Royer circuit - blown MOSFET? / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Royer circuit - blown MOSFET?

previous 1 2 3 next

Move Thread

LAN_403

Electroholic

Thu Oct 19 2006, 05:59AM

Registered Member #191 Joined: Fri Feb 17 2006, 02:01AM
Location: Esbjerg Denmark
Posts: 720

maybe you have too much cooling, and not enough thermo power in the plasma, pump more amps to it by lowering the primary turns on the FBT.

Colin 99

Tue Nov 21 2006, 04:32AM

Registered Member #192 Joined: Fri Feb 17 2006, 03:08AM
Location: Canada
Posts: 44

Well I now have this circuit just SMOKIN'

I tried different caps for the resonant cap and eventually put a 2 uf metal film poly cap (this made a hugh difference). I also added caps across the input to the supply, I used a 0.68uf metal film poly cap and a BIG electroylic cap (this improved the output too).

To prevent the flyback from flashing over I jumpered a couple of the secondary windings to joint them together since one set of inner coils were floating. This is a BIG old flyback!

I can now run 36 volts into the royer circuit and the output is amazingly powerfull. It make a real nice jaccobs ladder. I can't measure the current, I tried with two DC current clamps and they give erronius readings on a DVM (I used the Min - Max feature). I also tried with another DVM that can measure 20 amps for a short bit - garbage!. It seams to be drawing more than 20 amps though.

The 14 gauge wires to the flyback primary get warm. The heat sinks get hot and so do the L1 air wound coil. I used 18 gauge for L1

Any suggestions on how I can measure the current realibly?

Sulaiman

Tue Nov 21 2006, 07:52AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

The simplest way to measure the dc current is with a moving coil meter
which displays AVERAGE current.
Whenever I use my dmm near arcs/sparks it gives erratic readings.
I suspect rf pickup.
(connect both dmm leads to 0v and check the reading)

When no arcs/sparks are present a 'Royer' should have a clean sinewave oscillation
F=1/2.pi.sqrt(Cp.Lp)

Flyback transformers have a naturally resonant secondary, somewhere around 50 kHz
operating at this frequency gives best performance.

At 35 volts dc input you should get about 75 Vrms across the primary inductance/capacitance
with 0.68uF at 27 kHz the capacitor current would be about 8.7 Arms, ok for a good capacitor.
0.68 uF at 27 kHz implies a primary inductance of 51 uH
I like to use at least 10x Lp for the dc inductor, so to me 128 uH is a little low.

I also use 30N60 with great success.
I do prefer driving the igbt gate resistors from a separate supply of 12 to 15 Vdc
(no zener diodes required)
IF you do this it's best to connect the gate drive voltage first,
then the power via the dc inductor.
You will notice that with just the gate drive the 'Royer' oscillates due to parasitics.

Colin 99

Thu Dec 07 2006, 04:32AM

Registered Member #192 Joined: Fri Feb 17 2006, 03:08AM
Location: Canada
Posts: 44

Well with some work I was able to measure the DC current to the Royer circuit, I'm drawing 13 amps at maximum arc length with 36 volts applied to it. I first tried to use an old simpson VOM that I bought at a ham radio flea market. The thing was way out of calibration, but after recalibrating it, I pin the needle once I draw out an arc on the 10 amps range. Then I remembered seeing a old phillips current shunt in a cupboard at work, so I checked it and it works fine. I'm still using a DMM but I have it located a good distance from the circuit and I check it for noise pickup - no problem.

I now need a bigger inductor, larger diameter wire and more number of turns. 18 gauge magnet wire heats up pretty quick at these currents. Also I need bigger heatsinks as they heat up almost as much as the L1 inductor. Or maybe a fan would do the trick?

---edit---
I found another trick to get more power out of this circuit. I put a heavy duty capacitor across where L1 connects to the flyback primary center tap and ground (battery neg). The arcs seem hotter. I just used a 0.68uf pulse cap.

Does anyone have suggestions to make the royer better?

thanks for all the replies.

Shaun

Dr. Dark Current

Thu Dec 07 2006, 03:27PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Colin 99 wrote ...

I found another trick to get more power out of this circuit. I put a heavy duty capacitor across where L1 connects to the flyback center winding and neg. The arcs seem hotter. I just used a 0.68uf pulse cap.

I think that doing this bypasses the inductor, so I don't think it's a good idea

Sulaiman

Thu Dec 07 2006, 03:59PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Depends what result you want;

For a good sinewave at constant output voltage the primary 'Q' must be AT LEAST 2.4
('Q' here is VAR in the resonant capacitor divided by output power)
and a fairly large series (constant current) inductor.
This will ensure zero-voltage-switching.

For maximum power output squarewave 'hard-switching' will give the best performance.
(NO SERIES INDUCTOR OR RESONANT CAPACITOR)

Using a series inductor with some capacitance from the primary midpoint to 0v
gives more power output than a true cfpr ('Royer') at the cost of higher switching losses,
the series inductor will then provide soft startup and allow for some capacitive output load.

