Go Back   Pro/Forums > ProCooling Technical Discussions > General Liquid/Water Cooling Discussion > Water Block Design / Construction
Password
Register FAQ Members List Calendar JavaChat Mark Forums Read

Water Block Design / Construction Building your own block? Need info on designing one? Heres where to do it

Reply
Thread Tools
Unread 09-22-2003, 05:49 AM   #1
watercoolerking
Cooling Neophyte
 
Join Date: Aug 2003
Location: Dalian China
Posts: 56
Default even better than cascade?

cannt see what's the inside look like
Attached Images
File Type: jpg kuehler-66-1.jpg (16.7 KB, 1144 views)
__________________
MSN: watercoolerking@msn.com
watercoolerking is offline   Reply With Quote
Unread 09-22-2003, 06:36 AM   #2
dot
Cooling Neophyte
 
Join Date: Feb 2003
Location: Germany
Posts: 4
Default

Hi,

a guy from Germany called "Tommy the old" did that one. It's called the "Murks3 REV1.1". It was tested here: http://www.watercoolplanet.de/index....urks3%20REV1.1

(all DIY blocks in that list.)

You can find some more pictures here:
http://www.watercoolplanet.de/index....r_user&show=66
and here:
http://www.kaltmacher.de/viewtopic/t...c/start-0.html
(An earlier version, called "Murks2" is pictured. They also discuss the testing of the new version on page 3.)

All commercial blocks which were tested by watercoolplanet.de can be found here:
http://www.watercoolplanet.de/index....0Water%20Block

Although some folks over here question the accuracy of waterblock testing done by watercoolplanet.de it seems they found that Murks3 beats the Cascade by 1.4K.

Details about the testing system can be found here: http://watercoolplanet.de/index.php?open=17&show=52

A thread at kaltmacher.de, a German enthusiasts site can be found here:
http://www.kaltmacher.de/viewtopic/t-22487.html

To sum it all up: testing at watercoolplanet.de might not be as accurate as some people would like it to be. Somebody stated that they exchanged the heating element between tests, so results might not be comparable. :shrug:

So, even if they tested it to be the "best" at the moment, results might not be accurate...
(No offense, Tommy! )

But it seems that the "Murks3" beats a lot of commercially available blocks in Germany at the moment. I'm sure that some manufacturers will soon start their own versions of this type of waterblock. Quite a lot of people here in Germany are experimenting with variations of Cathar's design at the moment, i guess.

We'll see...


Regards,

some stranger

/edit: no pictures of the insides, sorry...
seems he doesn't want to show them just yet.

Last edited by dot; 09-22-2003 at 06:50 AM.
dot is offline   Reply With Quote
Unread 09-22-2003, 07:15 AM   #3
TerraMex
Cooling Savant
 
TerraMex's Avatar
 
Join Date: Nov 2002
Location: Portugal, Europe
Posts: 870
Default just my 0.01 eurocents.

>To sum it all up: testing at watercoolplanet.de might not be as
>accurate as some people would like it to be. Somebody stated
>that they exchanged the heating element between tests, so
>results might not be comparable.

I agree, it doesn't add up. I think you've hit the spot there.

One of the things i could see (and as far as i've seen/read on other threads) , is that loosing the center inlet , with an impingement_based block, kills part of the performance.

And everyone knows , as blocks tend to be very close in performance, misreads and errors (or even some fluke) tend to get an exponential boost. And give wrong results.

But that's just me.
__________________
"we need more cowbell."
TerraMex is offline   Reply With Quote
Unread 09-22-2003, 08:11 AM   #4
Cathar
Thermophile
 
Cathar's Avatar
 
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
Default

Changing the heating element would require the entire test-bed to be recalibrated, and effectively invalidates all new results in comparison with old results. The mere fact that the testbed setup has changed should ring alarm bells louder than an air-raid siren.
Cathar is offline   Reply With Quote
Unread 09-22-2003, 09:27 AM   #5
funktional
Cooling Neophyte
 
Join Date: Jan 2003
Location: Germany
Posts: 17
Default

The heating element was the same for the Cascade and Murks 3 REV 1.1.

As far as i can see the Murks koncentrates the Flow a bit more while the Cascade cooles a larger area which might give the Murks higher flow velocity. The jets will probably be closer to the Bottom. Cathars 4xID theory isnt proved yet and google cant find anything about it...
The Die simulater is 128mm┬▓ on an smaller die the Murks might outperform the Cascade even more
funktional is offline   Reply With Quote
Unread 09-22-2003, 10:19 AM   #6
gone_fishin
Cooling Savant
 
gone_fishin's Avatar
 
Join Date: May 2002
Location: Da UP
Posts: 517
Default

It's shameful that people will base performance on such unscrupulous methods.
Changing the testbench midway through testing just does not cut it.

