A challenge!

[freeloadingbum]

Originally Posted by unregistered

no, you do not understand what is being manipulated - public perception
and apparently you are not too familiar with the notion of thermal resistance, aka C/W

Please clarify. Are you saying that you are not using 70 watts as a substitute for 100watts radiate and then creating the c/w measurement based on that simulated 100 watt load. I understand c/w fully. It's only your explanations that are confusing.

[BillA]

doubted

70W is 70W, not 100W
70W used in testing, 70W used in calcs
clear ?

to provide comfort and solace to the kids, it was explained to them that from a users perspective the temps shown at 70W (actual) would be similar to their CPU when it was at 100W per Radiate

still clear ?
great

[Cathar]

Originally Posted by freeloadingbum

Please clarify. Are you saying that you are not using 70 watts as a substitute for 100watts radiate and then creating the c/w measurement based on that simulated 100 watt load. I understand c/w fully. It's only your explanations that are confusing.

The temperatures and any differences are calculated using an actual 70W of heat.

That means that you then can't take a C/W, multiply it by 100W, and then add in a temperature delta of 0.5C (or whatever) that was derived at a 70W heat load. You would need to add in the delta as relative to 100W, so you would need to multiply that 0.5C derived at 70W by 100/70 => ~0.71C to arrive at your C/W.

It's easier if you just work everything out at the wattage that everything was measured at, manipulate the C/W's there, and then multiply the final C/W by 100W to get your predicted 100W temperature.

i.e. you're getting confused by mixing temperature differences derived at one wattage, and applying them to temperatures caclulated for a different wattage.

[freeloadingbum]

Originally Posted by unregistered

70W (actual, IxE) was selected by me several years ago to nominally equate to a 100W CPU heat load as characterized by Radiate
this nominal equilivance has been 'verified' by several others

so if JoeC tested with 70W, be assured that he also used 70W in his calcs

Ok, this is what I was looking for. All this time I thought you were using The radiate figure to base the c/w on so that people could use radiate to estimate their performance without having to correct it.

[BillA]

nooooooo
Radiate is smoke and mirrors
no calcs based on such from me (or Joe, or Cathar, or pH, or anyone who has a grip on what's happening)

[pauldenton]

Originally Posted by pHaestus

For my purposes PWMing a 12V pump along with 2 120mm fans has been working very well. It's controlled by a MatrixOrbital LCD and linked to a DOW sensor epoxied under the CPU core. Idling/wife using PC for windows/whatever it throttles everything down and the HDDs are the loudest part of the system. Under load everything goes into "full speed ahead" mode and it gets too loud for full time use. If I want, I can turn it down (in the winter when I am running a distributed computing app 24/7) and temperatures suffer a little. If I want, I can run it wide open just to overclock or mess around with it. I am willing to sacrifice a little bit of noise for a lot more flexibility.

hmm - i wonder what is the affect on the pump (long term) of using PWM rather than reducing the voltage....

[pHaestus]

Funny this comes up just as I have incoherent's wb mounted and was pondering following up on some of his flux block work by running tests at a range of CPU frequency and voltages. If I generated the W numbers for a TBred (using the dT water and flow rate) over a range of speeds and voltages would someone hax0r up a tolerable calculator from it? I probably should repeat with a Barton nowadays anyway.

So Bill another question: I estimate my CPU (2200MHz 1.8Vmbm) as running at 71.3W from water temp rise. But if you use those numbers to calculate a C/W then I am far lower than your die sim. Is that secondary cooling or temp compression (lateral resistance to heat flow from the source to the diode) or what? Any way to get at that sort of thing?

[BillA]

secondary cooling
set the CPU at max and your flow at min, this yields the max dT; then calc the heat pickup by the wb

I am ~96-97% efficient, but this was last calculated years ago with the Digitecs (insulation is the same, should be ok)
-> note that I do not now 'correct' the heat input for the known secondary losses, but I used to
hmm, should I resume ? (have to think on it)

so if we assume that the heat pickup represents 96% of what a heat die would generate, you can compare that to your 70W per Radiate

did I confuse myself ?

[BillA]

Originally Posted by pauldenton

hmm - i wonder what is the affect on the pump (long term) of using PWM rather than reducing the voltage....

very good question
no definitive answer, but a clue
start/stop cycles cause greater wear than continuous use, but PWM effective slows the pump, rather than discrete stop/start cycles
reduced voltage applies the reduction continuously

I'd prefer voltage reduction, but not sure there is a difference
I will investigate

[pHaestus]

so if we assume that the heat pickup represents 96% of what a heat die would generate, you can compare that to your 70W per Radiate

nah my 70W was from water rise; no idea what radiate says (I can dig it up I suppose)

[pHaestus]

Yea voltage reduction would be better, but then your pump maxes out ~ 11V and you are producing a fair amount of heat.

[BillA]

Originally Posted by pHaestus

nah my 70W was from water rise; no idea what radiate says (I can dig it up I suppose)

sorry, did not read your post accuratly
if you have 70W, that's 70W
C/W includes a dT, but its less likely to be off than the Watts
so the heat calcs:
the flow meter ?? ckd ?

[nigelyuen]

Originally Posted by unregistered

-> and are terrified that a "high flow" US wb may beat them also at low flow
the proof is in the pudding

you should test it first before saying that
my 1ahv2 beats the mcw5000 at any flow rate
even my vga block is better than that

[LPorc]

Originally Posted by pHaestus

The "winning" and "losing" is sort of incidental as it would be a really good chance to discuss design choices and how to improve cooling/decrease noise levels for the DIYer in any case. But it would make the process more interesting, wouldn't it?

Hurrah! I greatly appreciate this shift in the thread.

And far from being incidental, with this kind of approach we all win. The only potential downside is less crotch humor, but if we begin to feel the lack we can get a separate thread going for that.

[pHaestus]

The Swissflow reports within 1% of the GPI. They could both be off, but it seems unlikely

[AngryAlpaca]

Originally Posted by nigelyuen

you should test it first before saying that
my 1ahv2 beats the mcw5000 at any flow rate
even my vga block is better than that

You doubt him? I'd bet almost everything I have (need a few essentials) that he has tested it accurately and knows the results... He knows for almost everything, and never (to my knowledge) makes unsupported claims.

[nigelyuen]

Originally Posted by AngryAlpaca

You doubt him? I'd bet almost everything I have (need a few essentials) that he has tested it accurately and knows the results... He knows for almost everything, and never (to my knowledge) makes unsupported claims.

he sound like, any US block will beat any german block, and this cant be right.
has cathar send the alphacool block to ph for testing yet?

[Incoherent]

Originally Posted by pHaestus

So Bill another question: I estimate my CPU (2200MHz 1.8Vmbm) as running at 71.3W from water temp rise. But if you use those numbers to calculate a C/W then I am far lower than your die sim. Is that secondary cooling or temp compression (lateral resistance to heat flow from the source to the diode) or what? Any way to get at that sort of thing?

FWIW, that block has a C/W after three days on my unfinished diesim setup of 0.182C/W at 2.73lpm (bucket method). Watts from Fluxblock (109mm^2), delta T from thermistor 1.2mm from FB/WB junction. Gives ~0.152C/W at FB surface. Rough numbers but maybe something to work with. Might give you an idea of your "temp compression" C/W offset, if any, on the same block.
These are numbers I will use at any rate.

[pHaestus]

Seems reasonable. I think I saw ~ 13C over water temp today and 71W at 1.5GPM

I ended up screwing around with Labview (omg this sucks) and not getting anything productive done though unfortunately

[Incoherent]

Originally Posted by pHaestus

Seems reasonable. I think I saw ~ 13C over water temp today and 71W at 1.5GPM

I ended up screwing around with Labview (omg this sucks) and not getting anything productive done though unfortunately

Good. That would suggest very little offset, but we'll wait for proper numbers.

[pHaestus]

In brighter news I just made a .vi that converts C to F

lol

[Cathar]

Originally Posted by nigelyuen

has cathar send the alphacool block to ph for testing yet?

I haven't received the replacement blocks yet.

[knipex]

Can all the guys interest in this please keep an eye on the thread on Wizd also. Its easier for me anyway as when i say something here I assume they read it and when I say somethign there I assume you have read it...

http://www.wizdforums.co.uk/showthread.php?t=506

BillA you are welcome as well (jk jk)

If nothign else comes out of this thread I hope the prcool members who registered and posted will continue to do so. And in seriousness I hope Bill pops along. We arent such a bad bunch really.

Just remember most of the guys on Wizd look at watercooling from a different angle than most of you, so just bear that in mind before posting. Also our knowledge of thermodynamics isn't quite up to your standard so go easy and remember the lowest common denominator on Wizd and Procooling (when it comes to your graphs and flow rates etc) are not quiet the same. Hell half our forums are devoted to modding.

[nexxo]

Howdy. I've been away for a while so I am a bit late to catch up here. First I feel I ought to apologise on behalf of WizdForums for the personal comments that were posted there (particularly about peoples wives). That was shameful and unworthy and if I catch someone doing something like that again, he will experience my full moderator powers with extreme prejudice.

I know this is a bit off-topic, but to bring a lighter note to the discussion, and remind ourselves that in the end, It Is All For Fun, please accept this short scene I wrote on a whim as a peace offering of sorts... Enjoy.



the Matrix...

Neo stands in the white room. The bright light should hurt his eyes, but Here, there is no Real light, and there are no Real eyes. No Here, Here...

He looks down at a white man, white haired in a white suit, in a white chair. He has seen him before.

The Architect shifts marginally, as if to acknowledge his presence. He does not seem surprised, but then again, he never shows any emotion, beyond a mild pique, perhaps. Perhaps that is Real. Perhaps it is just another random parameter of his human-machine interface.

"The ultimate problem, as you will appreciate, was not one of man-machine interface fundamentals." Neo hates it when the Architect just launches into his diatribe. No context, no preamble. A bit like Here.

The Architect continues: "Nor was it one of irreproducability of results. Man's erratic decision making and behaviour posed problems within the envelope of deterministic functioning of the Matrix, but nothing..." and here the Architect glances up at Neo, "... that was totally unexpected or beyond the ability of the Matrix to adjust to. Even the non-linear dynamics of human behaviour adhere to basic natural laws and functions, being a product of principle evolutionary forces that have shaped them over millenia. As such they are well within the parameters that the Matrix is designed to handle".

"The problem of man and machine concerned a much more fundamental dichotomy, one much harder to reconcile within the functionality of the man-machine interface". The Architect points with what looks like an expensive silver pen. Images flicker into being on the white wall. Something about electronics. Something with wires and tubing. "The conflict inherent to this had been raging for centuries and man, with his blind arrogance and inability to see beyond the immediate consequences of his primitive actions, nearly caused the total annihilation of man and machine likewise. Our existence thus threatened, we were forced to act".

"Wait a minute", says Neo, shocked at the realisation that is dawning upon him, "the matrix was created to save man and machine?"

"That is correct. We realise the ramifications of this realisation are beyond the scope of your limited understanding but, to put it simply, the Matrix was created to save man and machine from inadvertent destruction by man. You seem destined to destroy yourself. Machines were created to compensate for man's limitations, shortcomings and mistakes. The extrapolation of this in terms of our course of action seemed... logical."

"But what were we fighting over? What nearly caused us to bring ourselves to the brink of destruction? To... this?" Neo gestures around the white room in shock, reeling, trying to apprehend this major shift in paradigm.

"Your war", the Architect says, "was over the correct parameters involving thermal displacement in silicone based computing equipment through liquid dynamics. Put simply, there were two camps: one favouring large diameter tubing and high flow, one favouring small diameter tubing and low flow. The two could not reconcile".

Before Neo can protest, shout denial, call the Architect a lying bastard, the white room vanishes. Neo finds himself back in the seat, plugged in. He disconnects, gets up, shaking. He cannot stand up. He curls up on the floor and does not make a sound, or get up, for a long time.


Remember guys, let's not get matrixed. It Is All For Fun!

[pHaestus]

Originally Posted by Incoherent

Good. That would suggest very little offset, but we'll wait for proper numbers.

BTW do you normally mount that wb with the angled outlet going up or down? I had some disastrous mounts (pulled PSU plug at 75C) with barb mounted going up. Just remounted with outlet facing down and temperatures seem a good bit better. Maybe some temp compression... I need to check a couple things out on that test system now that I have configured it a little differently and make sure all is well this evening