Some New Blood On The Testbed :)

[nikhsub1]

5c cooler ambient temps over the previous testing? Uh, yea, I think that should make a difference, theoretically in a perfect loop you're looking at 5c cooler water temps.

Yeah but it is the delta between coretemp and water temp... in theory it should scale linearly, but maybe not.

[Cyprio]

REALLY cant wait for more results as my AC block is patiently waiting for me to complete my case build!

[Petra]

5c cooler ambient temps over the previous testing? Uh, yea, I think that should make a difference, theoretically in a perfect loop you're looking at 5c cooler water temps.

Remember, Scott is measuring the delta T (CPU temp minus water temp), so the data should still be good regardless of the difference in ambient temperature... However, he's also using TAT to monitor the CPU temperature--this is where things get interesting. During some early testing that I was doing for Enzotech, I noticed a discrepancy between my TTV data (E6400 w/CNC milled IHS and thermocouple embedded per Intel TTV spec) and delta T data taken using TAT... basically, the CPU-water temperature delta calculated from CPU temperatures provided by TAT was being effected by changes in ambient temperature, whereas my TTV data wasn't.

As far as I'm concerned, this is kind of a big deal... but I need to look into it further. Hopefully I'll have a chance to put together a crude controlled 'hot box' within the next couple weeks to experiment with testing in hot and cold environments. If I can manage to isolate where the point of influence is, then I may consider TAT/CoreTemp good tools for (non-enviro chamber) controlled testing again.

*EDIT* BTW, Scott... I have a few H2O-120 Compact kits now--let me know if you want me to toss one your way (since I remember you mentioning it a while back).

[nikhsub1]

Woops, I misread my own excel file... the first run was NOT 25.48 deltaT, it was 23.87! The 25.48 was air/cpu delta... silly me.

[Nate P.]

Wow, that's a good looking block, and pretty good performance to boot!

[SiGfever]

Scott, are you still using the Iwaki on your test bed? If yes then a test with a DDC-2 or D5 Vario would be nice to see for more comparisons for such a restrictive block.

[Kayin]

is it too early to predict that this may actually be a performer?

[ShoNuff]

Scott, I'm glad you are testing this block as I am sure many will be impressed with your results. Now that AC has moved beyond their 1/8" fittings I know the Cuplex DI will be competitive.

Anyway, in your first post you said that the block was restrictive. Was that your opinion while using your dual Laing DDC pumps in series?

I am considering this block BUT atm I am waiting on the results of EK new block using the more restrictive/high performance mid-plate.

Looking forward to your results.

[Jedda]

Remember, Scott is measuring the delta T (CPU temp minus water temp), so the data should still be good regardless of the difference in ambient temperature... However, he's also using TAT to monitor the CPU temperature--this is where things get interesting. During some early testing that I was doing for Enzotech, I noticed a discrepancy between my TTV data (E6400 w/CNC milled IHS and thermocouple embedded per Intel TTV spec) and delta T data taken using TAT... basically, the CPU-water temperature delta calculated from CPU temperatures provided by TAT was being effected by changes in ambient temperature, whereas my TTV data wasn't.

As far as I'm concerned, this is kind of a big deal... but I need to look into it further. Hopefully I'll have a chance to put together a crude controlled 'hot box' within the next couple weeks to experiment with testing in hot and cold environments. If I can manage to isolate where the point of influence is, then I may consider TAT/CoreTemp good tools for (non-enviro chamber) controlled testing again.

Kind of a big deal, indeed.
Will wait with interest for any light you can cast on why that's showing up.

[nikhsub1]

Anyway, in your first post you said that the block was restrictive. Was that your opinion while using your dual Laing DDC pumps in series?

No I don't use dual laings on the testbed, I use the RD-30 at 18V... let's just say that the swiftech micro res is usually pretty active, even with a nozzled fuzion or a storm. With this DI it is hard to see the water even churn in the res. I mean, very restrictive. I suspect this is why it performs the way it does (so far). I really do like this block a lot.

[nikhsub1]

Remember, Scott is measuring the delta T (CPU temp minus water temp), so the data should still be good regardless of the difference in ambient temperature... However, he's also using TAT to monitor the CPU temperature--this is where things get interesting. During some early testing that I was doing for Enzotech, I noticed a discrepancy between my TTV data (E6400 w/CNC milled IHS and thermocouple embedded per Intel TTV spec) and delta T data taken using TAT... basically, the CPU-water temperature delta calculated from CPU temperatures provided by TAT was being effected by changes in ambient temperature, whereas my TTV data wasn't.

As far as I'm concerned, this is kind of a big deal... but I need to look into it further. Hopefully I'll have a chance to put together a crude controlled 'hot box' within the next couple weeks to experiment with testing in hot and cold environments. If I can manage to isolate where the point of influence is, then I may consider TAT/CoreTemp good tools for (non-enviro chamber) controlled testing again.

*EDIT* BTW, Scott... I have a few H2O-120 Compact kits now--let me know if you want me to toss one your way (since I remember you mentioning it a while back).

Alex, do you mean that if ambient went up (or down) 4C that TAT would reflect this 4C change at the coretemp measurement? If so this is to be expected. There should be a linear rise and fall in coretemps with ambient temp changes (in theory). My first test with the AC DI block was with ambients of ~23-24C or so. Most of the other tests were taken a good 5C higher than that on average. Perhaps I should wait to do more runs until my ambients are closer to 28C or so again. I know I would feel better about it even though in theory it should not matter but then there is the whole matter of compression of temps and at what thresh hold that may start to affect thngs.

[Martinm210]

From what I read, it sounded like Alex may have picked up an oddity with TAT and while it shouldn't matter what your ambient is (when recording water/core deltas), he was seeing otherwise.

