Induction Heating/Smelting

Remember, this would be a potentially dangerous build.

Here are some links to various induction heating projects:

Some thoughts on the “theory” behind this…

Why do this?

It’s a bit like the argument of electric heaters vs heat pumps: An electric heater is 100% efficient at converting energy into heat. You can’t get better… so why bother with heat pumps? Because you can move heat with less energy than you need to create it…

So goes with an induction smelter. You could easily dump electrical energy into heating wires and heat the raw material by conduction and radiation. But you will also loose (a lot?) of the heat to other parts of the environment.

Induction causes the ohmic heating to happen in the material being smelted, or in a very close container, reducing heating losses to the rest of the environment.

So more energy is put towards doing what you want.

Electrical Theory

  • remember most formula you easily find are steady state, after the circuit is running
  • don’t forget initial, transient startup! E.g. initial energy stored in C & L of a tank, leading to very high circulating currents
  • voltage is energy/charge
  • current is charge/second
  • Power = I * V, or I^2 * R, or V^2/R. So current has a bigger influence
  • magnetic fields are also more affected by current
  • this sort of makes sense: one electron moving through a substance with more kinetic energy (higher V) is less likely to interact with nuclei (heating); and will create less of a magnetic field (fewer moving particles with quantum spin?) So more electrons moving, even if they have lower kinetic energy, are more likely to hit nuclei (heat it up), or generate a bigger magnetic field
  • Note: electrons actually move super slow, but because of their field, push one “into” the “top” of a wire, and due to electric repulsion, another electron will “pop out” the “bottom”. It’s not the same electron!
  • electric field does change the quantum orbital shells of electrons (eg Zeeman effect-
  • in theory, a capacitor can store and release energy in an electric field with no losses; an inductor can do the same in a magnetic field.

What Gives with all the talk of “ZVS” Drivers

  • an LC (inductor and capacitor) parallel tank have a natural resonant frequency
  • at resonance it appears to have infinite resistance- it draws no current from the power supply
  • BUT when you first apply power, it will take energy from the power supply and charge up the capacitor and inductor.
  • This initial energy will flow back and forward between the capacitor and inductor- it won’t be seen by the power supply! This circulating current can be very high!
  • you can see this on a scope: apply an electrical impulse to an LC tank and it will “ring”. The ring decays as there is parasitic resistance in real capacitors and inductors
  • remember in steady state analysis we say the L & C reactances cancel from the point of view of the external circuit connected to the tank… it’s easy to forget about the initial, transient energy pumped into the system.
  • no real cap or inductor is perfect; there’ll always be an equivalent internal resistance
  • need a way to drive the tank at it’s natural resonance
  • can be done by an external oscillator… but it has to be tuned to the tank, and the two might drift… leading to energy not being added back in at the right time, potentially even killing the resonance and circulating current
  • clever arrangement of transistors can be used that uses the voltage across the tank’s capacitor to alternately drive them on and off at exactly the right moments to add energy in sync with the resonance
  • have to be very careful when you first apply power to this circuit. It relies on no two transistors being exactly the same, and requires that one turns on fully faster than the other… otherwise current will flow through both, neither will switched off, and all the current will go through both of the transistors, quickly destroying them
  • have to avoid power supplies that have a “soft start” - lower initial voltage ramp can cause this situation to happen

So in effect the parallel LC tank is converting Voltage to more current… which is what we want for stronger magnetic fields, and more ohmic heating. Although it is oscillating at quite high frequency- looks like ~90kHz is common.

The “work coil” is connected across the tanks L.

When there is nothing in the work coil, the circuit uses the least amount of current (power) to maintain the resonance.

When a magnetically susceptible item is placed in the work coil, energy will be transferred to it. For all conductive materials there will be some energy transfer due to ohmic heating caused by the induced eddy currents. For ferromagnetic substances there’ll be a lot more transferred due to overcoming the magnetic hysteresis on each flip of the magnetic field- it results in more heating!

This energy reduces the current flowing in the resonant LC tank… which is what the ZVS driver adds back at just the right time.

So when something (conductive) is placed in the work coil, the current load on the power supply goes up.

Note: the banks of caps seen in videos are not to get really high capacitance.

  • strings of caps in series to allow for higher voltage (but reduces overall capacitance)
  • multiple strings in parallel to “get back” the capacitance