Made my own waterblock!

[bigben2k]

Originally posted by Graystar
You know alot about watercooling, but you have tons to learn about civility. *That* is something that you need to get a grasp of.

BillA comes with a warning, under his avatar. I think civility went the way of the do-do for him, a couple of decades ago!

I have yet to see anyone be able to razz him the way he has others, but I stick around for that day

But he's still right. You may come up with a number of variations of your block, but the only thing you'll be able to measure, is the relative performance among them, not the relative performance against a known block.

But hey, I don't see why we're even talking about it, you made it clear that you have no intention of picking up a commercial block, so that you can get a more accurate idea of what the actual performance is.

Maybe if someone had a block that they'd be willing to lend you for a while...

Either way, you can still optimize your design, and find out which design variation will give you the best result. I'd stick to it, if I was you.

[Graystar]

Originally posted by unregistered
nooo, really Graystar - trust me
you do not understand what you are babbling about

It is you who is babbling, most likely because you enjoy hearing yourself speak.

Apparently, you are under the delusion that I am comparing my block to others. However, if you were to *actually read* the posts, you would see that I never drew any comparisons to any other system, nor did I make any claims to absolute accuracy. All I did was simply record a starting point that will be used as a base of comparison for upcoming tests.

The power consumption will be the same for all blocks tested. As a constant, any number will do. The water temperature measurements are accurate enough for a hobbyist. The results will be perfectly valid as a comparative measure. But if my process offends you then feel free to never post in this thread again. I assure you, we'll survive without you.

[Graystar]

Originally posted by bigben2k
BillA comes with a warning, under his avatar. I think civility went the way of the do-do for him, a couple of decades ago!

Probably longer! Imagine having to live with him? (cringe)

I have yet to see anyone be able to razz him the way he has others, but I stick around for that day !

Yeah, he would actually have to listen to someone for that to happen. It ain't gonna happen.

But he's still right. You may come up with a number of variations of your block, but the only thing you'll be able to measure, is the relative performance among them, not the relative performance against a known block.

Yes, but we knew this already! Did I ever state otherwise??

If He had his way, we would all just sit on our hands until He spoke and told us what "Is." But that's not gonna happen. He's just not that important. And that upsets him.

But hey, I don't see why we're even talking about it, you made it clear that you have no intention of picking up a commercial block, so that you can get a more accurate idea of what the actual performance is.

Maybe if someone had a block that they'd be willing to lend you for a while...

Either way, you can still optimize your design, and find out which design variation will give you the best result. I'd stick to it, if I was you.

I agree. I will stick to it. I'm interested in seeing the effects of these variations. Grumpy old men can't get *me* down!

[winewood]

Personally I admire Graystar for hanging in there. Hes getting gangbanged.. in some cases for good measure, however he is doing what he should do, listening and working through it. A person of lower calibur wouldn't have done this.
Oh.. and one thing about Bill... This IS his being civil! Thanks for taking the time to explain BillA, I am enjoying your comments. The readership will learn alot from this thread for sure.

edit.. DOH! Graystar... I was so proud there for a second! Your letting them get to you and regressing. Oh well.. just cut the crap and learn something if thats what you want to do, you have some good participation from the good guys in the industry.. dont burn your bridges now...
oh.. and Grumpy men will get you down.. temperatures that is.

[Graystar]

Originally posted by winewood
DOH!

LOL!

Your letting them get to you and regressing.

Possibly. I just don't like when *my* stuff is taken out of *my* context, and then berated because it doesn't fit into someone else's context! The nerve!

But I will pull the reins on my regression and make you proud!

[UnloadeD]

Well, there is the iceberg block to compare it with, although I doubt that will tell us much. I think you should make your variations and find your best one, then send it to BillA for testing. Some people might consider having it tested as proving one person right or wrong. I look at it as throwing more info into the data pool. Sponsorship is available to help defer the cost of testing non commercial blocks, details can be found HERE. It used to be a sticky but has since been moved to "General Cooling Forum" and has slipped into oblivion 8( If you decide to go this route, I'd be willing to help and hopefully others would too. There hasn't been a block of this genre put thru the wringer yet. The results would be interesting.

peace.
unloaded

[Graystar]

Originally posted by UnloadeD
Well, there is the iceberg block to compare it with, although I doubt that will tell us much. I think you should make your variations and find your best one, then send it to BillA for testing. Some people might consider having it tested as proving one person right or wrong. I look at it as throwing more info into the data pool. Sponsorship is available to help defer the cost of testing non commercial blocks, details can be found HERE. It used to be a sticky but has since been moved to "General Cooling Forum" and has slipped into oblivion 8( If you decide to go this route, I'd be willing to help and hopefully others would too. There hasn't been a block of this genre put thru the wringer yet. The results would be interesting.

