AquaXtreme MP-05 triad Review by SystemCooling

Cathar · 10-01-2005, 01:03 AM

http://www.systemcooling.com/cooltec...xtreme-01.html

Eddy_EK · 10-01-2005, 01:31 AM

Quote:

Originally Posted by Cathar

http://www.systemcooling.com/cooltec...xtreme-01.html

What are your comments cathar, as it is almost as good as swifteck storm?
But it is more restrictive than storm.

I see some big difference in results between Overclockers and System Cooling setup. What can you say about that?

Cathar · 10-01-2005, 01:34 AM

The graph that I look to is the power vs C/W graph. Says it all for me really.

http://www.systemcooling.com/images/...image31big.gif

Eddy_EK · 10-01-2005, 02:03 AM

Quote:

Originally Posted by Cathar

The graph that I look to is the power vs C/W graph. Says it all for me really.

http://www.systemcooling.com/images/...image31big.gif

Yes, yes, it sows more difference.

But I really prefered some more ilustrative testing method, I havent have time to poore over hydraulic power comparison and thesis.
For example, results of C/W for at least 4 pumps (eheim 1046, 1048, 1250, and one very powerfull pump). So the viewers would se what to expect form pumps with simmilar characteristics.

Cathar · 10-01-2005, 02:32 AM

Uh, that's what the Hydraulic Power vs C/W graph tries to show you. Shows performance relative to a fixed pumping power, which at a glance provides a decent summary of the performance across a range of blocks with a fixed pump, or at least pumps of similar characteristics.

A pressure vs C/W graph favors lower pressure drop blocks.
A flow vs C/W graph favors higher pressure drop blocks.

A power vs C/W graph treads the middle ground. If you want an exact value you still need to plot the particular pump's pressure-flow curve against the system's pressure-flow curve and find a flow rate, and then look at a flow vs C/W graph to derive a performance figure. The power-C/W figure tries to present a generalised summary of that sort of information though. It is a generalisation, but it's a lot closer to hitting the mark for users who just want to look at a single graph and nothing else, to presenting an even handed appraisal of the efficiencies of various waterblock designs when matched with any given pump.

jaydee · 10-01-2005, 09:46 AM

Bah, read graphs wrong...

jaydee · 10-01-2005, 10:07 AM

Bah, I am reading the graphs wrong. For some reason I assumed he used the MP-05 SP with all 3 jets. Now I see he did the three blocks with their default nozzels....

I would have liked to see the MP-05 SP with all 3 nozzels done. I got the #1 much better than the #3 but more restrictive aswell.

dacooltech · 10-01-2005, 12:43 PM

Quote:

Originally Posted by jaydee

Bah, I am reading the graphs wrong. For some reason I assumed he used the MP-05 SP with all 3 jets. Now I see he did the three blocks with their default nozzels....

I would have liked to see the MP-05 SP with all 3 nozzels done. I got the #1 much better than the #3 but more restrictive aswell.

The MP-05 Pro BP has 270pins and he tested it with N01 slit (which is the default mid plate for the Pro)... You tested the N01 nozzle with MP-05 SP BP with 300 pins.

Now one other thing is Robotech e-mailed me to give me the heads up on 2 things.

1) Springs that he received are on the light side for A-64 2x studs mounting

I found out that he was right, and the last batch of springs that we received were a bit different than what we normally use. Thanks to Lee, I took care of that one before sending them to the customers... but couldn't send him new set of springs before he completed the review.

2) He also e-mailed me that he measured the channels on the base plate at .015" wide (going the long way) and .012" wide (going the short way).

I know the exact reason for that, and that's my fault really. I rushed out the first samples, and unlike the production units, that small variation causes the samples that I sent him being more restrictive. I should be more careful, but a lesson learned...

I pretty much perfected things now, but there's still more work to do...

Edit: wording

jaydee · 10-01-2005, 01:03 PM

You can see in that picture that the pins are not the same. That is definatly going to mess with the #3 jet if those holes get partially covered by a pin. That will hurt flow and performance some.

