RF power supply and impedance matching

[steve_rb]

I need help with a RF power supply design and it's impedance matching to a small cupper loop antena for creating a small plasma reagion. I will use H2 gas in a vacuum regoin. Any one has any experience with design or impedance matching to help?

 

[vfone]

Actually I think you are looking to match an RF Signal Source connected to a small loop antenna.
First of all you have to know (or to measure) the output impedance of the signal source and the input impedance of the loop antenna, and after that design the matching network.

The net is full of papers about impedance matching including some free calculators.

https://bwrc.eecs.berkeley.edu/Research/RF/projects/60GHz/matching/ImpMatch.html
https://www.advanced-energy.com/upload/File/White_Papers/ENG-WHITE18-270-02.pdf

 

[FvM]

I here from your question, that you don't have a RF power generator yet.
But you didn't even mention the intended operation frequency and RF power level.

 

[steve_rb]

No I don't have the generator yet. I am looking to purchase or to build one if I find schematic. Frequency is 13.56 and power about 2-3 KW.

 

[FvM]

A 13.56 MHz loop should be considered as coil rather than an antenna. Because it's operated far below wavelength, it's pure inductive with a very small real impedance. So the main action is to bring it into resonance with a parallel capacitor and then try to match the real impedance with the generator.

I doubt however, if a "small copper loop" can handle 2-3 kW. Copper tube coils, possibly water cooled, are used in ICP (inductively coupled plasma) generators and induction heating.

 

[steve_rb]

A 13.56 MHz loop should be considered as coil rather than an antenna. Because it's operated far below wavelength, it's pure inductive with a very small real impedance. So the main action is to bring it into resonance with a parallel capacitor and then try to match the real impedance with the generator.

I doubt however, if a "small copper loop" can handle 2-3 kW. Copper tube coils, possibly water cooled, are used in ICP (inductively coupled plasma) generators and induction heating.

1. I almost have no info about Rf networks. You have said 13.56 MHZ operate far bellow wavelenght. Could you please explain this a bit more?

2. In one case I have seen someone has used a ferrite core transformer primary is connected to 13.56 MHZ Rf amplifier via a 50 ohm coaxial cable and secondary connected in series to a variable capacitor+a choke+ 13.56 MHZ antenna (coil as you said). Secondar has no or a few turn winding as I can see from schematic. What do you think about this method?

I am uploading it' schematic:

 

[FvM]

13.56 MHz free space wavelength is 22 m. I guess, your coil dimension is about 0.1 - 0.5% of it. That means, it has effectively no antenna properties and can be described as an inductor without considering EM laws.

Your principle circuit can work, in general, as it doesn't have any quantitative parameters (winding ratio, L and C values).

I almost have no info about Rf networks.

I expect serious difficulties for the instrument design then.

 

[steve_rb]

I expect serious difficulties for the instrument design then.

I am relying on:

1-Help from people on this board
2-Internet informative power
3-Expert softwares
4-My own studies and talent
5-Luck

By the way as I can see and I understand in the circuit above it seems transformer is used for matching real Impedance between a high and very low impedance. Output impedance of Rf generator must be high and since input impedance of the antena part is very low the person decided to use transformer to match these real impedances. Valiable and choke are used for matching Imaginary part . Am I right?

 

[FvM]

The description is basically right, the problem is to know the real impedance.

 

[steve_rb]

As I undestant I must take following steps:

1- Mesearment of output real impedance of purchased RF generator (simply just use ohmeter to measure this when generator is off)

2- build or purchase a ferite core transformer with almost same real impedance on primary winding (check simply with ohmeter leads conneted to primary). secondary impedance will be obtained later

3- For finding exact real impedance of 2 I must add series resistance of 50 ohm coaxial cable as well

Matching of real impedance on primary side is finished. Now for Secondary side:

1- build antenna (coil) acording to watage required for induction power needed for ionizing gas around antenna (coil). Here I will need help on how to calculate this

2- measure real impedance of the coil simply using ohmeter

3- calculate current needed inside the coil according to calculated watage on 1 and also calculate number of windings on transformer secondary side acording to the calculated current.

4. measure output real impedance of secondary side of purchased or home made transformer simply using ohmeter

5. Now we have real impedances of antenna(coil) and transformer secondary side. For mathing these two impedances we can use a choke with proper real impedance. Calculate choke real impedance and build or purchase proper choke.

6. No measure imaginary impedances of coil, choke and secondary of transformer (which will be inductive type) and add a proper variable capacitance with proper capacitive type impedance to match this inductive impedance to have resonace conditions. this part looks to me will be lot harder and I will need a lot of help here

Above I have neglected primary side imaginary impedance matching. I assume this part will be lot easier that calculating second part imaginary impedance. Even I may say it may not be required to do this if I accept some power loss in primary side.

Please comment on the above steps and rewrite steps if there are many errors or just post corrections if main outlines of the above steps are OK.

 

[FvM]

I fear, your suggested steps don't fulfill their purpose, for various reasons. In my opinion, they sound like a more detailed version of your previous statement "I almost have no info about Rf networks". It's obvious, that you have neither the general electrical/RF engineering nor the application specific ICP background.

