[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [N8VEM-S100:2035] Voltage Regulator

On Tuesday, November 12, 2013 3:32:12 PM UTC-6, steve...@gmail.com wrote:
 That is why, I play safe to use expensive 3A regulator as considered as "Heavy Duty".

My gut level guess was that they use the same package so junction-case thermal resistance would be the same, assuming you'd hit thermal limit at 3 amps, the junction rating temp must be higher for the higher current device, but I was completely wrong the limit on current output is actually the short ckt protection hardware on the chip and the STM micro regulator has a surprisingly lower thermal resistance in the same package. (How'd they do that?)

I pulled a Fairchild 7805 and a ST Micro LM323T datasheet and compared them side by side and was surprised the 7805 short ckt protection hardware kicks in at only 2 amps or so, and the thermal resistance of the 7805 is 5 C/W junction to case and the 323T is only 3 C/W junction to case.  Also both junctions are limited to the same standard 125C.

So at 8V in and 5V out thats 3 watts in the junction per amp, so a theoretical 7805 run at 3 amps would have a junction a mere 45 degrees C above the temp of the heatsink, so if you can keep your heatsink below 80 degrees C you're all good (LOL) of course the short ckt protection kicks in at 2 amps or so, you'll never get 3 amps out of a 7805, at least not a Fairchild one.

True that the ST Micro at 3 amps would have a cooler junction at 3 amps only 27 degrees above the (up to 98C) heatsink temp, but what makes it a 3 amp regulator instead of a 1 amp regulator is the short ckt protection components don't kick in until 4 amps or so.

In some ways I think your plan will not work, a 7805 has a lower short circuit limit than the 323T and both are running so cool as a cucumber that they'll never thermally limit no matter what you do (assuming a sane heatsink), so you might want to actually use the 7805 if total current draw is low enough to get away with it if you're trying for overcurrent protection.

Of course just because the junction will be happy if the heatsink is almost hot enough to boil tea, doesn't mean the adjacent electrolytics would be amused at being broiled, so it would be good engineering practice to install the biggest heatsink you can fit in the space.

For a given junction temp it might be cheaper to install a slightly smaller heatsink on a 323T than a larger heatsink on a 7805, which is much more interesting if you're making a 100K board production run than 1 board.  Heatsinks get expensive and a buck can buy a lot of aluminum.

Of course a TO-220 mica insulator has a thermal resistance above 0, and ambient in the area of the heatsink might be warmer than drafty room temp, gotta be pessimistic with thermal calculations for safety...  Still I think you'd have to really screw up your heatsink design before you got thermally limited at 3 amps, its the short ckt protection thats limiting us.