0
\$\begingroup\$

I have designed a power converter for a travelling wave tube which regulates a high voltage output of 6kV and 3kVDC well, but when I move onto designing the control system transient response with pulsed load I am having issues with the regulation, and a large voltage jump occurs at the first transient event an it takes a while for it to return back to normal. I am unsure whether I am modelling the output load correctly - it is a travelling wave tube, which transitions from full load current of ~200mA to substantially zero load current (apart from resistor divider currents) at frequencies anywhere from 100Hz to 100kHz, at duty cycles anywhere from 50% to 1% respectively for each condition. I am using LTspice. At the moment, I am simply using a pulsed current source on the output as instructed by Analog Systems.

However I know that when the load is pulsed ON, there should be some voltage droop in the capacitor as it delivers charge to the load. Then when the load is pulsed back on, the capacitor should be recharged back to nominal. I do not seem to be getting this behaviour. I think I may not have modelled the output correctly - should the output caps themselves not be pulsed ON and OFF? Should I have an active switch at the top of the outputs to stop current flow to the load? I will show my current layout. If anyone has any alternative methods which better replicate a travelling wave tube / pulsed output would be greatly appreciated.

If it helps, my topology is a buck-current fed push-pull converter, which has a resonant snubber to reduce switching losses - but it is fixed frequency and regulated by the buck current - maybe this is what causes the large output voltage increase when the transient event occurs? Any advice on this would be appreciated also. See attached for pulsed load as of current.

pulsed load config.

\$\endgroup\$
  • 2
    \$\begingroup\$ Is the converter stable with sufficient phase margin at both load extremes, hi and lo currents, in steady state? \$\endgroup\$ – Verbal Kint Apr 8 at 16:47
  • \$\begingroup\$ How do I check this? It is stable at maximum load current, but when I step the load, there is a voltage jump. It eventually settles back at the reference voltage but it takes too long. I need to have a fast response so that the output voltage is back to a set point before the next transient load pulse occurs. Do you mean I should run the converter with no load current consumption, and see if it reaches the output reference (no pulsing?) \$\endgroup\$ – jvnlendm Apr 8 at 18:44
  • \$\begingroup\$ Non I mean you should extract the control-to-output transfer function of this converter at the maximum current and apply a compensation strategy with crossover and phase margin as design goals. Did you do this? \$\endgroup\$ – Verbal Kint Apr 8 at 18:51
  • \$\begingroup\$ I don't know how to extract control-to-output transfer function of converters. I have tried to ask for help before but it was too mathematical and overwhelmed me. Is there a way to do this in LTSpice? \$\endgroup\$ – jvnlendm Apr 8 at 21:20
  • \$\begingroup\$ LTspice can do it but I am not an expert in doing this. ADI has a landing page on the subject and you may find sources and examples elsewhere. Once you have the transfer function, you can have a look at a seminar I taught in 2018 which covers the subject for the boost converter. \$\endgroup\$ – Verbal Kint Apr 9 at 6:59

Your Answer

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

Browse other questions tagged or ask your own question.