Here are the pics of the 300 pin base on the one I have. As you can see the pins still very but I don't think they very as badly as Robo's. Next time I take the block apart I will look more carefully to see how the holes on the #3 jet line up. Last time I looked it did look like there was some overlap. I just about got my bench replumed and will run the block again with the new plumbing and silver TIM compound I switched to. Might not be all that usefull though if the retail blocks are now slightly different/improved.

RoboTech · 10-01-2005, 10:05 PM

Quote:

Originally Posted by jaydee

I would have liked to see the MP-05 SP with all 3 nozzels done. I got the #1 much better than the #3 but more restrictive aswell.

Hey jaydee,

I'm looking forward to seeing your test results - maybe you can help fill in some of the gaps I was not able to (like how the SP performs with the 3 different nozzle plates). There was a lot more testing I would have liked to have done, but doing the four CoolTechnica blocks took me almost four weeks and about 100 hrs. I'm trying to maintain a life beyond waterblock testing (family and my day job still come first).

I think jaydee and pH understand where I'm coming from...

Anyway I hope to see your results soon and maybe we can get pH back in the game one of these days.

jaydee · 10-01-2005, 10:41 PM

Quote:

Originally Posted by RoboTech

Hey jaydee,

I'm looking forward to seeing your test results - maybe you can help fill in some of the gaps I was not able to (like how the SP performs with the 3 different nozzle plates). There was a lot more testing I would have liked to have done, but doing the four CoolTechnica blocks took me almost four weeks and about 100 hrs. I'm trying to maintain a life beyond waterblock testing (family and my day job still come first).

I think jaydee and pH understand where I'm coming from...

Anyway I hope to see your results soon and maybe we can get pH back in the game one of these days.

It was my bad on my first post.
I hear you. Testing is a bitch. I got about 300hrs of data all scrapped right now on the MP-05 SP, TDX, and WWLE. Months of time and money burnt on the bench so far.....

I am completely re-designing my test bench right now. I have used your bench as a guide on how I want to re-create mine.

One thing I am not 100% clear on is how you are taking your differential pressure measurements? I see on your site you have a manometer and I see pressure gauges. What do you use to take the measurements? That is one area I am lacking.

Also do you feel the 14 x 14 die is the right size to use? I am using 12 x 12 right now and I get various answers from 100mm to 1" should be used.

Pretty sure I am going to stick with 12 x 12 for now. Just curious what you think....

One of these days me you and pH should start a thread and try and figure out why all our results are way different. For example in this graph of your work http://www.systemcooling.com/images/...image27big.gif compared to yours JoeC has different numbers: http://www.overclockers.com/articles373/wbsum.asp ph has different numbers: http://www.procooling.com/html/pro_testing.php and I have different numbers: http://forums.procooling.com/vbb/att...achmentid=4828

Anyway I am getting a new flowmeter and dP meter as soon as I can. Then I will make a new die sim and start over again.

Cathar · 10-01-2005, 10:56 PM

I understood that Robotech's die-sim was 14x14mm...

redleader · 10-01-2005, 11:16 PM

All his pictures say 14 mm square, which means 14^2 = 196mm^2. A bit bigger then most people, and I think a little larger then realistic, but not huge.

Talonz · 10-01-2005, 11:26 PM

Hi, I was wondering if anyone (especially RoboTech or Jaydee) has results to compare between the three accelerator nozzles. I have a MP-05 pro coming in next week, so I would be really interested in which one performs the best. This will be my first setup, so I don't really want to test my luck too much by testing and fiddling around with the mid-plates

As already mentioned, I agree that the MP-05 Pro with the unaligned pins might be a problem. I know you guys have been working for quite a while on the review, but if there are any numbers or even theories on the best midplate (#1 or #2) for the MP-05 Pro, they would be greatly appreciated. I imagine that midplates 1 and 2 would provide a similar performance change between the 270 and 300 pin versions of the MP-05 block.

jaydee · 10-01-2005, 11:43 PM

Edited my post above. I ment 14mm x 14mm and 12mm x 12mm for the dies. Doh... Long day....

jaydee · 10-01-2005, 11:44 PM

Quote:

Originally Posted by Talonz

Hi, I was wondering if anyone (especially RoboTech or Jaydee) has results to compare between the three accelerator nozzles. I have a MP-05 pro coming in next week, so I would be really interested in which one performs the best. This will be my first setup, so I don't really want to test my luck too much by testing and fiddling around with the mid-plates

As already mentioned, I agree that the MP-05 Pro with the unaligned pins might be a problem. I know you guys have been working for quite a while on the review, but if there are any numbers or even theories on the best midplate (#1 or #2) for the MP-05 Pro, they would be greatly appreciated. I imagine that midplates 1 and 2 would provide a similar performance change between the 270 and 300 pin versions of the MP-05 block.

I only have a SP 300 pin version. I would think the jets would still have the same performance ranking though.

Cathar · 10-02-2005, 01:40 AM

Thing is though, relative to the WW-LE, Robotech has the MP-05SP performing better than where Jaydee put it. I would perhaps suggest that the jets are aligned just fine on Robotech's item?

BillA · 10-02-2005, 07:42 AM

fine jet alignment between pins requires locating pins
or the results will be different between EACH assembly of the wb pieces (clearance in the bolt holes)
the results for this wb, just like the Nexxos, will show larger than normal variations
not a good short cut if aiming for maximum consistent performance
- all about the target consumers (price/performance/understanding)

jaydee · 10-02-2005, 09:46 AM

Quote:

Originally Posted by unregistered

fine jet alignment between pins requires locating pins
or the results will be different between EACH assembly of the wb pieces (clearance in the bolt holes)
the results for this wb, just like the Nexxos, will show larger than normal variations
not a good short cut if aiming for maximum consistent performance
- all about the target consumers (price/performance/understanding)

Alighment pins are good as long as the machining is accurate to begin with. It looks to me like the mill used on the MP-05's seems to not make a constant accurate move to the next channel. It could be in the coding, the mills controller or the mill itself (or all the above). Bruce says he has a handle on the problem though.

There seems to be plenty of room for a couple pins though.

dacooltech · 10-02-2005, 10:32 AM

Quote:

Originally Posted by Talonz

Hi, I was wondering if anyone (especially RoboTech or Jaydee) has results to compare between the three accelerator nozzles. I have a MP-05 pro coming in next week, so I would be really interested in which one performs the best. This will be my first setup, so I don't really want to test my luck too much by testing and fiddling around with the mid-plates

I recommend either N02 X or N03 Jets... N02 X nozzle works very good...
You'll get the best results with the N03 jets though...
If you're going to install the block in to an existing system, make sure to flush it first, and then refill it with distilled water + Zerex RSC @ 5-10%

Quote:

As already mentioned, I agree that the MP-05 Pro with the unaligned pins might be a problem. I know you guys have been working for quite a while on the review, but if there are any numbers or even theories on the best midplate (#1 or #2) for the MP-05 Pro, they would be greatly appreciated. I imagine that midplates 1 and 2 would provide a similar performance change between the 270 and 300 pin versions of the MP-05 block.

No worries... the production units are much better than the early review samples... The review samples were sent out in the early phase of the production... the prodcution BPs are also lapped flat with a Peter Wolters AC 500 lapping machine
The N01 mid plate works good with 270pins @ low flow... better with 300pins but it's more restrictive than 270 pins...

dacooltech · 10-02-2005, 10:35 AM

Quote:

Originally Posted by Cathar

Thing is though, relative to the WW-LE, Robotech has the MP-05SP performing better than where Jaydee put it. I would perhaps suggest that the jets are aligned just fine on Robotech's item?