Perhaps it does perform great, but we sure don't know it by that review.

I do not see the offset inlet not being directly over the jets as a good thing. Is it just so you can see the middle better or what? This adds a restriction before the jets.

http://www.kaltmacher.de/viewtopic/t...c/start-0.html

That link shows the murk3 having 25 cups in a non honeycomb pattern and they do not appear to be precisely lined up. That would be important for jet and cup allignment. Also a gasket material is shown, this would cause each assembly of the unit not to yield the same jet height.
gone_fishin is offline   Reply With Quote
Unread 09-22-2003, 10:48 AM   #7
Cossey3
Cooling Neophyte
 
Join Date: Jun 2003
Location: South London, UK
Posts: 62
Default

1.4k is a big drop therefore to do this would require something special not just more concentrated jets so it does suggest that the testing method isnt as accurate as the site suggests (they are also quoting the results in a way that should only be done if they can be accurate to 0.005k which they certainly are not, the testing page only shows them using a relatively basic setup. also they are using a tiny pump the will nowhere near be giving the sort of flow rate to make the blocks anywhere near what most users would expect (most eheims users would use a 1048 or 1250 which would give much higher flow rates eg a 1250 would give well over 3-4x the flow rate on a system like that)
Cossey3 is offline   Reply With Quote
Unread 09-22-2003, 11:04 AM   #8
dot
Cooling Neophyte
 
Join Date: Feb 2003
Location: Germany
Posts: 4
Default

Quote:
Originally posted by gone_fishin
It's shameful that people will base performance on such unscrupulous methods.
Changing the testbench midway through testing just does not cut it.
As funktional stated, the heating element was the same for cascade an murks3 REV1.1. So the results should be comparable. But I have to agree that changing the testbench (or atleast the heating element) during testing makes the results questionable.

Quote:

Perhaps it does perform great, but we sure don't know it by that review.

I do not see the offset inlet not being directly over the jets as a good thing. Is it just so you can see the middle better or what? This adds a restriction before the jets.
Do you know what "Murks" means in english?
http://dict.leo.org/?p=tLMk.&search=Murks
So don't take everything too seriously.

The offset inlet was chosen to be able to mount the block by a central pressure screw, I think.

The block in its first revision was quite far away from being mass-produceable.
Actually, it was meant to be easily reproduceable even for people who don't have expensive equipment. If I remember correctly, the Murks2 & Murks3 were made only by using a drill press.

Tommy has since changed quite a few things as far as I could tell from the pictures and what he states in the thread at kaltmacher.de.

Quote:

http://www.kaltmacher.de/viewtopic/t...c/start-0.html

That link shows the murk3 having 25 cups in a non honeycomb pattern and they do not appear to be precisely lined up. That would be important for jet and cup allignment. Also a gasket material is shown, this would cause each assembly of the unit not to yield the same jet height.
Yes, this is the first version of this block.
The version tested at watercoolplanet.de is REV1.1.
The block is now sealed by an o-ring, much like the Cascade.
If I remember correctly, the rubber gasket was chosen because you don't need to make a groove in the block for it to seal properly.

Regards,


some stranger
dot is offline   Reply With Quote
Unread 09-22-2003, 11:32 AM   #9
Cossey3
Cooling Neophyte
 
Join Date: Jun 2003
Location: South London, UK
Posts: 62
Default

Quote:
Originally posted by dot



Do you know what "Murks" means in english?
http://dict.leo.org/?p=tLMk.&search=Murks
So don't take everything too seriously.

Yes, this is the first version of this block.
The version tested at watercoolplanet.de is REV1.1.
The block is now sealed by an o-ring, much like the Cascade.
If I remember correctly, the rubber gasket was chosen because you don't need to make a groove in the block for it to seal properly.
i wondered why that name sounded a bit wierd but its too long since i "learnt" german to remember exactly what it meant

the real question is that doest the test setup constitute and average setup, the loop in the setup is basically the same as many people would use but it is a low flow rate system and as with a eheims 1250 the cascade jets are about 4.5m/s then with the setup in this test i would be only ~1m/s which is far too low to get proper inpingement
quote from cathat's cascade thread
"Your pump would probably push around 8 to 9lpm in a full setup at a guess. This is dependent upon the restriction of the block, but let's say 9lpm.

