Big server SMPS spiking / surging on load removal question

[Chris Wilson]

I am curious as to whether uncontrolled voltage regulation on the drains of big FETs used in a push pull low frequency (136kHz) RF amp run in quasi Class D may be caused by the 52V at 57 Amps HP computer server SMPS not liking sudden load removal? I haven't the knowledge to test whether the uncontrolled voltage waveform when the drive to the amp is suddenly removed is from the antenna and loading coil, or from the PS. But from a max of 130V peak to peak square wave on the drains seen on my scope when running at 52V supply I can sometimes see the pretty uncontrolled waveform go beyond 400V before decaying. Could it be these SMPS can't cope with sudden load removal, and is there a way to minimize or fix it? My only thought was a secondary, permanent current draw in parallel with the 20 odd Amps the amp draws, maybe 10 Amps, but that's a lot of heat to dissipate. Any opinions welcome please. Thanks.

[Chris Wilson]

Great reply, thanks very much for going to such trouble. Further testing (by me, someone that is a bit out of his depth here...) suggest it's inductance in the output transformer that is causing it. The choke that feeds the centre tap of the output transformer (push pull output) shows about half the level of the disturbances on its transformer side, and pretty much sod all on the PS side. The drain sides of transformer show the most disturbance.


Thanks again!!