2

I have a great open-source board for induction heating here from Fluxeon which uses a Royer oscillator. It does a pretty good job and I already had it delivering power up to 1.5kW. However I continue to get smoked IGBTs and main capacitors due to overvoltage.

schematic

simulate this circuit – Schematic created using CircuitLab

When I measure the drain-source voltage over M1 I get a voltage up to 1kV. I would like to lower it as much as 500V so the capacitor C1 and the IGBTs don't need to deal with such high voltages. I tried putting more load on R3 but this didn't bring the desired effect. The voltage lowered a bit but not nearly as much as I would need. I succeded putting a resistor in series with L4 but this also resulted in high losses on the resistor.

Maybe someone has got some experience and has a great idea dealing with this.

MrLeeh
  • 121
  • 3
  • Where does the 230 V HV supply come from? – Andy aka Feb 23 '17 at 09:03
  • This is net supply voltage (in Germany we have 230VAC) – MrLeeh Feb 23 '17 at 10:21
  • @MrLeeh: but you show HV as a DC source. As much as I know the "net supply voltage" in Germany isn't DC. So HV can't be the same. Where does HV really come from? – Curd Feb 23 '17 at 11:27
  • @Curd. You are right. This is net supply voltage rectified and smoothed by a bridge rectifier and an electrolytic capacitor. – MrLeeh Feb 23 '17 at 12:10
  • 2
    So the DC supply will in fact be 320 volts and not 230 volts. And what makes you think a standard AND gate is supplying enough current to charge up the gate capacitance quickly enough? Where do the logic gates get supplied from? – Andy aka Feb 23 '17 at 12:37
  • @Andy For reasons of simplicity I supplied the model I was using in LTSpice. In the real world circuit I use gate drivers here and of course the logic IC get's some supply voltage. – MrLeeh Feb 23 '17 at 12:47
  • 1
    Then your question is unclear. Always show what you actually implement. – Andy aka Feb 23 '17 at 12:48

1 Answers1

1

The pushpull Royer ZVs osc that you have will make big peak drain voltages when run off rectified mains .The peak DS volts are Pi * Peak DC supply volts .Remember that the peak of the rectified mains could be 1.414 * 230 if the load is light.The tailing of the IGBTs does distort the DS voltage waveform tending to make the peak higher .When you are heating your workpiece the currents flowing in the hot metal are relatively sinusoidal .The HF tank is having sinewave current drawn out of it .The current in your DC feed choke will be relatively constant during the AC cycle .This is because the choke is chosen to have relatively high impedance at the switching frequency and of course low DCR .The self oscillating switching action of the pushpull IGBTs feed the tank with pretty much a square wave of current at the resonant frequency .This means that some more distortion on the tank voltage is to be expected.Standard 1200V IGBTs are looking very borderline here .I have Always fed my ZVs ROYERS with a buck convertor .I measure the DS volts with a simple Diode peak detector and feed it back into the buck convertor control .I have used S trap and used Peak current mode and even voltage mode.You can use anything that has fast current limit and is stable into a relatively low output capacitance .If C out is huge its energy stored at bucked down rectified mains will be capable of blowing the IGBTs anyway despite having fast current limit on the Buck. C1 is under a lot of stress .Its not just the voltage but the frequency .The voltage rating of most caps falls with increasing frequency .This is due to heating caused by High Ac currents heating up due to ESR and dielectric losses .It is common to see many caps in parallel to reduce ripple current heating . Please use an isolation transformer if you start setting all this up .

Autistic
  • 14,235
  • 2
  • 27
  • 65
  • I like your idea of a buck converter. I'm aware of the principle but didn't gather any experience with them, yet. Can you recommend an specific buck converter controller or a schematic for the buck converter control as you described it? (Yes, I'm using an isolation transformer) – MrLeeh Feb 23 '17 at 13:01
  • I used a 600V 38 milliohm N chan mosfet .I detected peak drain current using a source resistor ,I used a 60V 30 milliohm n chan mosfet with a pot on the gate to make a variable resistor so I could move the current up and down .I did not use a chip because at the time the chips would not support my switching loss reduction scheme.I turned on the mosfet with 2 series connected mje 350 pnp bjt .I used a MUR3060 for the freewheel diode . – Autistic Feb 23 '17 at 20:19