TO92 package power dissipation

[exapod]

I'm design a small board and i need to regulate 5V to 3.3V, i was thinking of using the MCP1700 in the TO92 package because i don't have a lot of space. The maximum current on the 3.3V is 200 mA so the dissipation of the regulator is P = (5V - 3.3V) * 200mA = 340 mW . The datasheet says that the TO92 maximum power dissipation is 644 mW. Do you think that the TO92 is a good choice or i should choose a regulator with a TO220 package?

[digsys]

It will get quite warm, but still in-spec. Make the pads larger, that helps and one "trick" I've used is to -
Use a LDO version of the TO92 with a series 1N4004 diode (Input), to drop another 0.6V and share the heat.

[Witention]

TO-92 isn't meant to be a high-power-dissipating package. 180°C/W isn't going to get you too far, and there often isn't exposed metal to help get power out.

[David Hess]

I came up with 380 milliwatts maximum in a conservative design so the TO-92 MCP1700 will work fine.

[T3sl4co1l]

You can always use the SOT-23 instead

Which isn't as bad as it sounds; with a good solid copper pour on the heatsink terminal (which is... output, I think?), the thermal performance is as good or better than a TO-92, at a fraction of the height.

Tim

[exapod]

Thanks all for the suggestion, at the end i will use a to-220 regulator.

[amyk]

I came up with 380 milliwatts maximum in a conservative design so the TO-92 MCP1700 will work fine.

Datasheet says 131.9°C/W, at 0.38W that's ~50°C rise above ambient... maximum junction temperature 125°C, so there's still a pretty good margin.