Designing an Auto Match Unit for RF supply / General Science and Electronics / Forums

Forums

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

« Previous topic | Next topic »

Designing an Auto Match Unit for RF supply

1 2 3 4 next last

Move Thread

LAN_403

Ash Small

Sun Nov 14 2010, 05:51PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

I'm new here but I'm hoping maybe someone who has built an induction heater can help me.

I've read up on induction heater circuits (including several here). In particular the PLL circuits, so I've done a fair bit of homework.

These circuits work by matching the supply frequency to the resonant frequency of the tank circuit (work coil/capacitor). I need to design a circuit which will match the resonant frequency of the tank circuit to the RF supply, by adjusting the capacitor as the load on the work coil varies.

Most of the induction heater circuits I've read about operate at around 100kHz, my RF generator runs at 13.56MHz but I assume I can use the same method of current sensing as in the 100kHz circuits, ie toroidal current sensing transformers.

The plan is to adjust a sliding vane capacitor using a stepper motor to match the resonant frequency of the tank circuit to the 13.56MHz RF generator.

Am I going about this in the right way? Any advice will be appreciated.

(BTW, The work coil forms part of an ICP ion source, but the theory is the same as for induction furnaces.)

Steve Conner

Sun Nov 14 2010, 06:57PM

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

Do you really just need to adjust the resonant frequency?

Matching units usually have two controls: tuning and loading. Broadly speaking, the tuning control adjusts the resonant frequency (including deliberately going off-resonance to compensate reactive loads) and the loading control adjusts the impedance transformation so the power amp is properly loaded.

But when the matching unit is designed for similar input and output impedances, say 50 ohm to 50 ohm nominal, it's no longer possible to say which control is the tuning one and which is the loading: both controls affect both variables. And both of them may need adjusted in response to a change of load.

I've built a couple of PLL induction heaters, and have seen the guts of an auto-tuning RF source, but I'm not sure how it works. Building a SWR bridge with voltage outputs would probably be a good start, then your microcontroller can look at the voltages with its A/D converters, and twiddle stepper motors for maximum forward power and minimum reflected power. The schematics I've seen for HF SWR bridges use a couple of toroidal CTs, so it's not that different to what you were going to do.

Proud Mary

Sun Nov 14 2010, 08:31PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

If you were to set up a variable LC matching network to discover the order of L and C values that worked, it would make it a lot easier to automate later.

The circuit below shows how a so-called pi-match L and C network couples a power oscillator to its load/antenna.

1289766536 543 FT0 One Valve Transmitter

radiotech

Sun Nov 14 2010, 11:47PM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

Is the supply seperated from the work coil? If it is, it
connected with section of transmission line. The standing wave
ratio on that line will be 1:1 when maximum power is transmitted
So a a staning wave bridge measuring circuit could serve as the
servo source signal to the tuning mechanism.

gadget like this?

Dave Marshall

Mon Nov 15 2010, 03:13PM

Registered Member #16 Joined: Thu Feb 02 2006, 02:22PM
Location: New Wilmington, PA
Posts: 554

I don't know a whole lot about induction heaters, but a suggestion. You might look in to how the Automatic Antenna Tuner units used by ham radio operators work. They do something very similar, but instead of varying the frequency generator, they vary the Fres of the tank.

There are several homebrew sites out there that describe how they do it.

-Dave

Ash Small

Tue Nov 16 2010, 01:20AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Thanks Dave, I'd googled SWR etc but googling 'Automatic Antenna Tuner units' has helped considerably. Thanks also to Radiotech and Proud Mary.

Steve, your reply was very informative. all the AMU's I've seen have two servo motors, now I know why. I'm still googling but I'm beginning to get my head round what is required.

I'm still puzzling over the exact layout, but I'll try and make a bit more progress before asking more questions.

I have ascertained that the transmission line needs to be 3.6 metres (1/4 wavelength x velocity factor for polyethylene dielectric), and that a parallel tank circuit would probably be better than a series tank circuit. I've also found an SWR circuit that looks suitable, (there are loads of different circuits out there)

I think I need to match the resonant frequency of the tank circuit to the generator frequency, then match the impedance to the transmission line, I'll post again once I've made a bit more progress and have some more questions.

Thanks again.

Ash Small

Tue Nov 16 2010, 02:27PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Having been googling til 4.30 this morning, reading up on all the different antenna matching circuits until my brain was scrambled, I slept on it and woke up with the following plan in my head.

I'm not sure it's correct, but it's a start.

One of the problems I'd come up against was that all these circuits are for antennas. I don't have an antenna in my circuit. Another was that all the commercial AMU's I've seen only have two servo motors driving variable capacitors. As far as I'm aware they don't have variable inductors. Also, any more than two variables would create a very complicated circuit (from the logic point of view). I need to adjust for tuning and load using two variable capacitors only.

