Maker Pro
Maker Pro

13.56Mhz Class E Power Amp for plasma generation

keathow87

Feb 6, 2011
16
Joined
Feb 6, 2011
Messages
16
hi, friends...I'm currently working on my Final Year Project to construct a plasma needle and facing problems in construct my RF supply. I'm using MOSFET IRF540N in my circuit as switching device and got some question here:

1) Does this IRF540N actually require Vgs of 15V to on/off it instead of the 2.5V shown in the datasheet?

2) The current instrument available for me only manage to provide a 13.56Mhz, Vpp of 2.5V square wave to my MOSFET, and it is a function generator. Is it possible for me to amplify the signal to Vpp of 15V?? what circuit needed to do it??

3) Any gate drive circuit available to provide a Vpp 15V, 13.56Mhz square wave if it is impossible to amplify the signal from function generator?

I'm really in need of your kind help and run out of time, so please help me:(
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Vgs(th) is between 2 and 4 volts.

And I'd be surprised if that particular mosfet would be the best choice of device to switch at that frequency. You'd be needing to supply a lot of gate current.

You may find that a mosfet gate driver goes some way to solving your problems (but I wouldn't guarantee it)
 

trobbins

Jun 15, 2010
83
Joined
Jun 15, 2010
Messages
83
Yes you will definitely need a dedicated specialist power mosfet gate driver IC. Carefully read and implement the circuit implementation and recommendations for the IC. I doubt you pedantically need to use 15V drive level to achieve your goal.

As Steve indicates, the parasitic capacitances of that particular device may swamp or grossly non-linearise the ideal level of Cds needed for a tuned amplifier - you may have to adjust the tuned circuit parameters quite a lot, and the waveforms may not look ideal.

The best 'tune' is effectively the efficency of the converter - I'm not sure what a plasma needle is, but it sounds like you may have a problem measuring the output power level.

Ciao, Tim
 

keathow87

Feb 6, 2011
16
Joined
Feb 6, 2011
Messages
16
http://images.elektroda.net/28_1297149430.jpg
this is the ideal output i want to get, the MOSFET available in this simulation is IRF530N. However, the input from XFG1 is only a 13.56Mhz, 2.5Vp square wave and it works perfectly well. Why it comes to a problem when test in real practice?


http://images.elektroda.net/42_1297149430.jpg
btw, i tried to amplify the voltage from the output of function generator and that's what i get. I was manage to amplify it from Vp of 2.5V to 10V now but it's still useless when come to connect the C2 to the gate of MOSFET. Can any1 tell me why it failed??
 

trobbins

Jun 15, 2010
83
Joined
Jun 15, 2010
Messages
83
Why it comes to a problem when test in real practice?
Ahh, the joys of giving someone a simulator. What have you tested in real practise?
What result did you get? can you show us?

Can any1 tell me why it failed??
RC time constants.
 
Last edited by a moderator:

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Well, I can see several reasons why it may have failed:

1) Excessive dissipation caused by an average gate voltage above Vgs(th) and insufficient gate current to effectively turn the device fully off leading to a continuous current drain from V+ through the 33uH inductor to ground, high dissipation and subsequent failure.

2) Excessive Vgs when the mosfet switches off leading to avalanche breakdown between drain and source, high dissipation, and subsequent failure.

3) High Rds(on) caused by low Vgs leading to high dissipation and subsequent failure.

4) Saturation of the core of the 33uH inductor causing a lower inductance to be exhibited. This would result in higher current through the mosfet, excessive dissipation and subsequent failure.

5) Any combination of the above.
 

keathow87

Feb 6, 2011
16
Joined
Feb 6, 2011
Messages
16
http://images.elektroda.net/78_1297177801.jpg
Trobbins, this is the experimental result. The waveform is the output at L2, which originally should be the same as what I got in simulation channel 1. However, this waveform only give me a 400mVpp :confused: btw, may I know what is RC time constant? sorry but my major is mechanical, this project was assigned by lecturer and not chosen by me...

Steve, can you please teach me how to identify which is actual problem? any solution to it? thanks for listing the reasons...
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Aah, failed as in doesn't work, not failed as in blown up.

The 33uH inductor is going to present a very high reactance at 13MHz and that will explain the low voltage appearing on the drain (I presume) of the mosfet.

Can you explain this additional capacitor, the value and how it is connected to the gate. The last thing you need is *more* gate capacitance.

