Back to 34th place!!!

[PiLsY]

Admittedly ive over-reacted a bit. Posted just after I found out my dead geforce3 wasnt accepted for RMA so im going to have to take the shop to court. Not the best circumstances to make a post under, which I shouldve realised form the start. My apologies for going way over the top there.

I think however ive come up with a way to solve this problem. We both have different standpoints on whats important and whats not. The only way I can see to settle this is for both of us to test out the others configuration. I will specify what I belive will be the most efficient configuration for a WC system, and if dorrell agrees he can do the same. I daresay we both have adequate equipment to test out each others suggested config and reach this the good old fashioned way. Just try it .

One thing I must make clear tho is I have always approached watercooling from the standpoint of maximum speed. This is very different than using watercooling as an alternate to air merely to lower the noise. Maybe dorrell takes this stand on it, as reading thru his webby (keep working at it m8 - im sure the site will evolve into something useful for one and all - youre already on the right track ) he seems to run lower clock speeds than perhaps could be attained, which would indicate hes not in it for the overclocking but rather for a quiet stable machine. This i can respect as much as any other outlook. I have no doubt that when not pushing the envelope to the max then one water system is about the same as another. However, when dissipating up to and in excess of 200 watts the barely noticeable differences dorrell mentions become very noticeable indeed. Maybe this is where the "argument" is arising. Looking at what he tests with (xp1600 @ 1534 / 2.2v or something) this is putting out very little wattage, and id expect pretty much every combination of block/pump/rad to cope with this to about the same degree. When going to extremes however small differences tend to become much larger.

Anyway, looking back at this with a cool head I can see that I was very wrong to reply to him in the way I did and for this i offer my apologies. I was far too sensitive to his reply and viewed it as an insult at the time, something which I now see it was not.

I hope this thread can continue as a discussion as I would feel very bad if my input had caused a potentially very helpful and interesting topic to be ignored.

If you wish I will not post on this thread again.

PiLsY.

[PAEricksen]

I'm sure JD would appreciate that Pilsy, and when He gets back next week he may just take you up on that offer to test your ideal config!

I'm just glad we can get back onto the topic of Extreme Cooling! albeit not in the right fourm!

[PAEricksen]

Pilsy: what is your ambient temp with that cpu temp?

Athlon 1600+ @ 1907 (173x11x2.15v)
256mb Crucial PC2100 (2/2/6/2/4-way/4-level/1T) (3.15v)
Abit KR7A w/ vcore mod + vddr mod
R8500 water cooled @ 305/308mhz
Netgear FA311 NIC
Asus ISDN card
Maxtor 30gb 7200rpm HDD
Swiftech MCW462B CPU water block
Swiftech MCW372B GPU water block
Maxijet 750lph pump
Coolercore rad w/4 80mm 40cfm fans @ 7v
Enhance 460w PSU
Windows 2k Pro/98

28c idle / 36c load

[PiLsY]

Rad air temp = 9c
Ambient room temp = 19c
Water temp atm under seti = 12c
CPU temp atm @ 1907mhz/2.13v = 31c
GPU temp atm @ 305/1.95v = 16c

Am at roughtly 70% unit completed atm. Gets hottest at the start of a unit and usually goes up around 2 or 3 degrees.

PiLsY.

[SloNoMo]

both of you guys [ JD and PiLsY ]
need to be a little more considerate of your fellow man , and need to speak a little more precisely...
I'd be a crummy engineer [ only academic- lousy grades, never practicing; and not mechanically inclined ] , but neither really stupid nor illiterate........

??

HOWEVER, the higher the watertemp the less the difference. Colder water obviously has higher thermal absorbtion properties than warmer. I should imagine that given a cpu output of 100 watts that a water temp of over 25 degrees will show very little difference from flow rates. When water temps are more in the region of 10 - 15 degrees you will start seeing HUGE differences. My water temp unloaded is 7c, and loaded is 13c. CPU temp gets up to 47 degrees when i turn 2 fans off and run the other 2 at 7v. running flat out with all 4 fans at 12v cpu temp never exceeds 35c. This is with an output of 134watts from the cpu and 54 watts from the gpu. There are so many variables missing from your posts ...

