HV power supply vs. Dead time
Dr. Dark Current, Thu Nov 08 2007, 09:20PMI'm working on another HV PSU, this time I want overcurrent protection and adjustable power via dead time.
As I've fixed my GDT to work with dead time, new problems have arisen.
When using dead-time, the primary voltage waveform looks truly weird and the two HV transformers emit a hissing noise. The "tone" of this noise changes a lot with frequency, I can even make it to emit a low-frequency tone 8-o
The GDT primary voltage also distorts- it shows short burst of nasty ringing-like noise on every turn-on. To make it more strange, this ringing occurs only when switching on the lowside FET.
Here's a pic of the primary waveform w/ dead time captured using a 100:8 transformer:
I should note this is with open load.
I know my problems are usually a bit tricky/complex but if anyone could help I would be grateful
Re: HV power supply vs. Dead time
Marko, Thu Nov 08 2007, 11:38PM
It's OK and expected to see wierd looking waveforms ojn primary of your transformer, simply because transistors go high impedance and not short to ground like with GDT during dead time.
If your transformer is loaded well enough with resistive load you'l see waveform shape approaching dead-timed squarewave, with some smaller kickback after it which gets rectified to DC bus.
If you have lots of reactive energy you won't be able to see the deadtime at all and even with large deadtime the voltage will look like 50% square wave.
Marko, Thu Nov 08 2007, 11:38PM
It's OK and expected to see wierd looking waveforms ojn primary of your transformer, simply because transistors go high impedance and not short to ground like with GDT during dead time.
If your transformer is loaded well enough with resistive load you'l see waveform shape approaching dead-timed squarewave, with some smaller kickback after it which gets rectified to DC bus.
If you have lots of reactive energy you won't be able to see the deadtime at all and even with large deadtime the voltage will look like 50% square wave.
Re: HV power supply vs. Dead time
ragnar, Fri Nov 09 2007, 12:56AM
You *do* have freewheel diodes on the MOSFETs, right....?
ragnar, Fri Nov 09 2007, 12:56AM
You *do* have freewheel diodes on the MOSFETs, right....?
Re: HV power supply vs. Dead time
Steve Conner, Fri Nov 09 2007, 11:21AM
Looks like noise getting into your deadtime thingy and making the duty cycle jump around wildly. If you are using negative feedback, then it's unstable and needs compensation. If you're not using NFB, then the most likely cause is noise is getting into your deadtime pot or the leads to it.
Steve Conner, Fri Nov 09 2007, 11:21AM
Looks like noise getting into your deadtime thingy and making the duty cycle jump around wildly. If you are using negative feedback, then it's unstable and needs compensation. If you're not using NFB, then the most likely cause is noise is getting into your deadtime pot or the leads to it.
Re: HV power supply vs. Dead time
Dr. Dark Current, Fri Nov 09 2007, 05:56PM
Well.. I am having a lot of problems with this dead time power control, ringing and noise is everywhere... and the FETs get very hot even at quite low power.
I think I'm going back to the level-shifted gate driver (which will be done with optocouplers) ,which will allow me to control the power by changing duty cycle, without dead-time (actually I will include circuit to add little dead time to eliminate shoot through but I think it shouldnot make problems).
Dr. Dark Current, Fri Nov 09 2007, 05:56PM
Well.. I am having a lot of problems with this dead time power control, ringing and noise is everywhere... and the FETs get very hot even at quite low power.
I think I'm going back to the level-shifted gate driver (which will be done with optocouplers) ,which will allow me to control the power by changing duty cycle, without dead-time (actually I will include circuit to add little dead time to eliminate shoot through but I think it shouldnot make problems).
Print this page