I'm assuming that 'tuning' means tuning the tank circuit under load to 13.56 MHz, and that 'load' means matching the impedance to 50 Ohms, Please correct me if I'm mistaken.

I read that (I think) it is possible to tune a Pi match circuit, consisting of two variable capacitors and a variable inductor, by using the capacitors alone.

I also read that this works even if the antenna is shorted or open circuit. This is the circuit I have, ie no antenna.

I also read that the disadvantage of this is that you can burn out the inductor because all the power can be dissipated through the inductor. This, I think, is what I want, ie the inductor is the 'work coil'.

If I remove the antenna from a Pi matching circuit, I end up with what is, in effect, an inductor (work coil) with a parallel variable capacitor (ie parallel tank circuit) with a series variable capacitor between this and the transmission line.

I can adjust the parallel capacitor to give a resonant frequency of 13.56 MHz and (I think) adjust the series capacitor to match the impedance to 50 Ohms. (Please correct me if I'm wrong here).

This gives exactly the circuit I want, ie one adjustment for tuning and one for load, with no antenna and all power dissipated through the coil. (At least, I hope it does)

Most of the circuits I've seen online for induction heaters with a similar sized coil to the one I'm intending to use have an inductance of 1.5 to 2.5 micro Henries. Frequency of circuit is 13.56 MHz, RF generator is rated at 600 Watts. I'm off to do some maths, but any assistance would be greatly appreciated (I've never done a SPICE simulation before, so I've quite a bit to learn.)

Also, any comments regarding my reasoning above will also be greatly appreciated. (I've no idea whether the assumptions I've made are correct or not.)

Steve Conner

Tue Nov 16 2010, 02:45PM

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

Whether it's the radiation resistance of an antenna, or dissipation in a ball of plasma, the maths is the same. It's a resistive load.

It's not the coil that dissipates the power: it's the plasma. Coils are reactive, and (to a first approximation) can't dissipate anything. If there were no plasma, the matching unit wouldn't be able to find a tuning solution, or would arc over its capacitors, or some similar bad thing.

This of course raises the quesion of how the plasma gets ignited in the first place. Ralph Hartwell's Rife tube site makes some interesting reading in this respect.

The inductor in a pi-match only needs to be varied for big changes in operating frequency, like changing bands in a ham radio transmitter. I've built a few (manual) matching units that cover 1.8 to 30MHz, and they have a great big inductor with lots of taps.

I don't know whether what you did is valid. The plasma appears as a (lossy) shorted turn to the work coil, hence a resistor between the two ends of the coil. But the Pi-match is supposed to have its load connected across the output capacitor. It would probably still work, but it wouldn't be a Pi-match any more, it would be some other circuit with a different theory of operation. Like a L-match, T circuit or whatever.

Out of interest, is this a commercial, hobby, university research project or what?

Ash Small

Tue Nov 16 2010, 03:40PM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

Steve, Are you saying I need to connect the work coil accross the output of the second capacitor and ground? If so, I think this means I'll end up with two coils effectively in series with C2 connected accross them in parallel. This seems pointless. I'll post some circuit diagrams later to clarify the point I'm making if you like.

As far as igniting the plasma, provided you have sufficient power, it will ignite purely from the ICP source, no CCP igniter is necessary, especially if you adjust the pressure. As a last resort I'll use a CCP igniter circuit but I'm hoping this won't be necessary.

This project is a hobby. My interest in this stuff was re-kindled after working at Harwell/RAL on their particle accelerator. I've since had another job assembling plasma processing systems for semi-conductor manufacture, but this was primarily to further my reasearch into this technology. (most of the money I earned went on hotel bills while working away from home)

Like most of us, I do sometimes dream of making some money from my hobby, but this rarely happens.

I'm off work due to long term illness at the moment so I have some time on my hands to design/build an AMU for the HF generator that I aquired from one of the people I was working with recently.

In 2001 I built an HF TIG welder based on a hand drawn circuit that, I believe, was posted on the net by Arcstarter about ten years ago, (last time I was off work for some time) but this is my first electronic project since then.

Steve Conner

Tue Nov 16 2010, 03:55PM

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

Ash Small wrote ...

Steve, Are you saying I need to connect the work coil accross the output of the second capacitor and ground? If so, I think this means I'll end up with two coils effectively in series with C2 connected accross them in parallel. This seems pointless. I'll post some circuit diagrams later to clarify the point I'm making if you like.

No, I'm saying that if you do what you originally proposed (use the pi-match coil as the work coil) then it won't be a pi-match any more, because the load is connected in the wrong place.

It will probably still tune and match fine, but you won't necessarily be able to use the theory of the pi-match any more to predict what it will do and develop your autotune algorithms.

Out of interest, what are ICP and CCP? In my line of work ICP is In-Circuit Programming, and I've never heard of CCP. :)

1 2 3 4 next last

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