Many moons ago I have recreated the schematic of a vintage Bosch ignition circuit, and confirmed it's working as expected with two prototypes (one can be seen here: https://www.youtube.com/watch?v=qzFwiPDNBnE).

The downside of the original design is that the power switching circuit is based on a BJT (T5). My prototypes used BUB323Z and BDW42G, and both required significant heat sinking. In normal operation the entire aluminium case is used as heatsink.

Would replacing T5 with a MOSFET help here? Is there a common part that would fit well? I'm not that familiar with these, and don't want to use a part that looks "about right" (e.g. IRFP450).

Would any design changes around T5 be required/recommended after switching to a MOSFET?


simulate this circuit – Schematic created using CircuitLab

  • 1
    \$\begingroup\$ In that case, that's exactly where Bosch put it in the 1970s. \$\endgroup\$
    – ptashek
    Aug 18 at 22:51
  • \$\begingroup\$ Ah, I missed the capacitor between the collector and the R17/ZD4/ZD5 string. Never mind :) \$\endgroup\$ Aug 19 at 0:25

1 Answer 1


Yes, T5 could be replaced by an NMOS. Caveats:

  1. BJTs are current driven - and here, T5 has a rather low current gain. So it turns off quickly by itself, as the base current is turned off. But MOS devices are voltage driven, and they will not turn off by themselves: the charge from the gate has to be drained.

    It may be sufficient to add a pull-down resistor from the gate to ground. It will have to be a fairly low value, say sized to dissipate peak power = 2x rated average power of the resistor. This pull-down resistance forms a RC circuit with the gate capacitance - so make sure the time constant is only a small fraction of the on-time.

  2. An NPN BJT only needs about 1V of base voltage to fully turn on. This will not be enough for most power NMOS devices - they will either not do much, or operate in linear mode, where they'll dissipate a lot.

    Ensure that when the T5 mosfet is meant to be turned on, there's sufficient gate voltage, according to the datasheet. The gate drive circuit will have to be modified appropriately.

  3. The circuits around the NPN T5 may require the base-emitter diode to be present for correct biasing. In the present circuit, T5's base is a diode to ground. Replacing it with an NMOS gate will require figuring out the effects of the diode removal, and suitable fixes to the circuitry that drives the gate.

  4. Due to gate capacitance, the mosfet may be turning on slower than the BJT did, if the base drive impedance was high enough. Compare the output voltage waveforms between the BJT version and the NMOS version, to ensure they perform similarly both on turn-on and turn-off.

I suggest that you get the simulation working correctly, and then play with replacing T5 with an NMOS of suitable current rating, and suitably low channel resistance, and plot relevant circuit nodes to see what changed. You'll want the rest of the circuit to be biased the same, so this will take some adjustments - it's not a 1:1 replacement even if it may appear to work at first.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.