transistor selection

Why do you "have to"? What's the application? What's the load?

John

Probably because if he didn't he wouldn't get a good grade. ;-)

Cheers! Rich

First compute the heat developed in the transistor. That's equal to Vce * Ic. Assume VCE is a sine wave and Ic is a sine wave 180 out of phase. Either evaluate that function and allow for the 180 cutoff (transistor is off for half the cycle. Or dig out somethine like the Radio Amateur's Handbook and find a formula for power in class AB amplifiers. Given the power requirements select your PNP transistor. Holes have lower mobility than electrons so all things being equal the PNP transistor will have lower Ft than a corresponding NPN. Go for the best Ft you can find in a device that can handle to power. Then use 0ja (thermal resistance case to ambient) to compute junction temp on your selected device, with the heatsinks you may or may not need at the ambient temp you expect. I feel that junction temps should be less than 125C, preferably 100C, but others have run devices as hot as 200C. If you cannot show a safe junction temp, you need a bigger package or a bigger heat sink or both. Given a good PNP device, scan the selector guides for a reasonable match in NPN. You want the same case style, the same Ft, the same breakdown voltage and collector base capacitance. Don't worry about hfe or beta. Finally make sure the devices you have selected are really available. Can you get them from Digikey? Are they on the shelf?

David Starr

I assume by "best" Ft, you mean one adequate for the job, not the highest one. Picking one too high will just invite stability problems. Certainly I can get parts that have very high Ft's that will handle the power; doing so seems a waste and an invitation to trouble.

Cheers, Tom