Siddhesh Sali

← All work

electronics · simulation

Power Amplifier for B–H Characterisation

A transformer-less, direct-coupled linear amplifier that drives magnetic cores from 50 Hz to 250 kHz, so their B–H loops can be measured across four decades of frequency.

A populated single-sided PCB: black heatsinks along two edges, green wirewound resistors, blue trimpots and through-hole components on a yellow board.

To characterise a magnetic material you drive a toroidal core through its hysteresis loop and measure what comes back — B against H. Doing that honestly means driving it at the frequencies it will actually be used at, which for modern power electronics means well past audio.

That is the difficulty. A transformer-coupled amplifier is easy to build and useless here: the transformer imposes its own low-frequency limit and its own phase behaviour, and you end up measuring the transformer.

So this one has no transformer. It is a multi-stage, direct-coupled linear amplifier — an input stage, a voltage gain stage and a current gain stage — flat from 50 Hz to 250 kHz, with a gain of about 4.6, running from a ±60 V supply and holding 2.5 – 3 A continuously.

Linear, not switching, and that is a deliberate cost. A class-D output stage would be far more efficient and would put its switching noise directly into the measurement. Here the heat is the price of a clean loop, which is why so much of the board is heatsink.

The repository carries the KiCad and Proteus files, the toroid design calculations, the measured frequency response against the simulated one, and a user manual — enough to rebuild it.

Gallery

0505
A schematic divided into three labelled blocks: an input stage, a voltage gain stage and a current gain stage.
Fig. 01Three stages, direct-coupled. No transformer anywhere in the signal path.
The same board with coloured rings drawn on to mark the power input, the signal input and the output terminals.
Fig. 02The version in the user manual, with the connections marked.
A plot of gain against frequency on a logarithmic axis, a nearly flat line a little above 4.5 from 50 Hz to 50 kHz.
Fig. 03Measured gain into an inductive load: 4.52 at 50 Hz, 4.6 across the middle, 4.68 at 20 kHz.
A rendered view of the bare board: four TO-220 devices on heatsinks, three white wirewound resistors, and rows of axial parts on green solder mask.
Fig. 04The board before it existed. Four devices on heatsinks, and the two large emitter resistors.
A B–H hysteresis loop plotted in red, rising steeply through the origin and saturating flat at about ±2,800 gauss.
Fig. 05What the amplifier is for. The simulated loop of our toroid, plotted in MATLAB.