That would be interesting, because I've heard more than once now about some bug in TAT.

I still seem to have produced a good and repeatable sets of results after 5 runs or so using the logging function, but I have seen a few odd tests that I had to throw out.

If only coretemp had a logging function....

[Petra]

Alex, do you mean that if ambient went up (or down) 4C that TAT would reflect this 4C change at the coretemp measurement?

Nope, the CPU-water delta was being effected (even though it shouldn't have been) and that's all I was really paying attention to while testing at the time... which is why I need to investigate further.

From what I read, it sounded like Alex may have picked up an oddity with TAT and while it shouldn't matter what your ambient is (when recording water/core deltas), he was seeing otherwise.

Bingo.

If only coretemp had a logging function....

I could have sworn that it did... or that some version of it did. Anyway, I plan on taking a look at CoreTemp, too.

[headala]

From what I remember of thermodynamics, wouldn't the rad operate slightly more efficiently the lower the temps?

As much as I want this block to perform well, I think you'd better test it at the same ambients as the others....especially considering how much controversy surrounds just about anything AC puts out.

[Snyxxx]

Martin/ Petra,

OCCT 1.01.4b together with core temp produces a temperature graph versus time. You probably already know that though. It works nice.

[Erasmus354]

I want to see a test with an Eheim 1048, considering that is basically what a typical AC setup will be using with this block. Would be interesting to see how much performance you would lose.

[Pete]

Scott, thanks for going out on a limb and testing all these blocks for us and sharing your info here.

Amazing work as ever

[Xilikon]

Core temp does have a logging function, I use it all the time.

andyc

Yes, it does in a txt format. You can easily import in Excel to have some graph or do some work extracting some min/max/avg data.

[nikhsub1]

Guys, I hear what ya'll are saying but I can't deviate from what I've done with ALL the other blocks/mods I've tested thus far... They would all have to be RETESTED if I changed anything to have any comparison value whatsoever. Good or bad, I've got to stick with exactly what i've been doing.

On that note, I think I will discard my first run with the 24c ambients just so I can get the atmosphere more in line with the other tests that I've done. As I've always said, my testing is far from ideal and not the end all be all. What I do give though is attention to detail, repetition and unbiased results.

[Jupiler]

No I don't use dual laings on the testbed, I use the RD-30 at 18V... let's just say that the swiftech micro res is usually pretty active, even with a nozzled fuzion or a storm. With this DI it is hard to see the water even churn in the res. I mean, very restrictive. I suspect this is why it performs the way it does (so far). I really do like this block a lot.

You're right.
According to Radical_53, who has tested 16 blocks in the latest edition of the HWLuxx [printed], a German magazine, the Cuplex is a very restrictive block.
The Fuzion is still the flow king, with 5.82L/min.
As for the Cuplex, he measured 3.86L/min and it ended 3th last in the flow rate charts.
It performed very well on both a dual core and quad core though.

[Petra]

Guys, I hear what ya'll are saying but I can't deviate from what I've done with ALL the other blocks/mods I've tested thus far... They would all have to be RETESTED if I changed anything to have any comparison value whatsoever. Good or bad, I've got to stick with exactly what i've been doing.

...and there's nothing wrong with that. I'm not trying to invalidate anything that you're doing, nor am I trying to create more work for you. I brought the oddity up because I thought that it would be good for you and others performing testing to, if nothing else, just keep in mind while testing so that you might be more likely to spot any odd data and/or limit the impact by not collecting data from TAT when there are large differences in ambient temperature.

I still need to look into this more to gain some idea of what's going on.

[Vampire_Hunter]

I want to see a test with an Eheim 1048, considering that is basically what a typical AC setup will be using with this block. Would be interesting to see how much performance you would lose.

Or gain?

The problem here of course is the comparing of apples to oranges. Unless you test all the other blocks with the same pump you just have an interesting piece of information.

If nikhsub1 is willing I would send him the older model Aquastream pump that was being sold a few months ago that can OC up to about 72 Hz. I think new ones will OC up to about 94 Hz.

Edit 12-3-07:

Forgot to mention that a lot of people who use AC components use the DDC's of one variant or another. Last time I looked AC (and their distributors) sell both Aquastream's and Laing DDC's. On the German forums I think more people are using DDC's.

[ShoNuff]

You're right.
According to Radical_53, who has tested 16 blocks in the latest edition of the HWLuxx [printed], a German magazine, the Cuplex is a very restrictive block.
The Fuzion is still the flow king, with 5.82L/min.
As for the Cuplex, he measured 3.86L/min and it ended 3th last in the flow rate charts.
It performed very well on both a dual core and quad core though.

Do you know if any part of Radical_53's testing is available online? I would really like to see his results.

[Martinm210]

Martin/ Petra,

OCCT 1.01.4b together with core temp produces a temperature graph versus time. You probably already know that though. It works nice.

Been a while since I ran OCCT...will check it out..thanks!

[Martinm210]

Guys, I hear what ya'll are saying but I can't deviate from what I've done with ALL the other blocks/mods I've tested thus far... They would all have to be RETESTED if I changed anything to have any comparison value whatsoever. Good or bad, I've got to stick with exactly what i've been doing.

On that note, I think I will discard my first run with the 24c ambients just so I can get the atmosphere more in line with the other tests that I've done. As I've always said, my testing is far from ideal and not the end all be all. What I do give though is attention to detail, repetition and unbiased results.

Yeah, that's what I would do. As long as everything was tested exactly the same, the comparison will still be valid and very worthwhile. The couple of odd tests I had found were usually pointed out after making several runs.

Strength in numbers..