows the flow on a RD-30 compare to 2 x DDC-2 /w petratop.

Trying to resist upgrading the RD30, if i dont need to. But its hard not to touch my computer at least once a month.

[serialk11r]

no one ever said you needed to PUSH the water into the blocks


My loop has both push and sucking on the cpu. Its worked awesome like that for about 4-5 months.

Hence why i keep tellign everyone NO, im not going on dual loops.

No take a look, the pump's inlet is SUCKING a good few inches UPWARD.

[123bob]

Yup the pump is sucking up...(no pun intended). But also look at the discharge tube. It going into the container at roughly the same level down from the pump. The siphon effect on the downhill side should negate the uphill pull on the inlet side, right?

Anyway, I just checked the 2.5 gallon container to see how close it was to 2.5 gallons. Using a measuring cup, it is 41.5 cups from empty, right to the tippy top, overflowing point. (Nice to know I'm not getting shorted distilled water when I'm buying it. ) Therefore, do the math and you come up with me filling 2.5 gallons, plus 1.5 cups. So, my numbers are actually a touch low. Not enough low for me to go recalculate the spreadsheet (though I probably will anyway....).

For timing, I bought a digital stopwatch just for this test. My lab assistant operated the stopwatch while I yelled "go" and "stop". I think the times are good since there is fairly small variance in the numbers. I also double checked the math. It's good too. So, there you have it. The flow is the flow.

I think this also possibly shows that the FuZion kicks butt on flow?

I would like to see others get in the "lab" and do this. So far, I only know of Alex that has posted actual flow data. I would be curious to know how other blocks and configurations perform. Please point me to the thread if others have already done this. Thanks.
Regards,
Bob

[NaeKuh]

No take a look, the pump's inlet is SUCKING a good few inches UPWARD.

ahahaha... once u get a pump going to push water, its goign to suck.

The sucking power should be almost as = as its pushing power.

You cant blow, without inhaling. :P

[Petra]

I would be curious to know how other blocks and configurations perform.

You know, superimposing a component's pressure drop vs. flow curve onto the pump's performance curve will, pretty much, tell you that (it won't take tubing into account, though). Add the formulae from various part's pressure drop curves together, superimpose it on the pump's performance curve, and you'd have a pretty good approximation of what the system's flowrate would be. The trick is in pulling the data together.

On that note, employing the same idea, one could create a 'cooling system predictor' of sorts... but it would be a monumental undertaking, as far as testing goes (gotta get the data from somewhere, after all). I went more into detail about this in the [H] a while back, but I guess that the act of suggesting they do something scientific offended Kyle's sensibility...

[123bob]

Yeah Alex, that was kind of where I was heading with this. At least if we could predict a loop's flow with even rough accuracy, it would make part selection easier. A tool, or spreadsheet could be made for this. (It would also help with multi-loop design)

I don't expect to see the manufacturers putting this out anytime soon. I would think this type of info is a well kept secret, if they even do testing like that. (Except for you of course, you published your top data for all to see.. )

(Edit - Well after looking around, I find that some of this info is available. Maybe I'll try scooping some up and see if a spreadsheet tool is possible. I would be curious if the tool could get even close to my test results. As Alex suggested, I'll try an intersecting curve type approach. It will be one of those get-around-to-it things. Right now my issue is trying to cram all the stuff into an Antec 900 case, DOH! May have to get a new case....)

Note that this information, by itself, would not constitute a good part over a bad one. For example, if I'm not mistaken, the Storm is a more restrictive part, but a very effective cooler, for it's particular application.

The proposed tool will not accurately predict what temperature you'd get out of the loop, but at least you would know the flow. There are probably too many other variables at play for accurate temperature prediction, but I think good, calculated flow design would help max out the loop's potential.

Perhaps the next step might be to have a tool that can calculate the heat dump of a given radiator, with a given flow, for various temperature deltas. That would put us one step closer to being able to predict temps, or at least the ideal cooling capacity of the proposed loop. (So many radiators, and so many fans!! Just think of the permutations to test!! )

One problem I see is the variance block mounting has. You can get the temps to move all over the place just by paste and mounting attempts. I suppose that may be overcome as time goes on.

As it stands, my experience was to pick my parts, then just measure the whole mess. If I get a good number, well OK. If I don't, I'm forced to troubleshoot to the offending part. Of course, what is a good number? That's what I'm trying to get more info on. I do not have years of experience working with this particular type of system.

For the purpose of getting my quad taken care of, I'll call 2 GPM good and see how it does in the machine....

I have to admit, this was a fun test to do. I don't have a pressure measurement setup, so I'm limited to only max flow measurement.
Regards,
Bob

[InfamousChu]

thanks for the informative post. Should have a ddct01s on the way in a week or so!

[OdinEidolon]

nice adn informative comparison petra!

i have the possibility to get some DCC-2s for cheap... say 35$ each (i live in EU so its very cheap compared to how much we pay here)
is it wqorth the risk? i know that lots of them died in the past

[aspire.comptech]

nice adn informative comparison petra!

i have the possibility to get some DCC-2s for cheap... say 35$ each (i live in EU so its very cheap compared to how much we pay here)
is it wqorth the risk? i know that lots of them died in the past

If they're lightly used, then chances are they aren't going to die on you now.

The ones that did die, did so within the first month of use.

[Waterlogged]

Necro thread in progress detected!

[XS_RoB]

Man - I got excited - I thought this was the unveiling of Petra's NEW AND IMPROVED TOP - to reclaim the throne from XSPC!

Im going to go cry silently now......

[T_Flight]

Necro thread in progress detected!

bwoop bwoop bwoop All hands man your battle stations, All hands man your battle stations this is not a drill...I repeat this is not a drill.