GDT problem... / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

GDT problem...

1 2 next

Move Thread

LAN_403

RateReducer

Tue May 22 2012, 10:23PM

Registered Member #4603 Joined: Wed Apr 25 2012, 07:33PM
Location: Austria
Posts: 159

Hi!
I have a problem with my GDT. It drives a IGBT halfbridge in my DRSSTC. The datasheet of the IGBTs says 50nF
I use this circuit:

I get this at the output of the right GDT every time the burst is over...

The oscillation is also on the primary of the GDT!!!

Signal on the gates of the 2 MOSFETs:(left GDT)

How can i get rid of this oscillation?

Greetings

Neet Studio

Tue May 22 2012, 11:39PM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

Disclaimer: I am not the expert on this subject. Here is what I understand from ringing in switch mode power supplies:

Basically the L from the GDT and the C from the MOSFET (gate & interconnect capacitance) forms a resonant circuit.
See Class E amplify stuff here for an explanation of the resonant.

In your case, you want the oscillation to die down as quickly as possibly.

To damp out the oscillations, you'll need to introduce some lossy elements in your LC circuit to dissipate the stored energy. It is a transformer, so energy can certainly go back to the primary side.

Things like gate series resistance, some shunt resistance to the gate windings(?) would help some what. Also possibly a RC snubber across the primary winding with a time constant close to the period of the oscillation.

Steve Conner

Wed May 23 2012, 06:40AM

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

Yep, I think the RC snubber is your best bet.

The fundamental problem is that when the burst is done, both MOSFETs turn off, and there is nothing to define the voltage on the GDT. So it oscillates as explained above.

RateReducer

Wed May 23 2012, 10:06AM

Registered Member #4603 Joined: Wed Apr 25 2012, 07:33PM
Location: Austria
Posts: 159

Thank you both for the useful help!!

Neet Studio, the RC snubber on the GDT works perfect! The ringing is nearly completely gone. The rise-time is not affected in any way.
Steve, thank your for the explanation with the floating GDT primary. I have not thought about it enough...

Edit: The remaining oscillation on the gate seems to be powerful enough to switch the IGBT messing up the CE signal -_-

I think i am going to use the GDT output stage of Steve Wards driver.
Will it solve the problem with the floating GDT primary?

greetings

Dr. Dark Current

Wed May 23 2012, 02:02PM

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

Use a H-bridge at 30 volts to drive the GDT, this way you can short the primary out after the burst is gone.

Steve Conner

Wed May 23 2012, 02:11PM

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

That won't work any better, if this H-bridge is also driven by a GDT.

You need some different kind of driver that allows for a third state, with the two top devices or two bottom ones turned on.

You can do this with the popular UCC gate drive chips. I think most of the DRSSTC driver circuits out there will drive both outputs low when the interrupter pulse is off, which shorts the GDT primary.

RateReducer

Wed May 23 2012, 03:17PM

Registered Member #4603 Joined: Wed Apr 25 2012, 07:33PM
Location: Austria
Posts: 159

That is not good...
If i use a GDT the voltage rating of the UCC drivers is too low for my IGBTs.
I think i am going to use a halfbridge of 2 MOSFETs on every IGBT gate.
A new GDT with 4 secondaries driven by the left GDT driver. 2 coils for every halfbridge.
Than there is no coil that could oscillate with the gate capacitance.

greetings

Marko

Wed May 23 2012, 05:17PM

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

Yep, floating the GDT primary at the end of the burst is a big no-no and you need to always keep the primary shorted by a low impedance when you want it to be off.

Steve Ward has had success with level shifted mosfet half bridges as a gate driver solution, see his driver schematic for an example.

If you want a simple and robust driver solution that handles higher voltage, go for IXDN614 or IXDN630 parts. Their older brother IXDD414 (now obsolete) has served me and many other people well so far.

Marko

RateReducer

Wed May 23 2012, 05:34PM

Registered Member #4603 Joined: Wed Apr 25 2012, 07:33PM
Location: Austria
Posts: 159

I live in Austria. It is hard for me to get the IXDNs cheaply...

A few posts earlier i mentioned Steves circuit as it shorts the primary.
I will try it out. I hope it will work properly.

Thanks for all Your replies!

greetings

Neet Studio

Wed May 23 2012, 05:42PM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

I noticed this on page 24 of On Semi's App note: "Half-Bridge Drivers A Transformer or an All-Silicon Drive?"

Is the design finished?
-> No, not yet. Pay attention to the ringing among C C1 , C C2 and driver transformer when skip or UVLO. A diode and resistor to damp the ringing. :(

They have a similar situation (pulse skipping/UVLO) when the driver circuit is halted. The energy is drained via the diode to the push-pull driver pair with output held at low.

If it is not possible to have the gate driver held at a known state via weak pull-up/pull downs resistors, may be you could add an auxiliary winding that you can shunt or bleed off energy with a MOSFET when the driver is not active?

(need a diode in series so that the MOSFET body diode isn't shorting the winding for negative voltages).

1 2 next

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