We have experienced a field failure of a TAS5630B when connected to a loudspeaker which uses an autotransformer at it's input.
Our amplifier delivers 100V output through a step up transformer for commercial audio applications.
Normally they connect to isolated transformers on the speaker inputs to step the voltage back down for 8ohm loads and everything works fine.
However a brand of speaker uses an autotransformer instead of isolated transformers to step the voltage down and it is permanently killing the TAS5630B.
Setup:
Our system has an LC filter of 10uH and 680nF on the BTL outputs.
The amp output connects to an output transformer to 100V. Primary inductance = 39mH. Secondary: 253uH
The output transformer 100V is then connected to the autotransformer of the speaker which measures 1.5H on the input and 9.7mH at the output tap for the 8Ohm load.
When playing audio the amplifier switching waveform starts to massively overshoot and undershoot on the amplifier pin outputs (pre the LC output filter)
The autotranformer has a number of taps that change power delivery to the speaker and this changes the input inductance from 1.5H down to 195mH.
The power delivery is small, 7.5W to 30W depending on the autotransformer tap.
In all transformer tappings the amplifier dies, presumably because of the voltage overshoot.
See attached image of overshoot,
The voltage overshoot is measured with the scope probe ground spring to create a very small loop.
Can TI please comment if this is to be expected when connecting an autotransformer like this?
Should we be seeing the switching waveform start to overshoot?
Are autotransformers not meant to be "supported" devices