Your total jet-hole orifice area is 132mm^2. At our guesstimate 9lpm, the pump is pushing 150ml/sec, for an average nozzle velocity of around 1.1m/s.

1.1m/s is pretty sedate, and is slower than the rate with which water will flow through something like a Maze 3, which would typically be up around 1.6m/s for a pump like yours in a full system.

Compare that to my block where with an Eheim 1250, the average nozzle velocity in a full setup is around 4.5m/s. With the Iwaki MD-30RZ, it's up around the 7m/s mark."
so with the testing method used the cascade has been crippled whereas murk's block with less jets may well be performing nearer to is optimum if you put a larger pump in that system I would expect a large drop in temps for the cascade and a much smaller for murk's block.
i might be wrong but the test setup doesnt seem to be average for most users (look at the pump round up thread and a 1250 is considered to be the smallest youd want to go)
Si
Cossey3 is offline   Reply With Quote
Unread 09-22-2003, 05:50 PM   #10
Cathar
Thermophile
 
Cathar's Avatar
 
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
Default

The Cascade packs twice the number jets in the same sized area as the Murks 3.1.

What changed between Rev 3.0 and Rev 3.1 that necessitated the ~4C drop in temperatures? The two blocks look near identical from the pictures shown, in terms of jet arrangement and jet density, yet 4.1C is picked up.

Edit: Just did a translation on the Rev 3.1 and Rev 3.0 descriptions. The Rev 3.1 is exactly the same jet configuration as the Rev 3.0, the only difference as stated by Tommy, the block maker was the addition of O-rings.

So once again, how is 4.1C gained between two identical blocks with just the change of the sealing mechanism used for the block?

Last edited by Cathar; 09-22-2003 at 06:00 PM.
Cathar is offline   Reply With Quote
Unread 09-23-2003, 08:25 AM   #11
Tommy the old
Cooling Neophyte
 
Join Date: Nov 2002
Location: Earth
Posts: 5
Default

Quote:
Originally posted by Cathar


Edit: Just did a translation on the Rev 3.1 and Rev 3.0 descriptions. The Rev 3.1 is exactly the same jet configuration as the Rev 3.0, the only difference as stated by Tommy, the block maker was the addition of O-rings.

Faith never a translation by the computer!

The Murks3 REV.1.1 is diffent in nozzle geometry, number of nozzles, the soil structure, but the principle is however still the same like the Murks3.

Could somebody translate the german word "Kanalf├╝hrung" please. It`s also in little different.

Tommy

P.S.
Excuses my bad English
Tommy the old is offline   Reply With Quote
Unread 09-23-2003, 08:46 AM   #12
Cathar
Thermophile
 
Cathar's Avatar
 
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
Default

We have a few conflicting reports here. One thread talks about the heating element being changed and labels the tests conducted as invalid, another says it wasn't.

The pictures of the Rev 3, and 3.1 show the same amount of nozzles.

What makes me question this, is that I tried a great number of combinations of base-plate thickness, tube depths, hole widths, hole geometry, etc, etc. I would hazard a very good guess that whatever geometry you used on the 3.1, I had already tried, or had been very close to.

In all my testing I found variances of around 4C from best to worst, and what was chosen for the final Cascade was the best result of those tests.

Apart from the WCP tests, has there been a different testbed without the shadow of a changed testing environment, confirmed the difference between the Rev 3.1 and Rev 3.0? Even if it was a regular system?

I only ask because I am in a fair amount of disbelief that I could have overlooked a combination that would give the results indicated.

The other possibility I can think of, and it is a vague one, is that the test flow rate is so incredibly low that a certain combination of a thicker base-plate and closer proximity of jet nozzle to cup-base would indeed provide slightly better performance than final Cascade implementation. It is true that there is no one best design for all possible flow rate regimes, and for critically low flow rates (<1lpm) the Cascade will suffer as this is outside of its design parameters.

The second possibility it this. The Cascade was tested using different, more restrictive, barbs to those which it ships with. I was unaware of the test of the Cascade, and had absolutely no control over ensuring that the block tested was configured in the way that I would ship it - that being with its standard large-bore barbs standing off correctly from the manifold chamber, and ensuring that the jets were aligned properly in the cups.

Last edited by Cathar; 09-23-2003 at 08:56 AM.
Cathar is offline   Reply With Quote
Unread 09-23-2003, 09:33 AM   #13
Tommy the old
Cooling Neophyte
 
Join Date: Nov 2002
Location: Earth
Posts: 5
Default

>We have a few conflicting reports here. One thread talks about the heating element being changed and labels the tests conducted as invalid, another says it wasn't.

