Low/Medum frequency royer induction heater

[RocketScienceSmurf]

I modified an ebay induction heater this weekend in order to try and heat up a small brass pipe. I tried to search the web for information about what frequency to use but I couldn't find much info so I tried a few different capacitor banks to see if I could measure the difference and take it from there.

Firstly I made a new work coil by winding a enameled wire about 1mm in diameter around a glass tube. With the default cap bank of 2uF I got about 70kHz without anything in the work coil and about 80kHz when I had the brass pipe inside it.

I used a lead acid car battery that was charged to around 13V for all tests. "Idle current" was 3,3A and "working current" was 11A so I assume the power transferred to the work piece is roughly 100W. This is probably not very accurate but all I wanted to do was figure out I should aim for a high or low frequency.

Next I soldered an additional 6pcs of capacitors to that backside of the PCB doubling the capacitance. When I fired it up I got 48kHz instead at idle and 60kHz with the brass pipe inside the work coil. Currents were 5A at idle and 19A when "working" so I think that brass responds better at lower frequencies.

The final test was made with a 30uF capacitor salvaged from the welder I bought to get my large three phase transformer. Unfortunately I did not get any current measurements but unloaded work coil gave me 19kHz . When I inserted the brass pipe and fired it up it quickly broke down and one of the mosfets started smoking.

The 30uF is probably not going to survive very long under these conditions but it was fun to give it a try. I will try and replace the mosfets and other components and do some more testing.

[Mads Barnkob]

Great to see you back with more experiments

I think your test voltage is too low, you simply risk blowing the MOSFETs from not switching them fully on. Maybe try with 24V instead?

I seem to recall that the 23.9MB document here: http://kaizerpowerelectronics.dk/tools/file-archive/?drawer=application_notes*induction_heating contains some tables on frequency for different material types.

[RocketScienceSmurf]

Great document! Did you scan that?

Yes I think you are correct about my driving voltage. However, since these are ridiculously power hungry circuits, finding a 24V psu that can drive this thing is not easy or cheap.

I found a really good online calculator that may aid me choosing a frequency: http://www.nessengr.com/technical-data/skin-depth/

I would like something like 1mm penetrating depth in my brass pipe which suggest an oscillating frequency just shy of 20kHz. I am also somewhat limited by my coil design since I only want to heat about 10-20mm of the pipe so I can not wind a 100mm long 15mm diameter coil to get a reasonable inductance of my work coil. One solution could perhaps be to wind a much larger diameter coil and up the psu voltage to compensate for the distance between the work coil and the work piece.

[petespaco]

finding a 24V psu that can drive this thing is not easy or cheap.

Just add another 12 volt lead acid car battery.

[RocketScienceSmurf]

finding a 24V psu that can drive this thing is not easy or cheap.

Just add another 12 volt lead acid car battery.

I would if I had one.

[petespaco]

Getting a car battery:
  Many car batteries are replaced before they are totally shot.
Assuming that you do have a battery charger of some sort:
So, go to a car repair shop and ask if you can pick through their used batteries.  They get $5 or $10 in trade, right?  Prepare to pay that much.  If you tell them what you are doing with them, they may even sell them for less.
   Find one or two that still have some life left.  I know this is a rough "rule of thumb", but if the terminal voltage is above 12.2, it will probably take a charge, at least long enough for experimental purposes.