Although that Aliexpress aluminum project box would be very nice for a lower powered device, its a poor choice when you need to dissipate over 40 watts. Because the only flat sides are the ends, it would be difficult to mount a large enough heatsink to it neatly. You cant just bolt the heatsink to the top of the box and fit the MOSFETs to its interior, you'd actually have to cut an opening under the heatsink so the MOSFETs could bolt direct to it. The 76x35mm end plate size constrains the heatsink size you could mount to the back, and the restricted interior volume would make it difficult to mount a fan cooled heatsink internally.
However if the same box with a smooth top or if a taller version with ends large enough to take a CPU heatsink were available they would be very suitable.
Lets take a closer look at your IRILZ44N MOSFET thermal data:
The same basic information is presented twice in differing formats.
The first (highlighted in Tan) is the power dissipation at 25° C and the derating factor. That assumes a heatsink capable of keeping the MOSFET mounting surface under a particular temperature limit. For every degree the mounting surface is above 25° C by, subtract 0.3W from its 45W @ 25° C rating. That's convenient for quick back of the envelope calculations. e.g. if we can keep the mounting surface under 55° C, it can dissipate 45-0.3*30 = 36W, but is a PITA if you actually have or need data for the heatsink.
The second, (hilighted in Yellow) is the design data needed for a more formal solution. T
J_max is 175°C, and the thermal resistance junction to case is R
θJC of 3.3°C/W. (Ignore R
θJA of 66°C/W - its only applicable if you are *NOT* using a heatsink.)
Lets assume a maximum ambient temperature of 45°C (as you have Jordan set as your location), and that we want a 10°C safety margin on T
J_max. That means we can tolerate a temperature rise of 120°C. With an infinite perfect heatsink, perfectly bonded to the mounting surface, that gives us a dissipation limit of 120/3.3 = 36.4W which closely matches the result for 55°C from the first method (55°C = 45°C + 10°C margin), as expected. At this point we already know a single MOSFET cant handle your proposed usage - to get 42.75W dissipation without exceeding T
J_max, you'd need to keep the mounting surface under 33.9°C, with no margin, which is impractical without active cooling - its cheaper to add more MOSFETs.
If you can split the power evenly between N MOSFETs, the calculation for the thermal resistance (to ambient) for N separate heatsinks becomes:
R
θSA = (T
J-T
A)/(P
tot/N) - R
θJC - R
θCSPlug in N=2 , take R
θCS as 0.5°C/W (typical for a TO220 screwed down on heatsink compound to an anodized heatsink), other figures as before, and you get an max individual heatsink R
θA of 1.8°C/W - which is enough to start searching for heatsinks on and distributor's site that has a parametric search. If you want to put them all on the same heatsink, you need to divide the result for individual heatsinks by N, which would give 0.9°C/W.
However, without separate driver OPAMPs and separate source resistors Rs, the power will *NOT* share evenly. Lets guess that it may be out of balance so one MOSFET is taking twice as much power as another and see what that does to the calculations. We only need to calculate for the one that's hogging the power. 120/(42.75*2/3) - 3.3 -0.5 = 0.41°C/W
That's only a 56.7°C maximum heatsink temperature - which may be possible with a big enough heatsink shared between all the MOSFETs, calculating its temperature rise from its R
θSA and the *total* dissipation.
Heatsinks with R
θSA below 2°C/W tend to be somewhat pricey and, apart from mass-produced CPU heatsinks, anything below 1°C/W tends to be really expensive, and for anything at all below 0.5°C/W you'll cry when you see the price.
Home
Help
Search
Login
Register
