Some practicalities

No figures have been found for the noise (interference) levels that can be expected on the power source.

The following assumptions were made:

  1. High voltage spikes (possibly some KV high either way) have to be expected.
  2. The metals in supply transformers and motors are conventional transformer standard. The rates of rise will be commensurate with this.
  3. Sustained (more than a few milliseconds) voltage surges will be less than 1KV.


It is important to use tried and tested devices. An IGBT rated at 33A continuous current and 1200V sustained voltage is a standard part. Two of these in parallel can give us control of 50Amps with a source potential of 750 volts and some margin for safety.

Some inductance is needed between the switching device and the supply to reduce unwanted energy transfer either way. Without figures to work to the calculation became one of what we could reasonably achieve.

It was soon found that a 50A inductor was seriously unwieldy. Splitting it into two made it much more manageable and incidentally provided a very small resistance that would ensure current sharing between switches.

Inductance was eventually chosen on the basis that it resulted in the largest assembly that was easily handled by one person! The inductor cores could have been made of high grade metal, but since fast events are not expected a substantial cost saving was made by using standard material.

Flywheel diodes were added and anti-parallel diodes to cope with reverse voltage spikes. Such spikes will be of short duration, at least individually, and current will be limited by the resistance of the heating element

Snubbers (R/C networks) were added to reduce the dv/dt at the moment of switching..

Final Comment:

Provision for transient energy (either way) absorption has been the subject of some guesswork. At this stage of any project it is a perfectly normal situation. It is believed that the estimates made are realistic, but only testing in real conditions can confirm this.

Increasing the ability to absorb transient energy is not a problem, but it is a cost.