# Is 150mW a lot to dissipate from a TO-92 in 50°C free air without heat sink?

I'm designing a circuit in which a 2N3096 PNP BJT in a TO-92 package will drop 15V with $I_C$ of 10mA worst case, for 150mW heat dissipation. Unfortunately, the worst case is when the circuit is idle, which could happen for long stretches if left on continuously, overnight, etc. The internal case air temperature can get to 50°C and there is no fan or case vents. Cooling is through conduction via the aluminum enclosure. The equipment is designed for use in living areas, like a lab bench, so it's unlikely to see external air temperatures above say 35°C

The thermal resistance, Junction-to-Ambient, $R_\text{\thetaJA}$, stated on the datasheet is 200°C/W. I'm taking that to mean the device could get to $50 + (200*0.15) = 50 + 30 = 80°C = 176°F$, which is definitely hotter than I want to touch with my finger.

This strikes me as a bit hot to be running a TO-92 package on a fairly extended basis.

Am I being over-conservative or should I really think about making a change here, either to the current or the package?

• It depends. I have some 2SD471's that left scorch marks on the PCB after 20 years of running at nearly full power (1W). Still the thing (Uniden phone) ran fine. – SX welcomes ageist gossip Sep 26 '15 at 3:47
• Junction temp of 80 celcius is OK .For silicone transistors anything less than 100 C is good .The only way that you could be in trouble is if you have lousey biasing and your 10mA increases with temp AND power disipation is a lot more than 150mW .Thats thermal runaway which was common with Ge transistors but can still happen with Si devices . – Autistic Sep 26 '15 at 7:42