Induction heater

[AstRii]

Hello,
i found a very nice coil for induction heating, 8 turns with 1cm copper pipe. I would like to create some induction heater circuit in my spare time. I have a lot of different WIMA capacitors and some FGA75NY60 IGBTs i use in my DRSSTC. I was wondering if you guys could suggest me some schematics. It would be nice if it could be powered from mains and capable of drawing high power. Also i wonder, which is better for induction heating, series or parallel resonance?
Thank you.

[petespaco]

You might begin by setting some goals:

-What do you want to heat?
-How much power  will you need?

Basic assembled ckt boards  (without power supply, coolant pump, etc.):
150 watts? USD$10.00
1000 watts? USD$30.00
2500 watts? USD$50.00
7500 watts? USD$1000
and up, to many megaWatts.

The work coil is about the least complex component of the whole project.  Also, I wouldn't  a solid coil.  You need one that can be water cooled.

IGBT's?   I just read that they work best below about 20 kiloHertz.  If that is true, you may not find them a good choice for induction heating in the lower ranges above.  I hope others will comment on this point.

Start by googling "induction heating" or something like that and spend some time learning.

Pete Stanaitis
---------------

[klugesmith]

I'm also curious about series and parallel resonance mean in the context of induction heaters.

When we connect together a work coil and capacitor, we have a LC resonant circuit.
The two elements are in parallel, and in series, at the same time.  The tank has a resonant frequency and an impedance (V/I ratio) that depends on frequency. [edit] oops!  An isolated LC has only one frequency and one impedance (sqrt(L/C).

Doesn't the distinction between P and S appear only when we connect a third element, the power source?

For example, suppose the LC loop passes through a toroidal core, and we drive another winding on the same core.
A low impedance driver in series with L and C.  Seen from the driving element, the tank impedance is very high at frequency extremes and has a minimum at the resonant frequency.

The alternative is to connect a high impedance driver between the two nodes shared by L and C.
Now we have three elements in parallel.  Seen from the driving element, tank impedance is very low at frequency extremes and has a maximum at the resonant frequency.

Are both cases represented in the world of practical IH's ?

[T3sl4co1l]

Yes.  Note that the series case expects a constant voltage source, otherwise it's not really ringing up or down it's just doing whatever your current is doing; and the parallel case expects a constant current source, otherwise the same thing.  And to drive each with a square wave, to keep current or voltage within reasonable bounds, you need to use a CV or CC inverter, respectively.

There's also a third case, where the current is sourced from a series inductor, driven by a voltage source inverter.  This has the advantage of parallel resonant operation (the capacitor and work coil can be together) without the burden of a current sourcing inverter, but has the disadvantage of more complicated tuning (for a given work coil, you need to choose tank cap, series inductor and transformer ratio; often transformer ratio is kept fixed, putting more emphasis on the inductor), and an additional work coil that's not doing any work but can indeed consume comparable VARs to the work coil itself.  Some of which resonates with the tank cap, but what's left has to be handled by the inverter.  Therefore the inverter needs excess capacity.

The inverter needs excess capacity anyway, in order to offer any useful tuning range, so that's not really a big distinction.

You can save on capacity if you use a fixed resonance, variable supply inverter (rather than a frequency-shift control), but this does require a variable supply, usually a buck converter stage.  Alternately, you can PWM the inverter itself, but this only works out with a tri-level half bridge (the level must be forced to zero, it can't be allowed to freewheel between pulses), or a phase shift PWM H-bridge.

The Chinese induction boards are parallel resonant, with a sort of current-fed inverter.  The current is sourced from series inductors, but they are not resonant, they are part of the DC supply.  The current can be fairly constant from cycle to cycle, but changes in output voltage are quickly (on the order of several cycles) drawn through the inductors, so that increased output load is reflected as increased (DC) input current.  There's no control here, it's just flat out, whatever power the work coil draws at the given supply voltage.

Tim

[Quentief]

Hi,

I am interesting about induction furnace. I would like to know if it is possible to make a high frequency signal directly from the main 230 volt, and then reduce its voltage and increase its current through a transformer, before to power the working coil of the furnace.