Overall, for a given quantity of ferrite and wire I like;
cfpr ('Royer') invertor for constant voltage sinewave output
Flyback for constant power/capacitor charging/eht output
hard-switched push-pull for maximum power at constant output voltage.

Colin 99

Fri Dec 08 2006, 02:51AM

Registered Member #192 Joined: Fri Feb 17 2006, 03:08AM
Location: Canada
Posts: 44

Quote: Using a series inductor with some capacitance from the primary midpoint to 0v
gives more power output than a true cfpr ('Royer') at the cost of higher switching losses,
the series inductor will then provide soft startup and allow for some capacitive output load.

That would explain why my heat sinks were getting so much hotter after I added that cap across it.

Now that I know, I think I'll remove it. - thanks

Could you further explain the Q factor? I guess I need to know how to measure it. For the record I'm using the flyback the way it came, I didn't change the gap. I'm using 14 gauge wire for the 10 turn primary and I'm using a pulse rated (metal foil) resonant capacitor.

It looks like if I increase the series inductance, I'll get closer to zvs and less heat on the heat sinks.

shaun

Sulaiman

Fri Dec 08 2006, 08:59AM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

First, as Firkragg mentioned above;
I doubt that your flyback transformer can handle the power that you're trying to push through it.
I suggest you change to true flyback operation.
A 555 directly driving a 30N60 would work very well (A combination that I use)
This would give a MUCH higher output voltage to operate your Jacobs Ladder
and run for much longer on your batteries.
One 12V battery should be good for powering the 555 and the flyback.

To answer your question, the 'Q' of your 'Royer' is not possible to calculate
because you are not running at constant power output with a Jacobs Ladder,
and I'm pretty sure you're saturating the core anyway.
IF you were operating in zvs mode;
The primary VAR with 35 Vdc supply would be 75 Vrms x 8.7 Arms = 652.5 VAR
(for 0.68 uF at 27 kHz)
so the MAXIMUM output power should be 652.5 / 2.4 = 272 Watts
(sorry, I can't remember the reference for 'Q' >= 2.4 for zvs)

EDIT:
If you do convert to true flyback operation the resulting eht power supply would be good for
- Jacobs Ladder (with or without a diode)
- Small sparkgap Tesla Coil (with a diode)
- Plasma Globe (without a diode)
- Capacitor charging (Pulse discharge etc.) (with a diode)
- Lifter (with diode)
etc. etc.
A 'Royer' is ONLY really suitable for constant voltage output)

Marko

Fri Dec 08 2006, 01:06PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

Flyback operation actually wouldn't do that much difference, since current is also what helps the arc to remain hot and stretch many times it's strike distance with jacob's ladder.

Many people use this circuit and get some scary JL outputs with it.

It's simply the amount of power you get into the flyback (more power = bigger spark), and flyback has it's limits. If voltage gets too high, it breaks down. If it is overcurrented, it dies again.

Mazzili circuit is generally a great way to do drive anything that doesn't require regulation or high input voltages, and is incredibly simple.

If you are saturating the core you'l need to use a smaller tank cap.

colin: what sparks do you get at full power input?

Some losses are simply unavoidable, and when you consider pushing 20amps through IRFP460's (270miliohms) you automatically get 5,8 volts of drop (DC!) and 108 watts of dissipation

IGBT's are also going to drop a few volts, and may be somewhat more efficient at that currens; but you will still be burning lots of power on them no matter what you do.

Flyback converter may require even higher voltage device and would burn a heck more power on switching losses.

If I had to operate this curcuit at such monstrous powers, I would seriously consider to do it directly from mains (110V is ideal to use with 600V IGBT's).

You could get much more power while keping IGBT's happy.

YOur flyback, whatever it is, must be a veteran now

Colin 99

Sat Dec 09 2006, 03:21AM

Registered Member #192 Joined: Fri Feb 17 2006, 03:08AM
Location: Canada
Posts: 44

Thank you all for your replies. This type of circuit is new to me and so are igbts.

I scoped the royer circuit at the resonant capacitor (to battery neg) that I'm using which is 2uF. With 36 volts input: I get 84 volts peak at each end and it's a nice clean, smooth half sine wave. When I pull a long arc the wave shape changes to a slightly distorted, lower amplitude half triangle wave (in between a sine wave and triangle wave). I also tried at 24v input and I get the same shapes. If I was saturating I would expect flatenning or a high peak at the top of the sine wave which I'm not getting.

My big flyback has some corona problems, it hisses underneath where the inside winding come out but the flyback still lives on another day

The jacobs ladder has a small gap of 1/2 inch and the top spacing is 4 inches. When it's running the arc often goes about an inch above the top of the ladders.

I'm not using irfp460 any more, I upgraded to 30N60s.

I'm going to put that bypass capacitor back on and scope it to see if that changes the waveform.

later,

shaun

previous 1 2 3 next

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