I dont know anything about that! Only that there was/is a discussion at www.Kaltmacher.de

>The pictures of the Rev 3, and 3.1 show the same amount of nozzles.

could you please link the pictures.

>Apart from the WCP tests, has there been a different testbed without the shadow of a changed testing environment, confirmed the difference between the Rev 3.1 and Rev 3.0? Even if it was a regular system?

Yes, there was a other test with a different testbed, it shows the same result. (Murks better than Cascade)

>The second possibility it this. The Cascade was tested using different, more restrictive, barbs to those which it ships with.

The test with a different testbed was made with the same presure (14KG) and the original barbs (springs?) too.

Tommy
Tommy the old is offline   Reply With Quote
Unread 09-23-2003, 10:46 AM   #14
Cossey3
Cooling Neophyte
 
Join Date: Jun 2003
Location: South London, UK
Posts: 62
Default

Quote:
Originally posted by Tommy the old


Yes, there was a other test with a different testbed, it shows the same result. (Murks better than Cascade)

>Tommy
any details of the other testbed because the one mentioned above is hopelessly under pumped (the flow rate will be far too low to show the cascade at anywhere near its best) maybe this will cause the difference as the murk3 m,aybe tuned for a much lower flowrate which wouldnt work for a cascade

a barb is the bit the the hoses connect to.
eg
Cossey3 is offline   Reply With Quote
Unread 09-23-2003, 11:21 AM   #15
Tommy the old
Cooling Neophyte
 
Join Date: Nov 2002
Location: Earth
Posts: 5
Default

Quote:
Originally posted by Cossey3
any details of the other testbed because the one mentioned above is hopelessly under pumped (the flow rate will be far too low to show the cascade at anywhere near its best) maybe this will cause the difference as the murk3 m,aybe tuned for a much lower flowrate which wouldnt work for a cascade

a barb is the bit the the hoses connect to.
eg
Oh sorry I thought barbs are springs my english is really bad!

It could be possible that the flow rate is too low for the cascade. I thought it was a Hydor 20 - but im not sure!
Tommy the old is offline   Reply With Quote
Unread 09-23-2003, 12:34 PM   #16
dot
Cooling Neophyte
 
Join Date: Feb 2003
Location: Germany
Posts: 4
Default

Quote:
Originally posted by Tommy the old
Oh sorry I thought barbs are springs my english is really bad!

It could be possible that the flow rate is too low for the cascade. I thought it was a Hydor 20 - but im not sure!
Hi Tommy,

glad to see you posting here.

I agree with Cathar - the pump used by WCP is basically too weak to test this kind of block. Both the Murks3 and the Cascade were more or less designed for high flow systems - so it's pointless to test them against each other in a low-flow system. Even if it is necessary to use the same pump for every test (to keep results comparable), I think that nobody would want to waste the Cascade/Murks3's potential by crippling it with a Eheim 1046....

Silly Germans...

Do you have any links to the results of the other test with the different pump for us? I'd really like to see the Cascade tested against the Murks3 in a more "high-flow" setup.

Because, I think, that this is what people over here actually care about...
dot is offline   Reply With Quote
Unread 09-23-2003, 12:34 PM   #17
nOv1c3
Crazy Stupid
 
Join Date: Nov 2002
Location: Dallas texas
Posts: 149
Default

I find it funny that when that site was first posted a while back that know one had questioned the validity of the resuts our the test setup untill now .....
nOv1c3 is offline   Reply With Quote
Unread 09-23-2003, 02:21 PM   #18
Zhentar
Cooling Savant
 
Join Date: Apr 2003
Location: Madison, WI
Posts: 225
Default

Well, when they reach a surprising and unexpected conclusion the first response should be to check that it is a legitimate, accurate test set up.
Zhentar is offline   Reply With Quote
Unread 09-23-2003, 04:26 PM   #19
gone_fishin
Cooling Savant
 
gone_fishin's Avatar
 
Join Date: May 2002
Location: Da UP
Posts: 517
Default

Using such a low flow pump is very strange to say the least. Does it say anywhere on that site what pumps were used to test all the other blocks they have listed?
gone_fishin is offline   Reply With Quote
Unread 09-23-2003, 04:37 PM   #20
Tommy the old
Cooling Neophyte
 
Join Date: Nov 2002
Location: Earth
Posts: 5
Default

Quote:
Originally posted by gone_fishin
Does it say anywhere on that site what pumps were used to test all the other blocks they have listed?
Eheim 1046

see here
Tommy the old is offline   Reply With Quote
Unread 09-23-2003, 05:08 PM   #21
Cathar
Thermophile
 
Cathar's Avatar
 
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
Default

Well there's a lack of information on which to comment any further.

