kim_ladha wrote ...

Wow, lots of questions....
The feedback is definitly only possible from the actual tank circuit and only for inductor fed resonant tanks. I have 800 amps going through mine so it is hard to just use conventional techniques. To drop the current i put a loop of coax braid (any copper wire will do) soldered on one of the straight sections of the work coil. The connections are 1" appart and this 'resistivly' drops the current to 50 amps or so- this is simply due to the resistance of the copper pipe! Then i tried several closed ferrites but they all over heated. Finally I used a c core, the air gap stops saturation and it kept cool. This went into a 1:500 current xfmr (cheapo plasic type) to get a few milli amps which i use for my pll signal. (using the term pll loosly). I came up with the concept using pspice- the theory is sound and reliable.

As for the other things. 1). Use long plasic water pipes- water conductivity is not really a problem however
2). You can short out the work coil in my design, the current becomes square wave and switching losses are high- it doesn't break however. Same happens if you loose current feedback.
3). Grounding only represents a problem if the circuit fails and your touching the coil/water. I use an isolation transformer for testing so i don't die- no earth. The coil is live but safe to touch probably due to the high frequency. In a normal failure it will also be safe to touch. There is always a chance that the devices will short so that it is connected to live- although unlikely, I try and avoid touching it. For normal heating work you can just connect it straight to the mains- every device failure I have had in any equipment like this ALWAYS blows a fuse and isolates it. Of course- the time it doesn't could be your last so be careful!
4). No plans to upgrade circuit anymore. I am at the limit of what I can draw from a mains plug socket already! If I wanted to make a 5-6 kw version, the fgh50n6s2d looks like a safe bet.
Karim

GeordieBoy wrote ...

Sensing a continuous RF current in the work-coil that is in the hundreds or thousands of amperes is going to be troublesome. You can get around this in the LCLR arrangement by instead sensing the voltage across the tank circuit. You can do this with a small ferrite voltage transformer that can be conveniently located outside the alternating field of the tank circuit. Use twisted pairs leading to and from the VT to avoid pickup from the IH field. Although this voltage is 90 degrees out of phase with the current, you can compensate for this phase-shift in the control circuit.

Richie wrote ...
...or alternatively a method that Steve Conner suggested to me recently, which is to parallel the tank capacitor with a PP cap of much lower value and then use a CT or sense-resistor to sense the much reduced portion of the tank current passing through this "sense capacitor."

Tonskulus wrote ...


I have been using this method in some of my IH projects. Just o ferrite toroid that has some tens of turns primary winding, paralleled with workcoil. I made several single-ended self excited inductionheater oscillators using this method. Some commerical ind.heaters are using that kind of feedback too.