......

help me out here-- not sure what you were saying...

for water temps ["inlet" ] closer to CPU temps, you'd better have optimum flowrate, and terrific heat-transfer mechanics--no ??

for turbulent- flow devices: "optimum" would mean "faster"...

for systems designed for laminar flow, with veins, channels, or streamlining devices-- one would guess that "optimum" and "maximum" are not the same thing..
----------
and as your water source gets warmer [ e.g.- in summer], one would think that this [ "perfect" flow rates, heat transfer characteristics get really important..

and that a system designed for turbulent flow would be desireable......unless the laminar flow rates can be jacked up sufficiently without going turbulent, and compromising the heat-transfer characteristics of the device...
[it's been a really long time, but memory serves that laminar flow is primarily a low flow- rate situation.. if really high flow rates are required-- when water temp gets closer to CPU temps- turbulent is better..
-----------

and all things being equal [ course, they never are ]- turbulent flow regimes are better for heat- transfer than laminar....

may have to do with the laminar streamlines representing a temperature gradient themselves--like a series of isothermal regions, laid along side each other- each differing by a small T...with effective heat-transfer being related to the need to transfer heat across the small differential T's associated with those little "slices"??..
which is a non-issue, relatively, in a turbulent system...??
--------

SO... are we still in 34th place ?? -- havent looked in a while ...

[PiLsY]

Youve basically summed it up pretty well there m8 . The only thing you have wrong is the water temp thing, which admittedly I explained rather poorly.

With regards to explaining my point. I myself run rather cold water atm due to it being winter and my rad and res being chilled by outside air. If your system isnt optimised then it doesnt actually make much difference running with colder water over warm as youre not actually making use of the colder water. The bottleneck in the system is elsewhere. If your water is warm to start with (standard room temp or higher) then the bottleneck is ALREADY the water temp, therefore changing other variables will make little or no difference at all. This I believe is why Dorrell doesnt agree with my points on flow rate etc. To liken it to air cooling, if your air temp is 35 degrees celcius then no matter how powerful a fan or how big a heatsink you put on it can never go lower than 35 degrees. The bottleneck isnt in the efficiency of the system, its in the cooling medium.

Hope that clears it up for you .

PiLsY.

[EMC2]

Pilsy - both of your first two variables - energy to be absorbed and water temp - relate back to the same basic principles. Maybe a way you might sum it up is to revert back to thermodynamics

The higher the temperature delta, the higher the energy (heat) flux between them will be (given all other variables being equal). And the higher the thermal load, the more pronounced the differences in efficiency of systems will be.

Regarding the *cough*disagreement*cough* between you and JD - the simplistic answer is it depends on the overall system. In some systems it will make a small difference (once above a minimum threshold) and in other systems it can make a very large difference

Laters

[jaydee116]

Hello, just got home from the airport. Don't really have much to respond to as my main point got lost. I will say I have a design enginer that specializes in heat echangers, as that is all he does for work every day, and throws me plenty of info. In fact he has helped me setup the CNC mill and helped make the drawings and code for my first block.

It is real hard to make a decent review of a water block that is acurate enough to make any logical choice for one. One review site may show it to be worth it's wait in gold and another may shout it to be worth dirt.

But I have no arguments with what was said as it really didn't apply to what I was getting at in the first place, but it all sounds good.

Also I make my blocks to where they work well without and out side chilling. Just a rad pump and block. No chillers, pelts, ect... I have NO interest in that type of cooling as the performance gain dosn't justify for me. What my goes are is to build a block that can handle mild overclocking and still run plenty cool and stable and quite. Quite is becoming more important over time.

Also I would have loved to test PiLsY's stuff but it is 9 days to late. I already sold most of my test setup including he XP1600+. I am rounding up enough cash to pickup a decent laptop so I can not miss these conversations on the road!!!

Fixittt is working on a couple test setups that will probably be used to test both our stuff and whom ever else's. Keep in contact and we can send both our stuff to him when he gets it all setup. For now I won't even have a water cooled comp.


Hello EMC2 glad to see you over hear.

PiLsY: Not hard feeling at all here, I never ment to dis you in any way. It was 3:30am when I responded so it may not have came out to well.

I got to go figure out why www.dorrellco.com is offline.