An Engineer in Wonderland – Alice invents the thyristor

threetransistors-thumb.jpgDoing things with bicycle dynamos is a bit of am obsession with me and several times I have pondered how to supply voltage to things that do not need the whole 3W output.

The trouble with a permanent magnet generator, as these things are, is turning them off as they don’t like to stop generating.

In a bike dynamo, the internal inductance and resistance are chosen to give a 500mA constant current characteristic when feeding a 12Ω load – equivalent to a 6V 2.4W bulb at the front in parallel with a 6V 0.6W bulb at the back.

Output voltage capability is roughly proportional to speed.

So at high speed a large amount of voltage is available to stuff 500mA through what ever impedance is connected.

The esteemed Mr Kurt, a correspondent to this blog, has measured well over 100V from one of these things.

If you want less than 500mA, you have a tricky problem that a brilliant engineer could solve with a switching regulator of some kind.

Alice Towers does not have a brilliant engineer, so I decided that shunting the spare current is the best answer.

A simple voltage clamp will do it, if constant 3W dissipation (and associated pedalling) is acceptable.

More cunning is to short-circuit the generator completely when power is not required, which means 0V at 500mA = 0W dissipation to a first approximation.

Some of these generators have a healthy inductance – 10mH-ish as I remember – which means a great big inductive spike if the shorting switch opens part way though a half cycle.

To avoid most of the spike, and correct me if I am wrong here, the switch should only open close to the current waveform zero-crossing.

After some pondering, this was my first idea:

threetransistors.jpgIt is a bi-stable, latching the mosfet on when the output voltage is too high and only turning it off when the voltage is ok and the current is close to zero.

The zener sets the turn-on voltage for the shunt – effectively the output voltage, and the base resistor sets the turn-off current.

It does need a transistor that can take at least 500mA through its base, as almost all the current flows through it.

Replacing the third transistor with a pnp would save a bit of static power consumption.

Anyway, hours later, it crossed my mind that I had invented the thyristor (aka SCR), so the circuit simplifies to this:

onescr.jpgor, if potential di/dt lock-on problems are to be avoided, this:

twoscr.jpgA pleasant surprise was to find that not only are thyristors still made, but increasingly-forgotten devices are available in surface-mount packages and not just multi-MW pucks for power stations.


Should you feel the need, respond to with ‘thyristor’ in the title.

Don’t respond below because the inbox is overwhelmed with spam.

No email addresses are collected for marketing (or any other) purposes from responses to this blog. I will keep it that way for as long as possible.

Related Tech News

Share your knowledge - Leave a comment