The other day, I was sorting through some old boxes and found a small bag of old transistors. There were two types, both

of them PNP devices. One kind was an OC204 part, made by the old British company Mullard, the other a 2S3220 device made

by Texas Instruments. I presume they date from the 1960s.

 

OC204.JPG

 

gb.JPG

 

2S3220.JPG

 

I wondered if they still worked and what their characteristics would be like compared to a modern device so I built up

this circuit of a simple common-emitter amplifier on a breadboard (shown with a modern 2N3906 PNP transistor) so that I

could test them.

 

circuit.JPG

 

For the base bias resistor (R2), I actually used a pot so that I could adjust the collector voltage to be -4.5V (half the

supply voltage). Both devices needed just 13.6k for R2 to bias the collector to -4.5V, unlike the 2N3904, which in the simulation

needed 430k. The gain worked out to be approximately 7 (at a collector current of 2mA). Yes, that's not a typo, the hFE

really is just 7 for both parts.

 

Here are the waveforms of the 2S3220 amplifying a 10kHz sinewave - yellow trace is the input and the blue trace is the collector (AC coupled)

 

TEK00214.PNG

 

I was curious about the frequency response, so I plotted a crude graph of the amplifier's response for both parts. As you

can see, the response is starting to drop off at just 100kHz, so they'd have been ok for low frequency work - audio

amplifers, etc - but not much use above a couple of hundred kHz.

 

frequency-response.png

 

For contrast, here's what the 2N3906 does in simulation

 

freq-response-2n3906.jpg