BJT Symmetric Electrical Characteristics

In general I think the answer is no. Certainly the base-emitter breakdown is usually small (on the order of 6V); I think that would dominate with the collector junction forward biased.

True!

Only the ancient "INCH" chopper devices had symmetrical characteristics (they were _lateral_ NPN's).

Vertical devices are quite asymmetric due to the progressively higher doping densities as you move from substrate to the surface. ...Jim Thompson

e but

But he said the ones that are not mentioned in the datasheet. VEBO and VBEsat are mentioned.

I always assumed that they were asymmetric for performance reasons, i.e. that you'd want a different doping profile, geometry, etc., for each bit of the thing to make it work best. In fact, I distinctly remember being told that the collector of a transistor was bigger than the emitter because then "it would collect carriers better".

So in a perfect world where it didn't cost anything to have the part put together atom-by-atom, would a high-performance BJT be, in general, symmetric?

If you don't mind low beta. If you examine the beta equations, beta is _proportional_ to....

((emitter-doping)/(base-doping))

and

((base-doping)/(collector-doping))

There's also a "constant" in there which I can't remember without getting out some old text books ;-) ...Jim Thompson

I'm

If you are making BJTs in a vanilla CMOS process, they can be symmetric, with lateral current flow. There is also a vertical current flow to the substrate, but I think the lateral device would break down first due to higher doping levels. Generally these devices were only used where you needed low noise (better than your MOS devices).

So it isn't _just_ because of the way they happen to go together -- it's really intentional, to get performance.

Hey! A guy with a PhD was right about something!

They're good at deriving and writing equations, but not much good at building things that work consistently. ...Jim Thompson

No need to be so precise, they're perfectly possible -- just not often seen. Check out the datasheet for 2SC2878. It exists, I pulled a few from junk before.

Lots of old germanium transistors measure with identical breakdown voltages (typically 30V), probably constructed symmetrically (probably partly because they didn't know or couldn't make better at the time). I have a TO-3 power germanium that's good for about 10A either way...

Tim