Plasmasonic II build / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Plasmasonic II build

1 2 3 4 next last

Move Thread

LAN_403

indiana

Sat Apr 18 2009, 08:57PM

Registered Member #2009 Joined: Tue Mar 03 2009, 07:07PM
Location:
Posts: 43

Hi all, working on putting together the plasmasonic II board, this thread will be for my build, and how it all comes together. I'll post some photos of where I'm at currently if my wife ever lets go of the digital camera. :P

I'm having a bit of an issue right at the get go...

In the instructions, the fist voltage checks list pin 1 of U52 as having a voltage of 18V or thereabouts. I'm getting 10.4V.

The schematic (page 3) shows 12V going to that pin, and a filter cap (as well as pins 10 thru 18). All of these pins measure 10V, and if the schematic's right, it seems like that would make sense.

Typo in the instructions?

Also, the instructions state the freq. adjustment should range from 1KHz to 500Khz. Mine ranges from 100KHz to 500KHz.

Two things... it could be because I haven't installed an esternal 1K Pot. ? Seems like it should be able to adjust all the way down to at least 50 KHz, eh?

Thanks all! (also, progress being made on my plasma speakers as well. i haven't forgotten those! ;) )

schematic

instructions

HV Enthusiast

Mon Apr 20 2009, 11:44AM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

wrote ...
I'm having a bit of an issue right at the get go...

In the instructions, the fist voltage checks list pin 1 of U52 as having a voltage of 18V or thereabouts. I'm getting 10.4V.

The schematic (page 3) shows 12V going to that pin, and a filter cap (as well as pins 10 thru 18). All of these pins measure 10V, and if the schematic's right, it seems like that would make sense.

Typo in the instructions?

All the output pins on the U52 should have a voltage around 10V. 12V Vcc - Vfwd drop of the LED gives about 10V when the LEDs are off.

Sounds like you are getting correct measurements.

indiana

Sun May 03 2009, 08:01PM

Registered Member #2009 Joined: Tue Mar 03 2009, 07:07PM
Location:
Posts: 43

Thank you Dr.

I thought that might be the case. What about the 100KHz to 500KHz range? I'm ordering another 1k pot to try that out.

I've just done the tests on the GDT's I wound, and got some really interesting waveforms.

They're spiking at 15.6V with just over 10 V or so for the top of the square wave, then coming back down to -3V or so, and dropping all the way down to about -9V.

Seems odd. Does this seem normal?

I've lifted the GDT's off the board with a little armature, just to make it easier to connect. The wires are wound very well all around the cores, and they come off clean and tight. Short distances to their contacts after that. I've removed the GDT's from the armature to test it that way too, and it seems not to make a difference.

1241380872 2009 FT67963 Plasmasonicboard

HV Enthusiast

Mon May 04 2009, 03:05PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Waveform seems about right if i recall. Remember, there is deadtime between pulses which is what you are seeing. Important thing it to ensure you have the proper phasing, as indicated, in the instruction manual.

LithiumLord

Mon May 04 2009, 06:56PM

Registered Member #1739 Joined: Fri Oct 03 2008, 10:05AM
Location: Moscow, Russia
Posts: 261

Hmm, something feels wrong here. In fact I never managed to make a GDT drive FETs right with the deadtime induced so I always used a direct gate drive method, but still I can tell the waves look wrong here. ;) Well, the deadtime is perfectly translated, and the ringing is just normal (use TVS or Zeners in order to clamp it within the safer range, also there seems to be too much current for the present leakage inductance, either clamp the current with a higher gate resistor or reduce the leakage inductance by optimizing your GDT, this is not the problem here however) - but I have totally no idea why is the signal asymmetric. Looks like there's an issue with the source signal, as normally if using a single GDT drive mode the gate voltage during the deadtime should follow the zero line, dropping to -Vgs and rising to Vgs (same voltages) during the positive/negative branch. You surely should start off looking at the GDT input traces, in a proper setup they should be two identical square waves with on-time being a little lower (well, for non-modulated systems - here the "little" may get pretty much like 50% or more ;) ) then the off-time, one of those should be 180' phase-shifted. If the waves are non-identical, you should surely check the driver circuitry as that's not exactly how a PWM should behave.

indiana

Mon May 04 2009, 09:51PM

Registered Member #2009 Joined: Tue Mar 03 2009, 07:07PM
Location:
Posts: 43

yep the phasing looks great.

i was surprised to see ac pulses at the gate. how does the negative swing of the voltage affect the fet?

Lithiumlord: the input signals look clean and great. perfect square waves with adjustable frequency and duty cycle. they both only pulse positive, and the voltages look appropriate. this is the first time i've worked on a bridge like this, and i think i might be a bit confused as to how they work (the theory) i thought two of the fets would be pulsing positive voltage, and two pulsing negative voltage, switching. now i'm thinking that's probably not the case, based on the ac wave at the gate.

i'm also surprised to see that it fails to zero out though. once i see the ac signal i get it, but wonder why it doesn't drop to zero after ringing up to 12, then drop to -12 after. instead it immediately drops to -3, then to -9 or so. seemed strange to me as well. though it's dan's design, and he's had great success with it.

so as long as it's looking like his signals looked i'm sure it's sweet. very curious though.

LithiumLord

Tue May 05 2009, 02:31PM

Registered Member #1739 Joined: Fri Oct 03 2008, 10:05AM
Location: Moscow, Russia
Posts: 261

I'm presently @ my gf's, installed Adobe Reader today and watched the schematic at last. It's something close to what I used as my own driver a while ago before moving onto looped deadtime generator, the only difference besides the operating mode is that I used additional drivers after the GDT itself - however I never seen anything like the problem you mentioned. The reason is either a driver fault or measurement issues. Try testing the signal with a winding connected in reverse - if you are getting +3v deadtime voltage it's surely a no-go condition as your switches will not handle the heating from cross-conduction with additional high-resistance state.

uzzors2k

Tue May 05 2009, 02:36PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

Negative voltage on the gate is harmless, and helps prevent the mosfet from turning on out of turn. See the datasheet, the maximum gate voltage will be specified as +/- 20V. I agree that the unsymmetrical waveform is odd. Is the GDT loaded with the mosfets, if so how does it look without them?

LithiumLord wrote ...

Hmm, something feels wrong here. In fact I never managed to make a GDT drive FETs right with the deadtime induced so I always used a direct gate drive method...

A bit OT, but the key is to short circuit the GDT primary during deadtime.

HV Enthusiast

Tue May 05 2009, 10:11PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Your waveforms look about right. I wouldn't worry about doing anything different.

LithiumLord

Tue May 05 2009, 11:30PM

Registered Member #1739 Joined: Fri Oct 03 2008, 10:05AM
Location: Moscow, Russia
Posts: 261

Uzzors wrote ...
Negative voltage on the gate is harmless

Yep, but imagine connecting that secondary in reverse ;) +3v on the gate isn't fun at all ;)

Uzzors wrote ...
A bit OT, but the key is to short circuit the GDT primary during deadtime.

Hehe I know :P The primary was never floating, but I had something wrong with the core itself however so with a significant load it started ringing like hell :)

1 2 3 4 next last

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