WW-LE vs RBX results confuse me a bit though...
maybe it's because of the lighter springs... Robotech even tried cranking the springs all the way down though... and got better results...

dacooltech · 10-02-2005, 10:39 AM

Quote:

Originally Posted by RoboTech

There was a lot more testing I would have liked to have done, but doing the four CoolTechnica blocks took me almost four weeks and about 100 hrs. I'm trying to maintain a life beyond waterblock testing (family and my day job still come first).

thanks for the excellent review Lee

without a doubt you did great with the units that you have

Quote:

Anyway I hope to see your results soon and maybe we can get pH back in the game one of these days.

heard he's getting a new rosemount dP xmitter... so hopefully he'll be back in the game soon

RoboTech · 10-02-2005, 10:51 AM

Quote:

Originally Posted by jaydee

One thing I am not 100% clear on is how you are taking your differential pressure measurements? I see on your site you have a manometer and I see pressure gauges. What do you use to take the measurements? That is one area I am lacking.

Also do you feel the 14 x 14 die is the right size to use? I am using 12 x 12 right now and I get various answers from 100mm to 1" should be used.

Pretty sure I am going to stick with 12 x 12 for now. Just curious what you think....

I use the SPER Scientific manometer for taking dP readings across the block. I also have a couple analog gauges in the system for reduncy (and because I had them - and the kid in me still likes that stuff). The manometer is much more sensitive, can read out in multiple units, and has an RS-232 port if I ever get around to automating...

I feel two of the main weaknesses with the system I have right now are the potential accuracy of the Omega flow meter and my Haake chiller/home-made heat exchanger requires constant manual tweaking to maintain a steady temp during the tests! Of course the heat transfer rate thru the tube-in-shell heat exchanger changes each time I adjust the flow rate so I have to sit there the whole time to keep the inlet water temp within a +/- 0.010 deg C range. I'm looking at trying a different approach in the future.

Yes, my current die sim is 14mm square. I picked that number after a lot of discussion and input from various people. My thinking... start with a 12mm die (an approximate compromise in size between the different Intel and AMD CPU cores) and add 1mm to each side for 45 deg angle of heat spreading thru a 1mm thick IHS => 14mm sq. Several years ago, 10mm sq. seemed good, but these days most cores are larger than that. Going the full size of the IHS (like my last die sim that had a 32mm sq die) is probably too big and doesn't realistically represent the actual heat flux profile of most real CPU's. Another reason I like "bigger" is because it helps make the copper die more robust (less succeptible to deformation and damge over time).

Talonz · 10-02-2005, 11:33 AM

Quote:

Originally Posted by dacooltech

I recommend either N02 X or N03 Jets... N02 X nozzle works very good...
You'll get the best results with the N03 jets though...
If you're going to install the block in to an existing system, make sure to flush it first, and then refill it with distilled water + Zerex RSC @ 5-10%

No worries... the production units are much better than the early review samples... The review samples were sent out in the early phase of the production... the prodcution BPs are also lapped flat with a Peter Wolters AC 500 lapping machine
The N01 mid plate works good with 270pins @ low flow... better with 300pins but it's more restrictive than 270 pins...

Okay thanks, I think I'll give it a shot with the N02 X nozzle. I don't think the CSP-MAG I have coming in will give me better results with the N03 jet nozzle.

I guess the only things that are bothering me now are the results that jaydee posted between the N01 and N02 nozzles. Maybe the N01 nozzle performed exceptionally well on the 300pin version because the base was so restrictive?

Les · 10-06-2005, 10:13 AM

Some modeling :-

Edit: Added fourth graph

10-01-2005, 02:32 AM	#5
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	Uh, that's what the Hydraulic Power vs C/W graph tries to show you. Shows performance relative to a fixed pumping power, which at a glance provides a decent summary of the performance across a range of blocks with a fixed pump, or at least pumps of similar characteristics. A pressure vs C/W graph favors lower pressure drop blocks. A flow vs C/W graph favors higher pressure drop blocks. A power vs C/W graph treads the middle ground. If you want an exact value you still need to plot the particular pump's pressure-flow curve against the system's pressure-flow curve and find a flow rate, and then look at a flow vs C/W graph to derive a performance figure. The power-C/W figure tries to present a generalised summary of that sort of information though. It is a generalisation, but it's a lot closer to hitting the mark for users who just want to look at a single graph and nothing else, to presenting an even handed appraisal of the efficiencies of various waterblock designs when matched with any given pump. Last edited by Cathar; 10-01-2005 at 02:37 AM.

