OCers Radiator Testing

[greenman100]

http://www.overclockers.com/articles1135/

[superart]

I really dont thik such an elaborate wind tunnel setup is neccisary for this application. All you really need is to be able to provide a constant airflow to the radiator. That can be achieved through much simpler means.

Also it would be nice to measure air resistance of the radiator. A lot of people have their incoming case air go through the radiator, and it would be helpfull to them to know how much the rad is hurting their incoming airflow.

[BillA]

suggest you re-read the article, AND follow the links he gives to 2 additional articles;
then think about your post 'cause its all wrong

[BillA]

spending some time on the article reveals some deficiencies:

the air-to-coolant deltaT is not stated, w/o which the #s mean nothing
a dual fan rad was tested with a single fan, NOT representative of use - and the dissipation is WAY LOW
was the fan/RPM/output tested with or without being coupled to the rad ?
how was the blanked off fan opening actually done ?

a good start, but not a good test

[Guderian]

I was wondering how many iterations of the tests they did: the linearity of the 4500rpm graph compared to the other two seems a bit off to me.

a dual fan rad was tested with a single fan, NOT representative of use - and the dissipation is WAY LOW

I assume this was to attempt to aid in the comparison of dual and single fan rads; apples and oranges unite!

[Cathar]

For the CoolWave radiator test, at 30CFM, and assuming 100% radiator efficiency, to arrive at 255BTU (the 1.5gpm figure), the water-air temp delta would have to be at least 4.22C.

Assuming an 85% efficiency which is quite possible, this could imply that the water-air temp delta is held at 5C, however I would rather that it was stated rather than assumed.

I have not heard back from Joe though on the questions I asked.

[lolito_fr]

Following the Be cooling rad test, thought this may be mildly interesting:



(as read off the graphs - no calculations. Not sure what the accuracy of datathief is.)

Bill's tests were done with a 5°C dT...

[HammerSandwich]

Originally Posted by lolito_fr

Bill's tests were done with a 5°C dT...

Not that first test.

Coolant radiator inlet temperature continuously adjusted to maintain 10°C differential (+ or - 0.01°C max) with ambient air as test series progressed.

[lolito_fr]

Ah hah. Next time I won't just look at the pictures...

[Les]

Have also been playing with the BeCooling.
Only used Bill's data
A crude model for the BeCooling 5x5
Radiator is 27x127x19mm , 16FPI (Al fins I think) with 4 Pass 9mm(Diam) tubes
For air pressure drop modeled as 127x127x19mm,16FPI(0.4mm) with 4Pass(19x9mm)Flat Tubes
For Heat Transfer sums modeled as 508x22x19mm,16FPI(0.4mm Al) with Single Pass(19x9mm)Flat Tube




Considered the theoretical effect of varying the Fin Density .
Sums done specifically for a 127x127x19mm radiator with Flat 0.4mm thick Al fins.





Will have to learn how to do these thumbnail things

Edit: Added "Corrugated Fins" Model to Graph 2

[lolito_fr]

Not sure how you do this stuff with Kryotherm...?

The C/W minima for the Papst 4412FGL@12v looks like 16FPI so it would seem the rad is quite well optimised for this type of fannage?

[Les]

Originally Posted by lolito_fr

Not sure how you do this stuff with Kryotherm...?
.....

Dunno whether you want to know but have a go with these "word" things:-
Only considering the Air-side.
Pressure Drop
Considered the 9.525m round tubes as 9.5x19mm flat tubes
Used KryoTesc/Hotside parameters/flow/Heat exchanger thermal resistance(K/W) to set the parameters:
1) Length=19,Width=127,Number of fins=80,Fin thickness=0.4(Guess),Fin height=89(127-4x9.50)
2) Click " Heat emission coefficient" tab for calculations -- adjust "Forcer mounting" to "near the heat exchanger" input "Flow rate" and reset "Coolant"=Air. and set "at temperature" =30.
3) Gives "Hydraulic Resistance"(Pressure drop) among many others. Noted "Pressure drop" and "Heat emission coefficient"[h(convection)]
Plotted "Flow rate" v "Pressure drop"

C/W
For this 4 Pass case(multi tube single Pass I consider as parallel heat-exchangers)
Uncoiled the 127x127 and considered as a 508x31.75
Calculated C/W(convection) separately for each face of the 508x19x9.5mm flat tube.
Divided the "C/W" by 2 to give the Total C/W of the tube.
Used KryoTesc/Hot side parameters/flow/Heat exchanger thermal resistance(K/W) to set the parameters:
1) Length=19,Width=508,Number of fins=320,Fin thickness=0.4(Guess),Fin height=11.1((31.75-9.5)/2) Heat sink base thickness=1(Guess),Base irregularity coefficient=1
2) Set "Heat emission coefficient" to value noted in "Pressure drop" calculations,set "Heat sink material" to "Aluminum" to get "Heat exchanger thermal resistance (K/W)"(C/W(convection) for one tube face)
3) Click "Return" tab, reset "Heat exchanger thermal resistance (K/W)" to 1/2 its value. Input "Total rate of flow", reset "Flow" to Air".
Gives "Thermal resistance of heat exchangers per module (to inlet flow temperatures) (K/W)"(Radiator C/W).
4) Plotted "Radiator C/W" v Air Flow

[lolito_fr]

Ta Les for the tutorial.
Just had a quick play and noticed that you could choose water for the coolant made me wonder if anyone has used this to try modelling fin type waterblocks?

Have not thought about this extensively but imagine it would be interesting even if just to calculate PQ or fin efficiency...

[Les]

Originally Posted by lolito_fr

The C/W minima for the Papst 4412FGL@12v looks like 16FPI so it would seem the rad is quite well optimised for this type of fannage]?

Maybe, but it is a very crude simulation.
Perhaps more useful for some insight into the influence of design parameters.
Some different thicknesses of the same radiator:

Originally Posted by lolito_fr

Just had a quick play and noticed that you could choose water for the coolant made me wonder if anyone has used this to try modelling fin type waterblocks?

Often. Together with Waterloo and Flomerics it is my standard tool for wb modeling.

[Les]

Played some more with a model radiator.
Basic Radiator: 127x127x12.5.mm with 0.4mm Al Fins and 4 Flat 127x19x9.5mm Tubes

1) Introduced corrugated Fins(50% more area).Corrugated in direction of flow - fin spacing remained the same.



2) Considered thinner (0.1mm) Fins and then doubled the number of Tubes to 8; Tube thickness was halved to 4.75mm thus keeping cross-section the same.