Air cooler load tester

[eltinio]

I'm in the process of building an load simulator to test air coolers. The idea is from the sub-zero section load tester, this one: http://www.xtremesystems.org/forums/showpost.php?p=1487098&postcount=32

It will be an heat cartridge placed in a copper block of 40x40mm and 20mm thick, the block will be insulated to prevent any heat form going out from the sides of the block. The cartridge heater will be supplied with power from an variac power supply so the amount of heat is controllable.

My question is if this is a good way to test air coolers ore is it something that can only be used with extreme cooling.
And will the size of the copper block be a problem? because an Intel IHS is around 3x3cm and an AMD IHS is about 4x4cm.

[inCore]

Testing air coolers is normal too. Testing is usually done with standardized values for "idle" and "load". These values naturally depend on the CPU that you're trying to emulate. As long as you keep these values when testing various coolers it'll be a fair test.

Have a look at this here for some inspiration: http://www.xtremesystems.org/forums/showpost.php?p=3416885&postcount=420

[eltinio]

Testing multiple coolers on the same value was the idea, the exact heat loads was something I need to find out, I need to see how much power the power supply wastes and how much in transformed in actual heat.
I've been thinking about making a plate around the heat loader so i can hook up am2, 775 and 1366 coolers but I might be sticking to one socket with and IHS because it is easyer and there are few socket specific coolers.

And i did saw the topic and topic like that make me feel like switching my education to cnc machinist :P
But will there be an big difference in the heat displaced when it is just one solid piece instead of 2 smaller core like peaces?

[Vapor]

The more closely you can match the physical aspects of a processor the better. In fact, it's probably not worth it to make a load tester if you can't closely mirror the processors your trying to emulate (for air and water cooling testing--their design is well past being optimized to move the most amount of heat, now they're accounting for and optimizing for the physical properties of the processors).

What performs well on Kentsfield/Yorkfield may not do well on Conroe or Wolfdale or even Bloomfield....the opposite is also true in many cases. And who knows with AMD processors, the amount of public thermal testing isn't nearly as high.

Milling an IHS and placing a high quality temp probe into the channel is probably your best bet, tbh. Then to account for any variation in load, you run the tests multiple times.

[eltinio]

Hmm that gave me an idea to make it modular. An regular square heat loader with add on modules that simulate the core design of different cpus covered bij IHS from those cpus.

There is an site with cpu information like core dimensions and pictures but I'm still looking for that.

[DetroitAC]

Already been done...

Mine is the CNC milled block you see, with two 150W cartridges, "dies" to simulate a Quad, and an IHS that I milled off a dead processor. I use thermal grease in between the dies and the IHS, so the thermal resistance is likely higher than with a processor, I'm thinking about soft soldering it together. The idea though, was to make all of the heat flow through the dies into the IHS. Thermocouple probe in the load tester to enable A to B comparisons of cooling devices.

[eltinio]

I'm not looking to be the first one, it will be the first one I make and i want to do the basics right.
Do like your load tester especially the 2 load testers, the heat outputs looks to be something for the sub-zero cooling.

[DetroitAC]

I use this load tester for testing phase units and DI pots, and like you said in post #1 I use a variac to dial it down to whatever I want it to be. In fact, I've never tested anything at 300W, I don't know of anything that puts out that much heat yet... There is nothing about this setup that wouldn't work just fine with an aircooler.

[eltinio]

ok, i'll be waiting until my coper comes in and than make something like you but with a single heat cartridge.

[ThugsRook]

the tough part is keeping a solid ambient temp :wasntme:

[eltinio]

That will be another hurdle to take that haven't really though about. But I can keep the temperature fairly stable in my room witch is well insulated and no draft ore so.

Wasn't it so that 1 degree higher ambient resulted in temps that where 1 degree higher ore is it a lot more complicated?

[ThugsRook]

That will be another hurdle to take that haven't really though about. But I can keep the temperature fairly stable in my room witch is well insulated and no draft ore so.

Wasn't it so that 1 degree higher ambient resulted in temps that where 1 degree higher ore is it a lot more complicated?
its a little more complicated.

at some point 1*C ambient is gonna translate to 2-3*C CPU. im not really sure how thats possible, but i see it all the time in my (fan) testing.

:shrug:

[eltinio]

Al that was to be expected that it wasn't that easy :) . it will be something to keep in mind when looking for a location to test in my house.