# Use resistors to simulate heat from microprocessor?

Hello,

I am playing with fan placement on a computer i am building and i am currently calculating (with solid works flow simulation) how the fans will cool the processor heat sink.

I'm not really a theory type of guy, i like to test real world. Basically what i want to do is simulate the microprocessor's heat without having to attach a real motherboard and over heat the motherboard.

I'm thinking i can take some high power resistors, and attach them to the bottom of the same heat sink i am using for the processor. I will use with LabVIEW and a data acquisition board (because that's what i have laying around) and i will just run a current through the resistor until it heats the heat sink up to the temperature i want, then kind of do a pulse width current with some hysteresis to keep the heat steady. If i get this working i can then vary the fan placement and take note on what it does to the heat sink's temperature.

The problem is, i have a higher power processor (70-95 watts normal not over clocked) and i am assuming i want the resistor to attach to the heat sink with the same surface area as the processor would take up, about 1.4 square inch. because of this i am limited to some smaller dale type resistors, around 25 Watts (again trying to use stuff i have laying around).

So my question is, obviously if i throw more current through these smaller resistors to try to create the heat a 95 watt processor would emit, i would probably de-rate them to a higher resistance but i don't think they would blow up or anything. Do you guys have a better idea of how to heat the heat sink, like an inductive coil or something?

Is it safe to assume that 95 watts of processor power will probably need around the same wattage (roughly) of resistive power, to get similar heat quantities?

Please throw some ideas my way, crazy or pragmatic, as i am just pondering this at the moment for fun.

thanks!

-Hx

sophiecentaur
Gold Member
You can dissipate as much power as you want by using multiple resistors in parallel. Two of twice the required resistance or four of four times the resistance or ten of ten times the resistance. Does that help or is there some layout problem? You will need to clamp them to the heatsink and use heatsink compound to get a realistic simulation, I reckon.

Will you be using a 5V supply? That would require about 1/4Ω or four 1Ω in parallel to give you 100W of power. Or 12V supply with 15Ω for 96W (four 60Ω). I am telling my grandmother to suck eggs here?

It all sounds a bit steamy and potentially burned fingers. As I remember, it's really important to get the right thermal contact with the heatsink because the dissipation rating relies on a certain case temperature for the chip.
My experience was with RF power transistors rather than processors but the same basics apply.

You can dissipate as much power as you want by using multiple resistors in parallel. Two of twice the required resistance or four of four times the resistance or ten of ten times the resistance. Does that help or is there some layout problem? You will need to clamp them to the heatsink and use heatsink compound to get a realistic simulation, I reckon.

Will you be using a 5V supply? That would require about 1/4Ω or four 1Ω in parallel to give you 100W of power. Or 12V supply with 15Ω for 96W (four 60Ω). I am telling my grandmother to suck eggs here?

It all sounds a bit steamy and potentially burned fingers. As I remember, it's really important to get the right thermal contact with the heatsink because the dissipation rating relies on a certain case temperature for the chip.
My experience was with RF power transistors rather than processors but the same basics apply.

aaahh yes. I am a dumbas\$ and forgot wattages increase in parallel. that is basically a good thought i was looking for.

I was planning on using the gold anodized DALE high watt resistors because they have the nice mounting screw holes on them. Then i was going to drill and tap a screw into the heat sink to mount them.

The way i understand thermal paste is that it just fills in all the air pockets where there are deformities in the metal. so if you had two pieces of metal that were machined to a very high tolerance, you wouldn't need any thermal paste as they would mate together perfectly. Same reason if your engine block and head were in an ideal world they wouldn't need a head gasket.
im gona throw some paste on for sure, but all i have laying around is some stuff in a tube that may be older than me haha.