If one transistor is good, then two must be better

Ah, time for a second transistor. There are lots of things we can do with a second transistor. One simple thing, in terms of keeping the resistor count low, is to build a Darlington pair. I heard about these when I heard about the ULN2803A chip, which people sometimes use for driving motors. It turns out that a Darlington pair is where you hook up the two collectors of two transistors, and the emitter of one to the base of the other. So a tiny current goes into the base of the first transistor, a medium current goes out of the emitter of the first into the base of the second, allowing a big current to flow through the third. Usually, the two transistors are on the same chip – you can get single components called Darlington transistors with three legs just like a normal discrete BJT. However, because it’s always fun to make your own, it’s schematic time:

Here, I use a 1k resistor for R1 and a 1M resistor for R2. This gave a very shallow slope when I plotted a graph several posts back, using a setup similar to the one above, but with a smaller transistor. I still have the data for that graph, so let’s dig it out, and also plot a similar graph for the new setup:

Ah yes, that seems to have more-or-less the desired effect – the inverter is now far less wimpy, and produces a nice low voltage at the ADC when the DAC voltage is turned up. There are two things of note: first, the flat bit at the top lasts for about 1.2V rather than about 0.7V. To turn the transistors on, there needs to be enough voltage at the DAC to cope with the voltage drop across two transistors, not just one. The second thing – the voltage at the ADC doesn’t bottom out at 0V, more like 0.75V. I’m not quite sure why that is. It seems, reading the wikipedia page, that this effect is expected.

Still, let’s do some maths, with some quick approximations. The current going through that 1M resistor can’t be more than 5µA, whereas the 1k resistor could well have about 4mA going through it. Possibly if we used a smaller resistor, we could get an even bigger current, and still get 4-and-a-bit volts across it. This is, of course, the big idea behind a Darlington pair – to use a small current to control a large current.

A thought – the two transistors in the Darlington pair are doing different jobs, in that with the first transistor, you care about what’s going on at the emitter, whereas with the second transistor, it’s what’s going on at the collector that’s important.

Anyway, next time I should add some more resistors. It’s not as if they are expensive…