Class E SSTC thread / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Class E SSTC thread

first previous 4 5 6 7 next last

Move Thread

LAN_403

Marko

Mon Jul 10 2006, 12:08PM

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

I actually don't think that some devil gatedrive is really important here.
Steve Ward ran his coil with 10 ohm gate esistor and everything seemed fine for him.
Your driver BP, could now drive a fullbridge of IGBT's without much problems

At least cou can't blame your gate driver for problems anymore :p

Thing that happens to me is pretty wierd and happens certainly as result of something_getting_hot, as I can replug the coil and still have about 400V DS voltage.
If I let the mosfet and cap cool down I can observe the effect again.

Voltage needs just few seconds to rise up to 400V, after some time it can get to 450+ and eventually cause death of the mosfet.

Small video of what actually goes on:

(4.2mb)

Your heatsink is a bit bigger than mine but air-contact surface is about the same.
And my IRFP450 does also get very hot after some 30 sec's of continuous torture.

I don't know if this is too good but anyway in best case he has to dissipate some 10 watts

oh, and 6cm Somewhere I lost a sense of scale in your pics, I was thinking more like 2cm You are certinly beating my vttc efficiency (as close as I can come you your solid state leetness) of 12cm at 1.8kw

My current maxsparklength is 10-11cm for this coil, with something about 100W power input.
Anyway I can't get very stable arc since breakout continues on it's own from the topload.
I guess I would need an interrupter to mantain this and prevent streamers from sticking on the same spot and burning the topload.

You can see the sparks somewhere in the middle of the first page. Topload-streamers are also something about 6cm max long...

Steve Conner

Mon Jul 10 2006, 12:31PM

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

Some ceramic caps are designed to work in resonant circuits and take high RF voltages, and others are only really meant for bypassing and DC blocking. These latter types show drastic changes of capacitance with temperature (think a factor of 2). I guess this is what's happening with Firkragg's setup.

Marko

Mon Jul 10 2006, 02:12PM

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

update1*

I found an adequate MKP 7,5nF, 1500V cap.
'rising' effect has gone and coil works like nuts now.
I get some 160-170V max no matter what I do.
Even 'fat looking' ceramic wasn't good for the job, but polypropylene/polycarbonate caps work pretty fine.

*me gives a bunch of geek points to Steve Conner again.

I'm not sure if capacitance is too low but it works very good with breakout point.
without breakout I get a massive overdamped waveform (a bit counter-intuitive).

update2*

This overdamping when coil is unloaded gets really annoying.
I want it to be somewhat stable with and without spark, and I can't find any eason why waveform looks like this then. It can't be just spark detuning since wavefoorm repairs to sine-ish again.
Interesting is that if there is a breakout I can keep high coupling and impedance match without making trouble.

I actually have no adequate theory for this yet.

I can guess this is just a trait of class E, that you can't give it enough coupling for nice sparks without punishing the driver.

ragnar

Wed Jul 12 2006, 01:00AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

The 'devil' gate drive really is necessary -- I find the IRFP460 4.2nF gate capacitance very difficult to swing, even as a resonant circuit :P.

Here he is running at 3.55MHz, about 180W (50V in) with 5cm max breakout. The discharge is almost inaudible.

Very stable, but still not efficient enough for myyy liking.

...

Wed Jul 12 2006, 04:30AM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

wow bp, looks great! I think you need a longer topload though

Sulaiman

Wed Jul 12 2006, 06:18AM

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

Blackplasma,
is that a sheet of copper-clad board under the coil?
If it is, then I think replacing it with a non-conducting board would increase performance.

WaveRider

Wed Jul 12 2006, 10:52AM

Registered Member #29 Joined: Fri Feb 03 2006, 09:00AM
Location: Hasselt, Belgium
Posts: 500

Firkragg, if you have a good match when there is breakout, it is unlikely that there will be a match when there is no arc. With no breakout, your inductively coupled resonator is unloaded and will appear inductive (below resonance). Given the high-Q of the resonator, this corresponds to wide shifts in input impedance over small frequency excursions. Before breakout, you need to excite the coil with a frequency that is slightly higher than when it's running with an arc (plus...at startup, your driver must tolerate high voltages because of power reflected back from the unloaded resonator).. I believe the PLL-based systems try to track these frequency shifts on start-up and when the arc-loading conditions change.

If it's any help, my experiments indicate that the resonant frequency will shift down by 20-30kHz (for a 4-5MHz resonator) with approx 3cm breakout. Theoretically, this is mainly due to extra capacitive loading at the top of your resonator as a result of the now conductive "plasma globe" at the top of your coil.

Great work Firkragg/BP!

Sulaiman...the ground plane is needed to force the current peak in the resonator to sit near the base of the coil...instead of the middle.....assuming that the coil winding is connected to the copper base..

ragnar

Wed Jul 12 2006, 10:20PM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Yes, copper is connected to the secondary base. :)

This pic shows the sparklength in nice-ish proportion to the coil... CW so no long streamers of course

Marko

Wed Jul 12 2006, 10:50PM

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

Firkragg, if you have a good match when there is breakout, it is unlikely that there will be a match when there is no arc. With no breakout, your inductively coupled resonator is unloaded and will appear inductive (below resonance). Given the high-Q of the resonator, this corresponds to wide shifts in input impedance over small frequency excursions. Before breakout, you need to excite the coil with a frequency that is slightly higher than when it's running with an arc (plus...at startup, your driver must tolerate high voltages because of power reflected back from the unloaded resonator).. I believe the PLL-based systems try to track these frequency shifts on start-up and when the arc-loading conditions change.

I didn't actually do much work here last 2 days, as I was mostly cleaning my room

Also dug up some old interesting stuff.
I can't really have some serious frequency-tracking problems since PLL seems to work quite good.

class E amp actually doesn't like good coupling and impedance match, and what seems to be happening that circuit matches into impedance of unloaded coil better than it's impedance is 'increased' by breakout, considering that spark looks like a parallel resistor to resonator.

I'm not yet sure how can I calculate impedance for class E (wierd voltages/currents) wich also uses primary as inductor, but I'l rather go into 'experimental science than post my ramblings over here.

Reducing coupling would for example help but then I get small sparks. :p

I think I should reduce number of primary turns a bit, maybe some more leakage inductance and etc.
Too much posting crap and not enough work anyway.

Blackplasma: spark looks nice but still somewhat smaler than your first pic. (?)
Maybe you really should move the primary cm or two up from the ground plane, looks like you are making induction heater over there..

ragnar

Thu Jul 13 2006, 11:39AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Raising the primary increases the coupling a wee bit much making it hard to hold class-E critical damping on the shunt capacitor, but if he were also increased in diameter, then I could keep the coupling constant and move it away from the secondary... That'd also help with any flashover issues.

As for the pics (breakout into one plume, wire brush breakout, me at work, breakout into three plumes) ... meh, I couldn't help myself

Sorry 'bout the filesizes, guys =)
[edit]so I don't get into trouble, I've made the GIFs into links. Please, click them! :P[/edit]

animated GIF of breakout into a single plume

animated GIF of breakout into three plumes

first previous 4 5 6 7 next last

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