To mention a few points:

1. A generator has a specified output impedance, often 50 ohm. It's achieved at the generators operation frequency and can't be measured with a ohmmeter (although some wideband test generators possibly expose 50 ohm impedance downto DC)

2. A transformer can be specified by it's windings ratio. It transforms impedances according to the square of this ratio. If it has a real impedance on it's own, it's made up by winding resistances and core losses, both are unwanted properties of real transformers. Particularly for a kW application, the winding resistances (at the operation frequency, also considering skin and proximity effect) must be really low.

3. The most serious problem is with the coil or "antenna" impedance. As mentioned before, the coil is basically an inductance. Very similar to the transformer windings, it should have ideally zero DC resistance and also not produce RF losses on it's own. Ideally, the plasma should be the only real impedance loading the coil. That also means, you can't measure the coil's real impedance part without igniting a plasma.

An ICP expert can most likely give quantitative estimations when knowing your design.

 

[steve_rb]

yes I agree. I have no information about this work I am trying to do but have no choice.

Now I can see some light is shinning on the dark path I have to travel through.

Could you please list steps ,as I have listed, I must take with a small explanation for each step for the circuit above. For each step I will do neccessary studies until get it to work.

 

[FvM]

Before thinking about impedance matching, you have to know about the coil and the plasma zone geometry. This point is ruled by application specific parameters. I noticed, that there's a large amount on plasma engineering literature discussing various plasma sources, including ICP variants.

 

[steve_rb]

coil is made from 2mm copper inside a 4mm dia quartz tube (for water running inside for cooling) as show on the picture below. Coil diameter is 2.5 cm and lenght is 2.8 cm. Plasma chamber is 5 cm diameter and 8 cm long cylidrical shape. I measured coil with ohmeter and it shows about 0.05 ohms but is not very accurate. 5 KW rf generator already ordered for this purpose. Please advice how to proceed and what is the next step.

 

[FvM]

The better configuration would be most likely a copper tube, with cooling water inside. E.g. 1/8" - 1/4" O.D., as used for laboratory gas supply. It needs an isolated feedthrough, of course.

As previously stated, the coil is pure inductive without a plasma load (except for the skin effect losses). I expect an inductance around 200 - 300 nH. The interesting question is, if you can ignite a plasma by applying a field without an auxilary gas discharge., e.g. a spark gap.

 

[steve_rb]

Do I need to measure inductance both with and without load? If so I need to know:

1. procedure for measuring inductance without load?
2. Procedure for igniting plasma and measuring inductance with load or somehow calculate inductance without igniting?

Added after 3 hours 44 minutes:

biff44 has mentioned a procedure for measuring impedance ( here : ) .Would you confirm the procedure? If so still it is a bit high level for me and I need a bit more details about how to connect directional couplers (a hand drawn schematic perhaps will be very good).

As I understand I should connect RF generator to the coil via a let say one meter long 50 ohm coaxial cable (as mentioned controlled impedance transmission line) and two directional couplers and adjust generator power for a low level (0.5 kw for example). Then using two directional couplers while generator is on I should meaure toward and away levels of power. From this I can measure VSWR and from VSWR I can calculate impedance using Γ=(VSWR-1)/(VSWR+1)
RL=-20*LOG(Γ)
formulas but this will give real part of the impedance. Will this be enough for me to finish the design and building Ion Source? (at least to finish the design?)

 

[FvM]

Taking some elementary plasma physics knowledge as granted, ICP works like an air transformer with shorted respectively low impedance loaded winding. So in the equivalent circuit, the plasma load appears as a loss resistance in parallel to the coil inductance.

I know, that the plama properties and also the expectable load impedance can be estimated with numerical EM tools, but it's not my business.

It may be the case, that you can ignite a plasma at low pressures without an "igniter", because a glow discharge starts at rather low field strengths. Ask the literature what others are doing.

 

[chau_s79]

Hi, i have a load that is likely to be a low resistance and capacitive (in the region of 2-j77). Not found a way to measure it though. Need to match it an rf source of 50 ohm. The L configuration of a series inductor and a shunt variable capacitor does the job theoretically. But the problem is that it requires variable capacitor of the order of 1000 pf. I however do not have more than 125 pf with me. Any suggestions?
I am attaching the paper from where i have sourced the referenced L configuration.

 

[dee25pak]

dear steve_rb did you got the answer you were looking for. I am also having the similar kind of application. If possible kindly explain the same.

 

[steve_rb]

I am working on it. I have made a coil and tried to turn on a quartz torch. I coud turn on the plasma using spark gap but plasma was weak and only close to outer quartz tube. I realised peak to peak RF is about 2000V which means current inside the coil is not more than 1 amps. This create a weak mafnetic field and expaines plasma weakness. I had no returned power and used about 2000 W for thsi purpose. I have posted photo here:


Now I am trying to increase coil current = decrease RF peak to peak voltage on the coil. One thing I am working on it for this is to change quality of contatcs I have and see if there is any change. With 13.56 MHZ I am using may be I have too much impedance preventing current from flowing inside the coil but I am not sure about this.