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

previous 1 2 3 4 next last

Move Thread

LAN_403

Ash Small

Tue Nov 16 2010, 05:05PM

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

I appreciate the point, Steve. I realise the maths will be different. I also appreciate the point that the inductor in the Pi match circuit is only varied to accomodate different wavelengths, ie frequencies. I do now have a starting point, though. the next job is to work out where to position the torroidal current transformers (I'm assuming at this stage that I'll need four of them, two for signals for the tuning circuit and two for signals for matching the load.)

ICP is Inductively Coupled Plasma and CCP is Capacitively Coupled Plasma.

ICP uses electromagnetic radiation to excite the plasma (same as an induction furnace heats metal), CCP uses electrodes, or 'charged plates' and electrostatics to excite the plasma. (basically an electric arc)

radiotech

Tue Nov 16 2010, 06:06PM

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

Ash Small wrote " Also, any more than two variables would create a very complicated circuit (from the logic point of view). "

With your "plasma load" is the input energy level constant, or
does it flip betwwen values as you alter fuel input rates?

If that is the case you may want to try and design a broader
tuning scheme, because your maximum power transfer depends
on conjugate matching. With a high Q, your servos could easily
overshoot and destroy the parts.
Google PID controllers, used everywhere in industry, which have
a lot of software outthere. You can get a cheap PID module for a small PLC.

Steve Conner

Tue Nov 16 2010, 06:37PM

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

Ash Small wrote ...

ICP is Inductively Coupled Plasma and CCP is Capacitively Coupled Plasma.

Well, here's a brain teaser for you. You said that ICP can ignite a plasma where there wasn't one previously. But un-ionized air doesn't conduct, so there is nothing to induce voltage in.

Unless:

Every ICP is also a CCP, because the plasma can "see" the high voltage on the unshielded turns of the work coil

or:

The distinction between ICP and CCP is meaningless, because a time-varying magnetic field is impossible without a time-varying electric one and vice versa.

(Those two statements might be equivalent, but that's a matter for someone with a firmer grasp of EM theory than me.)

In any case, I argue that you don't need four CTs. You need one SWR bridge positioned between the RF generator's output, and the matching unit's input. A SWR bridge consists of two toroidal CTs and one voltage sensor.

Or maybe you do need more than just that. The SWR bridge gives a scalar measurement, it tells you how bad the mismatch is, but not which direction you should tune in to correct it. Some sort of phase detector might be called for to tell if the load is capacitive or inductive.

By the way, this might be a long shot, but did you ever do any work in Kamelian's semiconductor fab in Yarnton?

iJim

Tue Nov 16 2010, 08:33PM

Registered Member #2662 Joined: Fri Jan 29 2010, 10:14AM
Location:
Posts: 36

Ash,

As far as I can tell from your posts you have a conventional ICP plasma source...

Have a look in the 'Inductive output networks' section of this page:

and find the most suitable one for your set up. In an automated system I built I used a directional coupler to determine forward and reflected power, a PIC to read these (a few Vdc) and vary the steppers until the Forward power is maximised.

Steve is correct, when there is no plasma there is nothing to induce voltage in. Most small (<500W) Argon ICP systems I've seen use some kind of ignition system... a few kV from an ignition coil is enough to strike an Ar CCP in the tube, this is then sustained with the RF and the ignition system can be switched off. I think in bigger system there's a chance of producing some ionization around the coil which is sufficient to get things going, although I never managed to get this to work.

Cheers,

iJim.

Ash Small

Tue Nov 16 2010, 08:53PM

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

Radiotech, I will google PID later. I'll comment afterwards.

Steve, with ICP the coil doesn't contact the plasma, it's usually outside a quartz glass or ceramic tube containing the gas to be ionized. With CCP the electrodes are inside and in contact with the gas (plasma). With CCP it's possible to contaminate the plasma with the metal electrodes (re arc welding).

ICP induces a field in the gas which tends to pull the electrons in one direction and the nuclei in the opposite direction (CCP also does this electrostaticly). CCP can be AC or DC, but ICP doesn't really work in DC mode.

Ion sources usually also have extraction electrodes, so are, in effect, a combination of both.

I realise I only need two current Xformers for the SWR/load matching circuit, but (Ithink) I need another pair for the tuning circuit. I'm working on some circuit diagrams at the moment that I'll post later.

