Induction Heater Diy, How To Improve Performance?

[plat]

I have constructed a small induction heater using a pulsed-dc generator that fires a monster power MOSFET which feeds the induction coil.

I actually discovered this by accident when the core of an electromagnet that I was running with this setup burned me, when the windings weren't even hot! It sure heats up metal objects placed inside the coil, but they don't get super hot, barely enough to melt solder.


For some reason, it's just not pulling the power that it could be. The whole apparatus draws maybe 15-20 watts max, while 180 watts is available to it.


So, looking to make this more powerful -

• I have experimented with a variety of different homemade coils of different wire gauges, lengths, and turns, and found no improvement over the first coil. It seems small-gauge wire doesn't work, and more turns is better, but that's about all I can glean from these experiments.
• I have tried a variety of different frequencies. While every frequency I tried produced some heating, it seemed the best around 5kHz.
• I have discovered that placing a capacitor in parallel with the induction coil provides a substantial increase in performance(temp of heated object). I did this originally to absorb inductive kickback to protect the MOSFET and oscilloscope. But, sure enough, the induction heater works MUCH better with this capacitor in place. Anyone know why? The value doesn't even make much difference.

One thing I don't understand. Why does the current draw sometimes increase when I insert a metal object into the coil, and sometimes it decreases when I do the same? It seems to decrease when no parallel capacitor is present, and increase when it is present. Anyone know why this may be?


I have done some research and don't understand the design of commercial induction heater coils. They are usually only a few turns, and very, very large gauge. Wouldn't more turns induce more magnetic flux and work better?


I know the coil is limiting my heater's power, it just needs to draw more current! But shortening it to reduce resistance and impedance doesn't help(increases current draw but not heating power). The wire gauge doesn't seem to make a difference. Basically, if I could get the coil to draw 10 amps instead of 1, it may work much better.

 

[davenn]

I have constructed a small induction heater using a pulsed-dc generator that fires a monster power MOSFET which feeds the induction coil.

a circuit and some photos of your setup would be awesome so that people could see what you are dealing with

 

[kellys_eye]

The best induction heaters are designed for specific purposes/materials as they all rely on resonance to deliver the efficiencies they require.

The signal you use to drive the coil is critical to the system and must equal the resonant frequency of the circuit.

The coil is part of a tuned circuit (LC) and the inductance of the coil will change with any change in its physical properties - including the material used as its 'core' (nominally 'air').

Inserting various items changes the resonant frequency and if you design the circuit to be 'adjustable' you can peak the resonance and get the highest efficiency from it.

With careful design I reckon you could make a peak-resonance detector and have the circuit automatically adjust it's pulsed frequency as items are inserted in the coil and thus maintain a constant resonance. Keeping an eye on the maximum current consumption (tune for peak smoke!) is a very simple manual method though.

But, as mentioned, induction heaters are usually designed for a fixed purpose and peak resonance when that particular fixed core item is fully inserted.

There are formulae for calculating the inductance of a single layer coil wound in free-air - not overly complicated but daunting for the first-timer although there are also many online calculators such as : http://mustcalculate.com/electronics/singlelayeraircoil.php

- and armed with that information ( a figure in μH) you can then add capacitance (μF) and calculate resonance using:



Pulse your tuned circuit at that frequency to get peak power transfer.

You then need to re-calculate L using a parameter for the material used as the core. Then you have yourself a range of frequency that the tuned circuit needs to be adjustable over and can adjust the pulsed signal to get/keep the resonance required.

 

[plat]

From memory, this is more or less the circuit. I have played witth the resistor values some so these may not be exactly how it is setup now.

I can put up pics of the thing tonight.


Kellys, thank you for all that great information. If I'm understanding correctly, I need to tune C3 and L1 to resonate together, and then drive T3 at the same frequency? To achieve maximum power.

 

[duke37]

T1 and T2 will have their bases driven to -12V or so. The limit of base/emitter voltage will be 6V or so. To stop the bases being driven too far negative, put a diode in series with the base or emitter.

 

[kellys_eye]

A simple 555 astable will allow you to change the pulse frequency by varying a single resistor rather than having to adjust two capacitors.

I need to tune C3 and L1 to resonate together

not quite - C3 and L1 will have a resonant frequency according to the formula I posted above (assuming you know what the value of L is).

You adjust your oscillator to pulse at that frequency - put an ammeter in series with the supply and 'tune for peak current' - the peak current consumption being 'resonance'. The peak will be very pronounced and you should be able to see the current consumption rise then fall again as you tune 'through' the peak. This assumes your pulse generator has the range required to encompass the min/max freqnecy of your LC circuit.

When you insert your metal into the coil note the current drop as the LC circuit is 'de-tuned' by its presence. Re-adjust your pulse generator again for peak current.

 

[CDRIVE]

I'm curious. As per the link that Kelly gave you what value inductance does your coil calculate to? Also, what's the value of the parallel Cap?

Chris