I agree on the Iceberg1...just a hood ornament at this point.

I have enough material to make lots of these blocks. After the initial design variations, I might make a few to send to people for testing. Not as accurate as official testing, but certainly cheaper and of comparative value to the testers...and more fun too!

[nicozeg]

I just realized why this block created all that struggle; It defies one of the greatest non written laws of the USA: “Bigger is better”

I don’t care if it’s the performance king or not, the greatness of this design is in the ease of making. Lots of DIYers don’t even have a drill press, and really like a low tools design with decent performance.

[Graystar]

Originally posted by nicozeg
I just realized why this block created all that struggle; It defies one of the greatest non written laws of the USA: “Bigger is better”

Yeah. First it was "it will never work because of blah blah blah..." and now it's "it's not working as well as you think because of blah blah blah..."

I don’t care if it’s the performance king or not, the greatness of this design is in the ease of making. Lots of DIYers don’t even have a drill press, and really like a low tools design with decent performance.

You hit the nail right on that head. That was the original intent...something that's easy to make, cheap, and usable. One of the reasons I designed it this way was that I have 8 noisy processors here I need to cool! With 32 bucks worth of copper from OnlineMetals I have enough material to experiment and then outfit all my computers with a block.

I'm hoping to cool 6 computers with one big pump and rad. My main computer (with 2 processors) will get a smaller pump and rad. So far things are shaping up! And don't forget the fun factor in making your own block! I certainly hope people give it a try.

[bigben2k]

...but it wouldn't hurt you to extend the baseplate over the CPU pads.

[Graystar]

Originally posted by bigben2k
...but it wouldn't hurt you to extend the baseplate over the CPU pads.

LOL!! Okay...you're probably right.

[MadDogMe]

First it was "it will never work because of blah blah blah..." and now it's "it's not working as well as you think because of blah blah blah..."

Who said it'd never work?. I've seen no~one 'diss' your block, on the contrary everyone likes it, but sees improvments...

Did you come here for feedback on your block or to get your ass kissed?, ocers.com is probably your best choice for the later so I hope you came for feedback?...

_________________________________________

IMO those pads are'nt worth shit for core protection BB, It's the 4holes mounting that provides stability. A good even pressure on each corner is where it's at (by good I mean 25lbs+). A 'centre' mounting may allow it to 'tilt' under adverse loads on the tubes or block and crush the core, but saying that it should'nt ever happen under normal curcumstances should it?, you'd have to grab hold of the tubes and yank on them...

I'd use silicon tube so there's no extra torque on the block though, my old SK6 spring clip was crap, I had to bend it tighter myself cause the weight of the block used to 'loosen' it up after being on there a while, and PVC tube does have a stubborn 'memory' ...

[Graystar]

Originally posted by MadDogMe
Who said it'd never work?. I've seen no~one 'diss' your block, on the contrary everyone likes it, but sees improvments...

Sorry, I was mixing forums. I first proposed the design at overclockers.com forum and it wasn't well received.

IMO those pads are'nt worth shit for core protection BB, It's the 4holes mounting that provides stability. A good even pressure on each corner is where it's at (by good I mean 25lbs+). A 'centre' mounting may allow it to 'tilt' under adverse loads on the tubes or block and crush the core, but saying that it should'nt ever happen under normal curcumstances should it?, you'd have to grab hold of the tubes and yank on them...

I'd use silicon tube so there's no extra torque on the block though, my old SK6 spring clip was crap, I had to bend it tighter myself cause the weight of the block used to 'loosen' it up after being on there a while, and PVC tube does have a stubborn 'memory' ...

Yes, silicon did work well. I did have a little twisting problem with vinyl tubing. On the next versions I think I'll solder on some tabs that will use the mounting clip to maintain alignment.

[MadDogMe]

Is your clip central?, or is it over to one side a bit?. I'd make up a 'plate' out of polycarb or alu with two holes in it for the inlet and outlet and mount it using the four mobo holes (you'll have to put the tubing on after though). You can apply as much pressure as you want then instead of guestimating with the clip. AMD recomend 24PSI of pressure between heatsink and core IIRC, and I'm sure HSF clips don't provide it...

[Graystar]

Originally posted by MadDogMe
Is your clip central?, or is it over to one side a bit?.

The clip is slightly offset so that the pressure point is in the center of the die.

My board doesn't have holes for 4-point mounting.

[Graystar]

I've had some odd expected and unexpected results from various setups.

Starting point - Original Ahanix Iceberg1 system with WhatBlock?
Tubing - silicone 1/4" ID
Pump - 3 ft. head 150 gph.
Radiator - 3.7" x 5.3" x 1.8"
Fan - YS-Tech Edge Drive, 5800rpm, 36.8cfm, 3.84 watts, .32 amps, 12 volts.