I didn't work in semiconductor fab, I was building the machines/systems that the semiconductor/microchip industry use to produce silicon chips, HB LED's, etc.

(I can e-mail you the details if you like, but would rather not post online)

BTW, I was wrong earlier, connecting a work coil in place of the antenna in a Pi match circuit won't mean they are in series accross C2, as they are earthed between them, so they will actually be in series with C2 accross one end. I think this layout will incur more losses though, due to the extra inductor, but I'm not certain about this, just intuition.

klugesmith

Tue Nov 16 2010, 08:56PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

>> You said that ICP can ignite a plasma where there wasn't one previously. But un-ionized air doesn't conduct, so there is nothing to induce voltage in.

I respectfully disagree.
A varying magnetic flux induces a fixed voltage per turn, even in a vacuum!

As a specific example: consider pulsed ion thruster with a flat pancake coil.
Geometrically this is just like a ring or disk launcher,
but the "projectile" is a puff of gas (e.g. argon) in a layer close to the coil face.
When a capacitor is discharged into the coil, the tangential electric field ionizes the gas, which then conducts a current in that direction, and is accelerated in the axial direction.

Something I have long wanted to try with my can crusher is to replace the can with partially evacuated clear tube or jar, and look for a flash of light.

Ash Small

Tue Nov 16 2010, 09:59PM

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

Very interesting and relevant link, Jim.

The circuit I'd decided on is the 'inductive output T network' but I'll have to do the maths and see if the reactance of the coil is below 50 Ohms.

I've a hunch that at low pressures it may well be, though.

Steve Conner

Wed Nov 17 2010, 10:40AM

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

wrote ...

A varying magnetic flux induces a fixed voltage per turn, even in a vacuum!

But there are no turns of anything in a vacuum, this is my whole point.

wrote ...

When a capacitor is discharged into the coil, the tangential electric field ionizes the gas

What tangential electric field, where did it come from? If there is an electric field, the field lines must terminate on some conductors with suitable voltages that agree with the integral of the field. So, I argue that the electric field is caused by, indeed is, the voltage on the coil conductors, hence it is capacitive coupling.

I think it follows that an ICP work coil with lots of turns and a high voltage would ignite a plasma better than one with a single turn, even if they both had the same amp-turns and so generated the same H-field. With the high-voltage coil, you essentially get a free dose of CCP with your ICP.

I don't believe that "pure H-field" could ignite a plasma: for a start it doesn't exist. The minimum E-field you can have with your H corresponds to the voltage between the ends of a single-turn coil.

If you tried to make an experiment where you "shielded" the plasma from the E-field of a coil, to ignite it with H-field alone, you would short the H-field too unless you cut a slot in the shield, and it would then be a single-turn coil with a corresponding voltage across the slot.

And of course in the far field the E and H components come in the proportion dictated by the impedance of free space. When you ionize air with a laser pulse, it's this E component that pulls the atoms apart.

But these 13.56MHz plasma things are all near-field, the impedance can be anything you want.

Ash Small

Wed Nov 17 2010, 05:38PM

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

Steve, Here is a quote from the wikipedia page on ICP:

"When a time-varying electric current is passed through the coil, it creates a time varying magnetic field around it, which in turn induces azimuthal electric currents in the rarefied gas, leading to break down and formation of a plasma. Argon is one example of a commonly used rarefied gas."

The page is here:

BTW, It's the changing current that produces the electromagnetic field, not the voltage. The voltage does create the plasma in CCP plasma though, an example of this is the high voltage HF that initiates the arc in TIG welding.

Hope this helps to clear things up.

Steve Conner

Wed Nov 17 2010, 09:18PM

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

That's the thing, though. I'm familiar with the explanation on Wikipedia, but I don't believe it, for the reasons I stated. You can't induce currents in the gas before it has started to conduct, therefore induced currents can't be the cause of the breakdown. Only the electric component of the EM field can cause it.

Sure, Maxwell's equations require a time-varying magnetic field to produce an electric one, but the amount of it is something like dB/dt divided by the square of the speed of light. I think that must be a very small quantity even at 13.56MHz.

It has never made any sense to me, and I was hoping to spark some debate, as it were.

previous 1 2 3 4 next last

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