The pictures for the Murks 3 and 3.1 can be find linked to in the "Gallerie" section of the respective sections on this page:

http://watercoolplanet.de/index.php?open=4&show=2

nov1c3, the Kaltmacher thread linked to is full of discussion by the German enthusiasts over the accuracy of the testbed.

All I can say is that the results of the Murks 3, from what I saw of its pictures and description, were about what I expected. To then achieve >4C gain on top of that, is in my mind, nigh on totally unbelievable. I'm not saying that it absolutely is not possible, that would be silly, but right now my level of doubt is bordering on extreme.

Tommy, also I had asked if another testbed had verified the size of the difference between the Murks 3 and 3.1.
Cathar is offline   Reply With Quote
Unread 09-23-2003, 05:54 PM   #22
Tommy the old
Cooling Neophyte
 
Join Date: Nov 2002
Location: Earth
Posts: 5
Default

>Tommy, also I had asked if another testbed had verified the size of the difference between the Murks 3 and 3.1.

No sorry! The original Murks3 works after the test on WCP until now in my computer.

But believe me the Murks 3.1 is inside rather differently to Murks3! It only looks outside similar.
Tommy the old is offline   Reply With Quote
Unread 09-24-2003, 12:19 AM   #23
Cathar
Thermophile
 
Cathar's Avatar
 
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
Default

Quote:
Originally posted by funktional
Cathars 4xID theory isnt proved yet
I still have the 2xID, 3xID, 4xID (shipping), and 5xID blocks here.

As a result of the above I've taken to throwing them back onto the testbed after a quick cleanup.

I can reconfirm that 2xID is "about" 0.5C worse than 3xID, and 3xID is "about" 0.5C worse than 4xID. Actually slightly less than 1C separates 2xID from 4xID. 5xID performs about the same as 4xID (no statistical difference), but 4xID is cheaper to manufacture (less milling time), so that's why I stuck with it.

Flow rates used are ~4LPM and ~10LPM, representing an Eheim 1048 and an Iwaki MD-30RZ (50Hz) pump pushing through the Cascade blocks and two Camry cores in series.

Baseplate thickness does have an impact too. I didn't rerun the baseplate tests again. I have 4xID baseplate thicknesses in 0.7mm, 1.2mm, 1.7mm and 2.2mm. 1.7mm and 2.2mm perform worse at the flow rates I test with. 0.7mm favors the higher flow rate, but 1.2mm catches up at the lower flow rate, so a final base-plate thickness was chosen that met the target flow rate design criteria, being anywhere from 4-10lpm.

Other tests previously done focussed on hole geometry (honeycomb/grid), including hole inter-spacing. Also tested were jet ID to cup ID ratios ranging from 1:1.75 to 1:3, for which over a number of different hole geometries a singular consistent value kept showing up as being "ideal".

Also tested were various manifold chamber depths. Also lapping finishes which is still an ongoing development for me and some improvements have been made since the first batch of blocks from the one tested at WCP came from.

So you can see that there was a wide range of variables and combinations tested. As said before, in all that testing I saw about a 4C variation from best to worst, but never saw anything that could ever account for more than a further 1.5C over the final Cascade, which itself represents a practically machinable peak of that work.

Oh well. At least this thread has caused me to rethink and triple check my earlier work. I still have my grave doubts.
Cathar is offline   Reply With Quote
Unread 09-24-2003, 11:12 AM   #24
Cossey3
Cooling Neophyte
 
Join Date: Jun 2003
Location: South London, UK
Posts: 62
Default

Tommy> can we have a look inside the 3 v1.1 then please
Cathar> well atleast this is keeping you on your toes.
Cossey3 is offline   Reply With Quote
Unread 09-24-2003, 01:59 PM   #25
leejsmith
Cooling Savant
 
leejsmith's Avatar
 
Join Date: Mar 2003
Location: notts uk
Posts: 408
Default

Is there any one that could do another test of the cascade and murks 3.1 besides billa ?
__________________
Folding , Folding , Folding !
leejsmith is offline   Reply With Quote
Reply


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

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT -5. The time now is 07:39 PM.


Powered by vBulletin® Version 3.7.4
Copyright ©2000 - 2018, Jelsoft Enterprises Ltd.
(C) 2005 ProCooling.com
If we in some way offend you, insult you or your people, screw your mom, beat up your dad, or poop on your porch... we're sorry... we were probably really drunk...
Oh and dont steal our content bitches! Don't give us a reason to pee in your open car window this summer...