10-01-2005, 09:46 AM	#6
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Bah, read graphs wrong... Last edited by jaydee; 10-01-2005 at 10:43 PM.

10-06-2005, 10:13 AM	#25
Les Cooling Savant Join Date: Oct 2001 Location: Wigan UK Posts: 929	Some modeling :- Edit: Added fourth graph Last edited by Les; 10-06-2005 at 10:16 PM.

10-01-2005, 01:03 AM	#1
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	AquaXtreme MP-05 triad Review by SystemCooling http://www.systemcooling.com/cooltec...xtreme-01.html Last edited by Cathar; 10-01-2005 at 01:21 AM.

10-01-2005, 01:34 AM	#3
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	The graph that I look to is the power vs C/W graph. Says it all for me really. http://www.systemcooling.com/images/...image31big.gif Last edited by Cathar; 10-01-2005 at 01:56 AM.

10-01-2005, 10:07 AM	#7
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Bah, I am reading the graphs wrong. For some reason I assumed he used the MP-05 SP with all 3 jets. Now I see he did the three blocks with their default nozzels.... I would have liked to see the MP-05 SP with all 3 nozzels done. I got the #1 much better than the #3 but more restrictive aswell.

10-01-2005, 01:03 PM	#9
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	You can see in that picture that the pins are not the same. That is definatly going to mess with the #3 jet if those holes get partially covered by a pin. That will hurt flow and performance some. Here are the pics of the 300 pin base on the one I have. As you can see the pins still very but I don't think they very as badly as Robo's. Next time I take the block apart I will look more carefully to see how the holes on the #3 jet line up. Last time I looked it did look like there was some overlap. I just about got my bench replumed and will run the block again with the new plumbing and silver TIM compound I switched to. Might not be all that usefull though if the retail blocks are now slightly different/improved.

10-01-2005, 10:56 PM	#12
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	I understood that Robotech's die-sim was 14x14mm...

10-01-2005, 11:16 PM	#13
redleader Thermophile Join Date: Jun 2001 Location: The deserts of Tucson, Az Posts: 1,264	All his pictures say 14 mm square, which means 14^2 = 196mm^2. A bit bigger then most people, and I think a little larger then realistic, but not huge.

10-01-2005, 11:26 PM	#14
Talonz Cooling Neophyte Join Date: Aug 2005 Location: Pittsburgh, PA Posts: 11	Hi, I was wondering if anyone (especially RoboTech or Jaydee) has results to compare between the three accelerator nozzles. I have a MP-05 pro coming in next week, so I would be really interested in which one performs the best. This will be my first setup, so I don't really want to test my luck too much by testing and fiddling around with the mid-plates As already mentioned, I agree that the MP-05 Pro with the unaligned pins might be a problem. I know you guys have been working for quite a while on the review, but if there are any numbers or even theories on the best midplate (#1 or #2) for the MP-05 Pro, they would be greatly appreciated. I imagine that midplates 1 and 2 would provide a similar performance change between the 270 and 300 pin versions of the MP-05 block.

10-01-2005, 11:43 PM	#15
jaydee Put up or Shut Up Join Date: Dec 2001 Location: Spokane WA Posts: 6,506	Edited my post above. I ment 14mm x 14mm and 12mm x 12mm for the dies. Doh... Long day....

10-02-2005, 01:40 AM	#17
Cathar Thermophile Join Date: Sep 2002 Location: Melbourne, Australia Posts: 2,538	Thing is though, relative to the WW-LE, Robotech has the MP-05SP performing better than where Jaydee put it. I would perhaps suggest that the jets are aligned just fine on Robotech's item?

10-02-2005, 07:42 AM	#18
BillA CoolingWorks Tech Guy Formerly "Unregistered" Join Date: Dec 2000 Location: Posts: 2,371.493,106 Posts: 4,440	fine jet alignment between pins requires locating pins or the results will be different between EACH assembly of the wb pieces (clearance in the bolt holes) the results for this wb, just like the Nexxos, will show larger than normal variations not a good short cut if aiming for maximum consistent performance - all about the target consumers (price/performance/understanding)

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)