"Electric flame" HF VTTC / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

"Electric flame" HF VTTC

previous 1 2 3

Move Thread

LAN_403

Arcstarter

Tue Sept 08 2009, 09:15PM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

I would say beyond a reasonable doubt tubes are easier to make a circuit for. You don't even need any kind of actives, and only 3 passives (excluding the filament bypass cap) which is resonator, RF inductor on the input, and the feedback, which all are very few turns. Getting it to work took no time at all. Though, i started with 5 feedback turns, and ended up with one. Also, if your mosfets get hot you can 'sink 'em. When using tubes you do not have that luxury, the plate is all that is there to dissipate the heat. Unless you have the radiator on the plate, which *sticks thumb up*.

I did use more passives, though. I used a RF filter on the plate, and i have used a pair of 2.7nf 2kvdc capacitors from the plate of the tube to sort of help protect from having a MOT shorted through you, but as long as the oscillations continue, this offers no real protection. I had the cap sitting right in front of me, and some schematics use it, so i thought why not.

Running mosfets at 10Mhz? Really? Wow. You make it sound like it is no big deal. I always just figured mosfets, such as the IRFP series, would not run 10Mhz without a destroyed gate or something like that.

Also, if you are familiar with the IRFP350 (no suffix), would you happen to know if it would run around 4-6Mhz?

How would you drive the gate of MOSFETs at this kind of frequency? Forget about galvanic isolation right now. Tiny totem pole pair of BJTs driving a bigger totem pole or a mosfet half-bridge or something?

GeordieBoy

Tue Sept 08 2009, 09:45PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

> Running mosfets at 10Mhz? Really? Wow. You make it sound like it is no big deal.

Many modern switchmode power MOSFETs can switch efficiently up to around 15MHz or so using single-ended topologies like the Class-E amplifier. Small-signal MOSFETs running well into the GHz region are also common in RF applications and modern microprocessor technology.

> I always just figured mosfets, such as the IRFP series, would not run 10Mhz without a destroyed gate or something like that.

Small die devices like IRF510 can run efficiently up to around 40MHz, but Fmax for efficient operating decreases with increasing die size. As you suggested, driving the gate effectively is the first challenge. The second is minimising switching losses by using drain-current shaping or drain-voltage shaping to minimise V/I overlap at the switching instants.

> Also, if you are familiar with the IRFP350 (no suffix), would you happen to know if it would run around 4-6Mhz?

Should be good for about 400W of power at those frequencies in Class-E. Take a look at the Class-E papers by Sokal, and the work by Rutledge and Caltech.

> How would you drive the gate of MOSFETs at this kind of frequency?

At frequencies in the range 1 - 8MHz I would use a Mn/Zn ferrite transformer to provide impedance matching into the gate, and cancellation of the MOSFET gate's capacitive reactance. At the high end of that frequency range, the leakage inductance of the GDT can cancel the capacitive reactance of the gate. At the low end additional variable inductance may be required. At the high end Mn/Zn starts to get quite lossy.

From 6MHz to about 16MHz I would us Ni/Zn ferrite instead for the GDT, possibly with additional capacitance in series if the leakage inductance of the transformer is too high.

Above 15MHz I would use a reverse L-match network to match directly to the gate of the MOSFET without galvanic isolation.

Regardless of the arrangement chosen the gate-drive circuit basically encompasses the MOSFET gate in a resonant circuit that allows it to be efficiently driven with a nice big beefy sinewave at the intended switching frequency.

Some info about my 4MHz Class-E HFSSTC is here:

and a similar device running at 8MHz is shown here:

-Richie,

Arcstarter

Wed Sept 09 2009, 12:46AM

Registered Member #1225 Joined: Sat Jan 12 2008, 01:24AM
Location: Beaumont, Texas, USA
Posts: 2253

Thanks alot, Richie! Very informative.

I am now confident that a HF sstc is within reach. Hope to only go through a few mosfets :P sarcasm(oh yea, that is rather wishful thinking, ain't it?)/sarcasm

I have read up on your 4Mhz sstc, which is where i originally got the idea. I plan to use a crystal, because of its stability, and simplicity. The more difficult it is, the harder it is to fix. Also, a CD4046 will only do 2.7Mhz, which would still be interesting, but 4Mhz is what i decided on. I need to read up on your page again, though.

Time will tell...

GeordieBoy

Wed Sept 09 2009, 12:20PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

I'd definitely recommend using a ready-made crystal oscillator module. Then you know that the frequency is fixed and the timing is rock-solid and stable. You design everything to work at this spot frequency, and tune the TC resonator to match.

4MHz isn't too difficult. The 8MHz HFSSTC was quite finicky for tuning, and the 18MHz one kind of worked once I think!

previous 1 2 3

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