Impedance matching woes / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Impedance matching woes

1 2 3 next

Move Thread

LAN_403

Dr. Shark

Wed Apr 25 2007, 09:16AM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

I am playing with a little series resonant induction heater setup. Problem is, I don't know how to control the impedance my half bride sees when I tune for resonance. What happens right now is that the heater "sort of" works when I am off-resonance and hard switching, but as soon as I hit resonance, I am drawing an insane amount of current and my impedance matching transformer melts.
I figure that I want my half-bridge to see more of the resistance in the work piece, so I increased my impedance matching transformer to a 40:2 turns ratio. While this is a lot more than Steves

12 turns, it makes sense since I am running off recitified 230V mains, and he was running off 30V or so.

Any solution to this problem besides increasing the turns ratio even further, like playing with leakage inductance and series capacitor values I a non-random fashion?

Here's a view of my li'l halfbridge

Steve Conner

Wed Apr 25 2007, 10:09AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

You could try going to 40:1 or even more. The only other thing I know of that makes a difference is tighter coupling to the workpiece and a lossier workpiece. Increasing the characteristic impedance of the tank circuit (more turns on work coil, smaller tank cap) might help too, but I'm not sure.

Dr. Shark

Wed Apr 25 2007, 11:29AM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

Whlie you are at it Steve, what is the reason you had two resonant capacitors, one in the primary and one in the secondary side of the transformer? From my understanding an induction heater is not a Tesla Coil, so I don't want coupled oscillators with the energy sloshing back and forth. I just want a single resonant circuit so I can boost the current (not sure if I really want this though

) and more importantly do soft switching on my MOSFETs.
And since the current is a lot lower in the primary side I assume I am a lot better of having the resonant capacitor there - after all a 5V 1kA RF capacitor is not exactly an off-the-shelf item.

Another thought: What is the common method to keep the heater from blowing up when there is no work piece in the coil? With a non-magnetic metal I cannot rely on the frequency going off resonance, so the current in the resonant tank will explode. Is there an easier way than using some kind of overcurrent detection to run an interrupter?

Marko

Wed Apr 25 2007, 11:52AM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

What I don't understand is why you guys keep using a ferrite transformer and series resonant circuit wich is poor choice for application where workpiece is to be changed.

Removal of workpiece with series LC will cause massive increase in Q, impedance drop and blowage of the bridge.

For any proper induction heater I would really prefer richie's design:

Paralell LC is so much more grateful... It if buletproof for any change in load or practically any short or open in main LC-circuit.

Resonant rise in the LC increases the current to massive values required for heating, while driver needs to feed very low current at high voltage.

Matching inductor enables us to pump any amount of power we want into workpiece with constant supply voltage (it makes the LC look ''more series'').
It is small and carries only supply current.

It also provides immunity for any short or open in the tank crircuit, if it happens by a chance.

Dr. Shark

Wed Apr 25 2007, 12:17PM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

Hi Marco, only reason I went with the design is because I am running directly off mains and I wouldn't be able to scope / touch the output if I did not have some kind of insulating transformer. I might be a lot better of just using a 1:1 transformer for that and Richies series induction design. The only drawback I can see is that I need two big chunks of ferrite instead of just one, but paralleled flyback cores should hopefully do the trick.

Still I wonder if I could have best of both worlds by using an impedance transformer with a big air gap and weak coupling, or should I say a matching inductor with two windings? In fact I might try that tonight.

Marko

Wed Apr 25 2007, 12:27PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

You can use a little current transformer for isolation purposes.
When you get it working you probably won't need a scope at all.

You could even use a current transformer on output of the bridge to provide feedback and make the circuit self-resonant, wich would completely eliminate any need for tuning.
You wouldn't anymore be protected if your tank capacitor shorts out, but still it's worth.

Isolating transformer is just a big waste of ferrite while it does nothing except introducing additional losses.

Dr. Shark

Wed Apr 25 2007, 01:01PM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

I guess you are right with the current transformer (why did I not think of that?), but in case the heater ever gets water cooling isolation is a very good idea. If I ever get the heater to do more than get a piece of iron red hot, which surprisingly it does very well even when wildly out of tune, having copper pipe filled with water as the work coil will be inevitable. And since the water tap is grounded...

Steve Conner

Wed Apr 25 2007, 05:43PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

The capacitor on the primary side of my circuit is a DC blocking capacitor, it plays no part in the resonance.

As for why not to use a LCLR circuit, well I was messing around, I tried the series resonant circuit and it just worked great first time! So I roasted some bits of metal until I got bored with it. I stopped it from blowing up in the absence of a workpiece by running it off a current limited power supply.

Is Richie's design really immune to all those fault conditions?

Marko

Wed Apr 25 2007, 06:11PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Is Richie's design really immune to all those fault conditions?

A short or open in the tank circuit will cause the circuit to walk off from resonance, and current will mostly be limited by matching inductor.

Since impedance of paralell circuit is proportional to it's Q it won't draw significant current at resonance without load.

If we used direct feedback, the circuit wouldn't anymore have tank circuit fault immunities because it would then attempt to go in series resonance and blow up... but think this could be fixed by using PLL, if really important.

Marko

Tue May 01 2007, 11:17PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

I just ran again in this and tought how I never actually built an induction heater, and just came about playing with LCLR.

I rolled a coil out of some copper wire, soldered a dozen of CFL mica caps to it and constructed a paralell induction heater in about 15 minutes. I hoked the mess to my favourite halfbridge.

I didn't bother to turn on a scope, I just tuned it by hand and added capacitors until I got a glowing bolt.
Faulty OCD was limiting my power to some 150..200 watts, but I wasn't bothered to fix it.
Just to show how easy it is to get something as such working.

My camera seems to interpret orange light as violet:

1 2 3 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob