Last week I designed a PCB for the practice keyer. Today, I’ll dive into the nitty-gritty of how it works for real this time.

Here’s the circuit from last time:

This is almost right. I’ll make it fully right next time. The key difference is R1 is labeled as 10K when it should be 100K. I’ll get to that in a bit though… otherwise it’s spot on.

So, how do I know how it works? If I were really much better at this I would have been completely right last time… but I wasn’t. I resorted to throwing the entire kit and caboodle into Cadence’s schematic capture program OrCAD and ran it through its built in spice module.

Zoom in on that to take a better look.

What you see there is a bit more than one full cycle of the oscillator. The output to the speaker is the green line. It’s a roughly 280 Hz signal that’s essentially a pulse waveform.

The key to see in the above graph is the red line that’s sloping up. I’ll get to that in a moment however.

Here’s a zoom of the part with all the action:

Now I’ll get to the good part.  :-)

  • Green – Voltage across the speaker
  • Yellow – Voltage on Q2’s base
  • Red – Voltage on Q1’s base
  • Purple – Current flow out of Q2’s collector
  • Blue – Current out of C1 into Q1’s base
  • Orange – Current flowing out of Q2’s base into Q1’s collector

Basically what happens is current from the battery slowly filters through the 100K R1 trying to forward-bias Q1, but it has to charge the capacitor first. Now Q1, being an NPN transistor will conduct from the collector to emitter if the voltage from the base (middle) to emitter is greater than a certain threshold — around 0.5V in this case. The PNP Q2 is the same, but opposite: it conducts if the base is around 0.5V lower in voltage than the emitter.

Once this happens the magic begins!

  1. Current starts to flow from the collector to emitter in Q1
  2. This pulls Q2’s base low (it was effectively floating before)
  3. Q2 starts to conduct as well which applies power to the speaker
  4. Additionally, that line is connector to the other side of the capacitor and it’s coupled over to the other side since it’s transparent to transients.
  5. At this point we’re almost stable. The current through this new system will flow only as long as there is current flowing on the base-emitter junction of Q1. There’s not enough current supplied through R1, however to keep it going forever. Eventually, the capacitor drains. As it drains the current flowing on the BE juction slows as well. (It drains quickly because there’s no resistor between it and Q1’s base so it has a big current spike draining it out pretty quickly) Likewise, if there’s no current it’ll also turn off.
  6. When the voltage on Q1’s base falls below the threshold, the cycle reverses.
  7. The current in both Q1 and Q2 suddenly stop. Since now the only thing on the right side of C1 is the speaker connected to the negative side of the battery it has a sudden spike to ground.
  8. As before, that transient is coupled to the other side of C1 which pulls it down to around -2.1V.
  9. Now, C1 again starts to charge through R1 until the process begins anew.

A bit of an aside: the timing of this circuit is controlled by both C1 and R1. It’s a classic RC circuit who’s timing is controlled by their values. In a RC circuit like this the time it takes to charge C1 to 63% is the RC time constant. In this case it’s 1e5 Ohms * 5e-7 Farads. This is 0.05s. This is awfully close to the predicted 0.0035s. Why is it different? Because it doesn’t have to charge as much. It only needs to go from -2.1V to 0.5V; full range would take it all the way up to 3V.

Next time: more circuit design — this time in Cadence

If I messed anything up with that explanation I would love to get corrected!