Radiator Heat Dissipation Questions

I'm trying to figure out what radiators are sufficient, assuming the hottest processors out there. I'm asking for some help in pushing this along, because I only know so much, and can only find so much information that I understand, so bear with me and correct me if I'm wrong

Using the data here, a Pentium 950 D (3.4 Ghz Dual Core) consumes 242 Watts.
http://www.gamepc.com/labs/view_content.asp?id=pd900&page=4&cookie_test=1

The Swiftech MCP655 12v DC Pump consumes 24 Watts
http://www.sidewindercomputers.com/swmc12vdcpu.html


1 watts = 0.86042065 kilocalories / hr (Google.com)

242 + 24 = 266
266 * .86042065 = 228.8718929 KCal/hr

So this means that I need a radiator capable of handling 250 KCal/hr to cool the system? I really feel like I'm missing a whole lot of information, but I have no idea what, Ive been looking through web pages for a couple hours now, and I'm drawing blanks.


The reason for this is that the Black Ice III Extreme is rated for 2445KCal per hour (9702 BTU per hour).

Also, Btu x 0.293 = Watts/hr

Therefore 9702 * .293 = 2842.686 Watts

Under what conditions can the BIX3 work like this? It seems like it would be overkill. MAJOR overkill.

Is there anybody who has done testing to find out what the real life application of a BIX3 yields? For it to dissipate that much heat, they must be running near boiling water through it? What fans are they using? Is it mounted to the front of a rocket?

I also understand that overclocking components increases the heat output, does anybody know of a way to tell how much the heat output will rise?

I feel really lost and I was hoping there was somebody out there that could help straighten my reasoning out

You raise some very interesting questions there. In all honesty, the BIX3 is a great radiator for many applications and you could not go wrong with it. As far as all the calculations you have done, I believe they may be a nulled out because of all the other variables that exist in water cooling, ie. Ambient Temp, Flow Rate, Resistance, Air Flow, Air Pressure, you name it. I believe that with a good Block like the Storm or Apogee coupled with a D5/MCP655, you should reach great temps with the BIX3. You probably know all this but what I am really trying to say is, forget the theoretical numbers because they are exactly what they are, theories.

Get yourself some watercooling goodies and post back with pics!
Bryan D

BIX3 will cool anything computer related. remember, the coldest something can get is ambient, unless you use a chiller of some sort.

I do realize that the computations are all theory, but hey its a start

I'm just trying to understand the concepts behind water cooling, and to buy myself some major components that I won't be looking to replace for some time.

I DO however, have access to some tools that are used for stress testing, flow testing, thermal testing etc. military airplanes (f-22 etc.) and I really wanna know what the real life applications of these expensive *** watercooling components are compared to cheaper counterparts.

I guess all I can say is expect alot of good information from me in the next few months as I buy parts, if anybody wants to donate/loan old parts to the cause, I'd be extremely happy to start the testing sooner. Anyway, if I was looking for donated radiators/water blocks/pumps do you think people would go for this? I could return them when I was done.

I'm just a poor college student so buying all these things out of pocket is not an option for me right now, my dad is a mechanical engineer though so I can use the facilities need be.

I WANT YOUR OPINIONS!!!

Your calculations bring out a point that is often overlooked and which thorilian and others over and over point out. A single 120 rad will cool most setups just fine. It just won't achieve the better temps that you can get with a larger rad like a BIXIII. It depends on what your WC goals are. You may be able to use a 120.1 rad and get the same overclocks as someone with a 120.3 if overclocking is what you are after. If temps and/or silence are more the concern, then a larger rad allows for slower, quieter fans and better temps. A BIXIII is overkill for just about any setup, but it will cool anything that you throw at it.

I understand.. I'm still interested in real life number on the BIX3... anybody know if this has been done already?

http://www.overclockers.com/articles481/index03.asp

I found that, but it appears to be from before there were radiators for cpu cooling in mass production.

I would love to see some hard numbers comparing the BIP.1 and BIX.1 to similar sized copper heater cores people commonly use. From what I gather on various sites on the net including here, a car heatercore will have higher performance and lower head/flow loss. Question is how much, or indeed, IF this is actually the case.

pay absolutely NO attention to what someone says a radiator is rated at . its marketing hype to fool those that dont know any better.

its not applicable because all tests that give the rating are done in lab at the most effecient possible situation and do not reflect real world results of what you will get in your computer.

its simple fact that a single 120 rad can effectively cool EVERY component in your PC with acceptable noise levels . for those that OC you want more surface area to lower your DT ( thats where most people confuse temps with being important ) to disipate more heat and hit higher OCs

keep in mind heat is not always the limiting factor in OCing so while you may get a slightly lower temp your noise may increase dramatically and your OC slightly to none.

this is where more surface area has a benifit UP TO A POINT . diminished returns can rruin the loop if your pump isnt suited for the loop resistance.

more surface area lowers your temps and dt but adds resistance. where most people benifit is that they can add more fans moving less air ( and much less noise )
2 fans at 7 volt are usually quieter than 1 fan at 12 volts ( this is subjective but most people with real world experiance do agree and brand plays a big part in this)

I can support what Thorilan is saying here. I originally thought that my single Chevette H/C would be insufficient for my new Dual processor + 7800GT setup. I was wrong. This rad is perfectly acceptable, and I even my push-pull fans slowed down a bit with a Zalman fan controller (don't know the voltage they're currently at, but def. less than 12). I can't hear the fans AT ALL. Max temp while playing BF2 for several hours is only 43, and I"m overclocked on both the 7800 and the 175 Opty. Adding a second GPU is in the near future. I wouldn't be surprised to see higher full load temp, but in reality so long as I stay well below the max temps of the components it won't matter at all. It's still gonna be the same BTU's regardless of how large the rad is. It did take some adjustment for me, though, to see my idle/load temps at 10 degrees higher than my old rig. Once I got over the unnecessary emotional "loss" of sub 40's temps, I was fine.

But does everybody agree that if I could get numbers on some popular radiators it might help you guys out?

I mean, I've learned so much from reading the forums the past few months, I really wanna give something back

Its hard to say if it would help. I would tend to think that those numbers would become useless each time you try and apply them to a different configuration.

Of course I could be wrong.

citronym said:

Its hard to say if it would help. I would tend to think that those numbers would become useless each time you try and apply them to a different configuration.

Of course I could be wrong.

Click to expand...

I was thinking about more along the line of graphing results for different flows/pressures and temperatures etc.. I still have yet to layout exactly what I want to do, but when I buy my radiator for my setup, I'll post the results I get with that and I'm sure somebody will chime in whether or not the data is useful and I should continue with other models

your missing the point. its not possible as there are other factors oither what the rad can disipate that contribute

Head over to procooling.com and ratch thru the Radiator forums over there, and note any testing methodology you come across...

The complexity of controlling the variables is the reason why what you're asking for hasn't already been done for every brand.

Then you have the problem of which figure is useful... the radiator's C/W vs liq flow vs airflow, or a Btu rating...

ThermoChill HE Testing Data & loose methodology -http://www.overclockers.com/articles778/

Some reference threads on why it hasn't been done yet for all rads:
http://forums.procooling.com/vbb/showthread.php?t=12280
http://forums.procooling.com/vbb/showthread.php?t=10000&page=5

As you can see by the extensive graphs etc, pressure drop, flow rate, turbulence must all be accounted for solely in the liquid flow, before u even look at airflow, and all these variables must be accurately defined to a decent resolution for any results produced to be considered useful. Heck the procooling guys go so far as to account for the friction / flow resistance applied by the fittings used on the rad, tubing, cross-fittings etc for the temp and pressure probes...