Twelve TO220 resistors will overheat when mounted on heatsink with fan ? (20W each)

[treez]

Hello,

I have twelve TO220 resistors all screwed to the same heatsink face (with heatsink paste) , each resistor is dissipating 20W (240W total). The resistor legs are mounted on a PCB which is 145mm by about 50mm. (you can see the resistors and the PCB below)

The resistors are LTO030F22R00JTE3 resistors as follows.......

LTO030F22R00JTE3 resistor datasheet.....
https://uk.farnell.com/vishay-sfern...-film-22r-5-30w-to-220/dp/9567011?Ntt=9567011

Do you think that since these resistors are so closely mounted together, they will overheat..?

The heatsink is a "350AB" type from ABL Heatsinks (25cm long) , as follows ......
https://www.abl-heatsinks.co.uk/index.php?page=extrudedproduct&product=163

There heatsink is vertically mounted with a fan blowing air up from undeneath it.....
The fan is as follows...(4710KL-05W-B50)
https://www.nmbtc.com/pdf/dcfans/4710kl.pdf


PCB comprising the 12 TO220 resistors (with dimensions)
https://i41.tinypic.com/2dlix08.jpg

....you can see the resistors body lying flat as they would be screwed to the heatsink face.

 

[mtwieg]

Okay so say your fan puts out its rated 115CFM of air, which translates to 0.05428m^3/s. Now let's say all that air flow is within the fins of the heatsink (50mm x 125mm is 0.00625m^2), so that gives an air speed of about 8.7m/s. That's very optimistic though, since it doesn't take into account the pressure in the air channel. I would estimate the actual flow to be less than half that, maybe 3m/s at best. According to your heatsink calculator, that gives a conductivity of 0.165C/W. So for 240W that gives a temperature rise of 39.6C. Now for the resistors, the Rjc is 4.2C/W, so you get another 84C of rise there. So at that point you already have a total temperature rise of about 125C, so even operating at an ambient temperature of 25C you'd be right at the SOA of the resistors (150C). Your bottleneck is the interface between the resistors and the heatsink, so using more resistors of the same package or using a bigger package (TO247) would be your best option.

 

[andre_luis]

I think that selected heatsink model has not enough base thickness deep, necessary to distribute this so big amount of heat from resistors to the tabs efficiently.

+++

 

[kam1787]

Are you using thermal grease?

I can not see any photos of the project, but it sounds like the fan is not blowing across the heat sink fins..............

 

[treez]

the fan is fixed beneath the heatsink as it stands vertically, so it should blow along the central gully and the fins.
yes we are using thermal grease

 

[chuckey]

The fan should have a sheet metal duct (like a tapered trough) sitting over the heat sink, to keep the air flow against the fins. has this been done?
Frank

 

[treez]

I took mtwiegs advice and now use 10R resistors, and six heatsinks, so its total of 72 10r resistors (12 per heatsink).

I dont have the duct, but rather the heatsinks will all be in a funnel, and i will try to keep this as tight to the heatsiks as possible.....i intend for the fans to blow air up from beneath the heatsinks and up through the funnel, in which the fins are located....i will try not to have gaps in the funnel

I will try and make "legs " (stilts") for the heetsinks so that they stand high and air can be blown up from under them....also, the fans , i will try and make it so that the heetsinks do not crush the fans which will be under them.

so now each resistor dissipates 9W max, and thats 108W per heetsink. So total heat rise will be 40degC + (9 * 4.2) + (108 *0.306) = 110degsC

That considers a 40degC ambient.

I cannot find a 3D mechanical drawing package for me to use to draw out the hardware, so ill have to just bodge everything together with wooden blocks, bits of cardboard, and aluminium bars etc.