Strange results measuring beta

[PCB.Wiz]

It took me a while to solve the puzzle of how to supply both NPN and PNP base current without some complicated switching scheme.

A current regulator also makes this less supply dependent.

You can do it with a single current regulator, by just switching the current path, if you do not mind moving the meter.

It can be useful to also bring out the BASE terminal, to allow you to also measure VBE, as often that is of more interest when matching.
That also allow you to measure diodes, LEDs and Zeners at 10uA test ( 50uA zeners are becoming more common these days)

[Analog Kid]

It took me a while to solve the puzzle of how to supply both NPN and PNP base current without some complicated switching scheme.

You can do it with a single current regulator, by just switching the current path, if you do not mind moving the meter.

Yes, but I was trying to avoid that.
I figured out that I could eliminate the switch by using a different socket for PNP & NPN transistors, but I'd still need a switch, since the current-sense resistor needs to go to the collector (I'm measuring I_C instead of I_E). So this way worked the best.

It can be useful to also bring out the BASE terminal, to allow you to also measure VBE, as often that is of more interest when matching.
That also allow you to measure diodes, LEDs and Zeners at 10uA test ( 50uA zeners are becoming more common these days)

I'm going to be building a more complex test rig that does measure V_BE as well, as part of my continuing exploration into the fascinating world of the transistor.

[PCB.Wiz]

I see the LM334 has a slight temperature variance.
TI give a circuit that reduces that, at the cost of 1 diode drop added.

[Analog Kid]

You must mean this, yes? (from Nat'l Semicircular*)



* Someone I knew who worked at National used to call it that.

[TimFox]

Note that that ap note uses a 1N457 diode, which is quite different from a more popular 1N4148.
Pease had a discussion somewhere about the choice.

[David Hess]

Note that that ap note uses a 1N457 diode, which is quite different from a more popular 1N4148.
Pease had a discussion somewhere about the choice.

The significance of the 1N457 is that it is not gold doped, unlike the 1N4148.  A high conductance low leakage diode is called for here, with modern examples being the 1N458 (still available) and FDH300ATR.  A small signal transistor base-emitter or base-collector junction would also work well.