Also, how much current can you get out of your signal generator. You will need a significant current capability to handle the gate capacitance. Have you measured the signal at the gate? Does your model assume a zero output impedance for the signal source (try putting a resistor in series with it (in the model) to reflect its output impedance
 

trobbins

Jun 15, 2010
83
Joined
Jun 15, 2010
Messages
83
Your class e circuit has L2-C3 across R1, and C1 is parallel with R1 - can you explain why you used this form of impedance loading - is it related to the plasma needle application - is the load voltage output meant to be the voltage across C3 (if so then are yo missing a further resistive loading element - can you confirm it is the correct circuit configuration?

Your drive voltage amplifier circuit is based on charging and discharge the gate capacitance of the FET - which ends up being the resistance from gate to the supply voltage for the HI state, and gate to 0V for the LO state - which effectively achieve waveforms with RC time constants. You really need a specialised IC driver, with specialised active devices to pull-up and pull-down the gate capacitance.

You can change the drive frequency slightly above/below the value you have simulated to determine which side of the correctly tuned state you have, and then use that information to modify component values (if you need to keep the final applied frequency at the set value you have simulated) - or use the frequency that gives you maximum efficiency as the best tuned circuit configuration for what component values you actually have.
 
Last edited:

keathow87

Feb 6, 2011
16
Joined
Feb 6, 2011
Messages
16
steve,
http://images.elektroda.net/51_1297255744.jpg
i have change the value of 33uH inductor but it doesn't affect the output much.

which capacitor? the one at output? what does it mean i need more "gate capacitance"? should I add a capacitance at the gate of MOSFET?

I'm using a Good Will Instrument, GFG-813 function generator to provide the square wave to the gate of MOSFET. Sorry but i don't really know how to measure the current of function generator, the lab technician even told me function generator can't output current but gain only :confused:

btw, these are the result when i add resistor to the gate...
http://images.elektroda.net/10_1297257102.jpg
http://images.elektroda.net/44_1297257102.jpg
http://images.elektroda.net/76_1297257102.jpg

tim=trobbins=schematic?? thanks for your reply in other forum as well =)

L2-C3 shouldn't be included in the simulation according to the paper, but here is what I get after remove it.
http://images.elektroda.net/38_1297257102.jpg

L1 and C1 is act as a resonant circuit to amplify the voltage signal instead of current signal according to the paper.

but speak truly, now what i hope is just to build a plasma generator that can supply my plasma needle. I will be very satisfied with it if I can see the purple glow on the needle neglecting its characteristic...have you guys got any experienced on building it?? my deadline is 8 weeks from now and I'm really stressed :( the ideal plasma needle is look like this http://iopscience.iop.org/1367-2630/6/1/149
 

keathow87

Feb 6, 2011
16
Joined
Feb 6, 2011
Messages
16
Alright, I'm not sure is it a good news or bad news. My instructor finally agreed to buy some component for my project, but the budget can not be too high. 2 things i need now are IXYS RF-DEIC420-IC,DRIVER,MOSFET DE275 and IXYS RF DE275-501N16A, if I can only choose one, which one is most needed in my project?

1) Is it possible to generate the plasma by using the DEIC420 to drive the IRF540N MOSFET? i heard that a RF MOSFET like DE275 series is more suitable in generating plasma but the problem is out of budget. The MOSFET is even expensive than the MOSFET driver...any suggestion please?? thank you...
 

trobbins

Jun 15, 2010
83
Joined
Jun 15, 2010
Messages
83
My initial feeling is that you need a standard class e output impedance circuit configuration - series LCR - to start with in simulation. Get that going to start with in simulation, and in a practical circuit, so that you have a fair understanding with which to move on.

You then need to read that article by Kieft et al very carefully, and determine what power level and operating parameters you want to achieve. Then determine the equivalent circuit (Rn, Cn, Rd) values you want to use. You may need to substantially increase the input DC voltage of the class E stage, which will affect your choice of FET.

The matching circuit they show is between a sine wave and your needle load. A class E won't be a perfect sine wave. You need to simulate each generator providing input to your needle circuit to understand the differences. Your needle circuit effectively replaces the R in the class E output circuit (it also modifies the final value of C).

At the moment you appear to have little understanding of what you are doing. I suggest you talk it over with your instructor and try to get a better understanding of the whole concept and what is happening in each section.

Ciao, Tim
 
Last edited:
Top