PLL CW SSTC Gate Drive / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

PLL CW SSTC Gate Drive

1 2 next

Move Thread

LAN_403

MRacerxdl

Sun Aug 28 2011, 11:40PM

Registered Member #989 Joined: Sat Sept 08 2007, 02:15AM
Location: SÃ£o Paulo, Brazil
Posts: 476

Hi all, I'm building a PLL SSTC to run in CW Mode with FM Audio Modulation.

I already built a board to test it and its working very fine.

My secondary coil runs about 400kHz with an IRFP260N Half-bridge. It draws about 10A from 127VAC mains (about 158VDC on the bus) and the irfp260N can handle easy.

The only BIG problem is the heating at the driver. I got the two TC4429 too hot to put my finger on it on only 5s running. I know that CW stress too much the driver.

I dont think two TC4429 in parallel will solve it (it will reduce the heat, but driving two IRFP260N at 400kHz isnt easy) and also I cant increase the gate resistor because my switch rise/fall times are already bit high (about 200ns for each).

I tried to find IXDD414 here, and the UCC drivers, but I didnt found both. I only see the Microchip TC's here :/

Is there any other way to drive? I was thinking about a IRF540/9540 fullbridge to driver the gate.

Thanks!

Mr.Black

Mon Aug 29 2011, 03:39AM

Registered Member #3960 Joined: Mon Jun 20 2011, 06:27AM
Location: Wuhan Hubei China
Posts: 33

I think you can try add a transistor totem pole
in the TC4429â€˜s output
this method is usually very effective
and the transistor best choose is TIP41/42

MRacerxdl

Mon Aug 29 2011, 04:01PM

Registered Member #989 Joined: Sat Sept 08 2007, 02:15AM
Location: SÃ£o Paulo, Brazil
Posts: 476

But the TIP41/42 is too slow (us switching time), I'm running at 400kHz

...

Tue Aug 30 2011, 12:43AM

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

Your coil sounds a lot like mine, I subverted the problem by switching to the TO-220 type tc4421/2 devices on a small fan cooled heatsink (I believe it was from a computer motherboard, cooling one of the support chips)

MRacerxdl

Tue Aug 30 2011, 03:53AM

Registered Member #989 Joined: Sat Sept 08 2007, 02:15AM
Location: SÃ£o Paulo, Brazil
Posts: 476

Hmm, these 4421/2 worked good switching that? its really almost yours coil indeed, I also used the Steve Conner PLL driver with a few mods.

...

Tue Aug 30 2011, 11:05PM

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

The switching waveforms weren't the greatest, but they did the job. I had the fets mounted on a huge heat sink from the initial high power testing with the ixfx55n50's but I have run the coil for hours at a time and abused it with all sorts of mistuning, primary hits, melting secondary windings, wonky grounds, etc, and have yet to manage to kill a mosfet after I replaced the IXYS parts due to the heatsink arcover.

One thing that I did end up doing was ditching the PLL and using the 4046 as a simple VCO. I left the current feedback transformer wired in if I ever decide to switch it back to closed loop mode, but I don't have the secondary ground wire run through it anymore.

I just have a knob on the back that lets me adjust the free running frequency of the coil, which works as a crude volume control of sorts, by detuning on the low side the sparks get smaller, running slightly detuned (pulling maybe 5A instead of the max 8A) gives the best audio quality, running right at resonance gives the largest/hottest streamers, and just on the high side causes it to work like a DRSSTC would, the audio quality is terrible but the streamers get longer/shorter with the peaks of the music (this mode works exceptionally well with midi songs that were synthesized with square waves, which acts like an interrupter). The bridge was designed to run at much higher currents, and the mosfets I used had good reverse diodes, so all seems to be well despite the extra stress that the bridge sees running way out of tune.

Dr. Dark Current

Wed Aug 31 2011, 08:29AM

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

The parallel drivers solve the problem, as well as increasing gate resistance moves the heat from drivers to resistors.

MRacerxdl

Wed Aug 31 2011, 07:10PM

Registered Member #989 Joined: Sat Sept 08 2007, 02:15AM
Location: SÃ£o Paulo, Brazil
Posts: 476

I put two TC4429 in parallel for each side (so its 4) gave me a few more secs without getting massive heat. I'm using 5R Gate Resistors (Half-bridge) the waveforms aren't already good, as the peak current at the mosfets is about 30A I'm worried about killing the mosfet bridge. I'm using two IRFP260N at 160VDC...

Here is a video of audio modulated:

Also ... , my pll is running at closed loop but I'm using antena feedback. But I think for that application antenna isnt good.

Mr.Black

Thu Oct 27 2011, 09:26AM

Registered Member #3960 Joined: Mon Jun 20 2011, 06:27AM
Location: Wuhan Hubei China
Posts: 33

TIP41/42 totem can work well in the high frequency of 1000KHZ
I think you should probably try this suggestion
This is my HF-SSTC driver PCB 139 14505 295e0c3ad0f72c3

and it is my HF-SSTC 139 14505 D9bc4a315b6c4d0

Inducktion

Thu Oct 27 2011, 02:47PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

Just a quick question; You have a DC blocking cap in series with the GDT, right?

1 2 next

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