1. From 1/4" silicone tubing to 3/8" vinyl tubing - temperature dropped about 2C

2. From small radiator to Pinto Heatercore with Vornado 510 fan on low - temp dropped 1C

3. Vornado on high - temp dropped another 2C

4. Switching to inline Danner Supreme Mag-Drive 250 pump - no change in temperature.

That last one has me perplexed. The Danner has more than twice the flow and pressure but it didn't make any difference. Some might say pump heat but the other pump was submerged and the Danner was inline. This is something I need to investigate.

I also made a new block with scoring on the base to promote turbulence. It seemed to have no effect on temperatures.

Finally, I added a second waterblock to see the effects on temperature. The Ahanix components were overwhelmed from the heat of the two processors and the computer had to be shut down.

[Tuff]

Quote: That last one has me perplexed. The Danner has more than twice the flow and pressure but it didn't make any difference. Some might say pump heat but the other pump was submerged and the Danner was inline. This is something I need to investigate.



I think I might be able to answer this one...It is my understanding that you can only draw so much heat from the cpu to the waterblock...ie..efficiency of the waterblock. You will have to change the design of your block or Tweak it a bit more.

OR

The water currently being used has drawn all the heat it can from the surface area of the block.

So if you have a pump that meets the spec's already for the removal of heat...the higher rate and pressure pump will not remove anymore heat and would be considered overkill.

Basically it sounds like you have hit the peak performance of your waterblock.

Hmm...I will try to explain this better.

Lets say your cpu is capable of producing 50 watts of heat...and your water block is capable of removing 30 watts of heat under the right conditions...IE water flow...pressure...Rad Perf and water temp.

At this point no mater what outside changes you make other than maybe original water temp..you will not be able to remove anymore heat by changing flow or pressure.

I am going to go out on a limb here....and say...Each waterblock that is made will have a predetermined amount of heat it CAN remove under excellent conditions.

I may be wrong..its just a theory.

Tuff

[Skulemate]

Originally posted by Tuff
I may be wrong..its just a theory.

That's about all you got right with your post
Our CPU cooling problem is primarilly a steady state problem. As such, all of the heat produced by the CPU will be removed by the watercooling system (or HSF, whatever). If this were not the case the CPU would continue to increase in temperature until it fried.

[Graystar]

Originally posted by Skulemate
That's about all you got right with your post
Our CPU cooling problem is primarilly a steady state problem. As such, all of the heat produced by the CPU will be removed by the watercooling system (or HSF, whatever). If this were not the case the CPU would continue to increase in temperature until it fried.

Yup.

I think I know what's up now. It seems like the flow through the block is about the same for the two pumps. I guess my design might be too restrictive. I did intend it to be used with a high pressure pump. I got a hold of one and I'm waiting for the right fittings to hook it up.

Edit - I checked this again and the flow is not the same. The Danner is really moving the water through much faster than the other pump. Yet, temperatures remain the same....

[bigben2k]

Did you resolve the problem of the leaking of the solder? As I remember, you had an issue with your first prototype, where the solder completely clogged the block.

I think that your scoring demonstrates how fins work better.

If I had the tools, I'd make and send you another base, where you'd have a bp of 2-3 mm, and fins that stand 3-4 mm tall, out of a single block of copper.

[Graystar]

Originally posted by bigben2k
Did you resolve the problem of the leaking of the solder? As I remember, you had an issue with your first prototype, where the solder completely clogged the block.

Yes but I still have a clean-up issue. It's hard to get inside the block to clean up the base.

[Graystar]

bigben2k sent me a 1/8" thick copper base to try. After testing, I found that the thick base made no difference.

I had a motherboard running with a standard WhatBlock? and swapped it out for the thick-base WhatBlock? and the temperature remained the same. This is pretty much what I was expecting.

As soon as I gather the parts for a diode reader I'll test again.

[theetruscan]

Have you had the chance to make a block with the same cross-sectional area as your barbs/tubing yet?

[MadDogMe]

But it was'nt less?...

Why not melt it apart and try one with 2mm deep channels in it?, cut them with a junior hacksaw lenghtways about 1mm apart...

[Graystar]

Originally posted by theetruscan
Have you had the chance to make a block with the same cross-sectional area as your barbs/tubing yet?

The only way to do that is to increase the height of the channel, but I don't see how that would benefit. I've yet to make a version with a wider channel. That would increase the area. I'll probably try that next.

Originally posted by MadDogMe
But it was'nt less?...

Well, it was less than the air cooler. But it wasn't less than the other other blocks. If anything, it seemed closer to the next degree up than the other blocks.

Originally posted by MadDogMe
Why not melt it apart and try one with 2mm deep channels in it?, cut them with a junior hacksaw lenghtways about 1mm apart...

Well, that is now getting more complicated than I wanted. But if you want to send me a base I'll try it

Almost forgot...a picture!