1 MHz GDT problems / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

1 MHz GDT problems

1 2 3 4 next last

Move Thread

LAN_403

uzzors2k

Sun Oct 07 2007, 11:29AM

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

I'm putting together a simple gate driver for my SSTC. The frequency is a mere 1.3 MHz, so I thought common components should still be usable. The driver works ok with a 100nF load, but with a real GDT the problems start.

This is the scope shot of the primaries / 100nF cap.

Not very good compared to 100 kHz, but it's usable.

Now, I expected it would be a simple matter of winding a few turns and putting the GDT to work but I was wrong. The output is a massive and CCPS like. Given that the GDT ratio is 1:1 and the drive topology it's quite odd that the voltage isn't halved. I tried two different GDTs, but the waveform only improved marginally with the one. The dummy load on both is a 4.7 nF cap with a 1.2 ohm resistor. The large GDT is a beefy ferrite toroid, the small GDT just a smaller toroid. Both GDTs have more than enough turns for the frequency, since they were used for flyback drivers previously.

Here is the output waveform for the small GDT.

And the large GDT.

I measured inductances and leakage inductance of both GDTs.

Large GDT
Pri inductance : 260 ÂµH
Sec inductance : 300 ÂµH
Leakage : 0.6 ÂµH

Small GDT
Pri inductance : 54 ÂµH
Sec inductance : 54 ÂµH
Leakage : 0.28 ÂµH

So what could be wrong?

EDIT: A larger version of the schematic is now linked.

Dr. Dark Current

Sun Oct 07 2007, 11:44AM

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

Uzzors wrote ...

Both GDTs have more than enough turns for the frequency, since they were used for flyback drivers previously.

So what? Because your frequency is now 20 times larger, you can remove many turns from your GDT. I'm sure this will improve the waveform a lot.

Also for me, high current gate driver ic's (ucc37321/tc4422 etc.) were more than enough to drive a GDT up to high frequencies to moderately large gates, so if you have some of them I would suggest to replace your discrete output stage.

Marko

Sun Oct 07 2007, 11:51AM

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

If your leakage inductance measurements are right it's not a surprise since even your small GDT
gives 4.2Mhz resonant frequency with your mosfet's gate.

You should try a much smaller mosfet, like IRF730.

Even then, you will find it very hard to design a GDT with low enough leakage inductance for Mhz range.

You should calculate the least number of turns possible, and possibly use shielded cable (those small double-cored coaxes are nice), with shield for your primary, and cores for secondaries.

If you still aren't happy you can use a discrete driver.
You can use UCC to directly drive a mosfet, with GDT providing power and signal in CW, or use separate PS.

This technique is often used and is very grateful for operation into several Mhz.

This is a 2Mhz waveform from a single UCC driving a IRFP450:

ragnar

Sun Oct 07 2007, 11:57AM

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

- Use the MOSFET you actually intend to drive, as the "dummy load"
- You should need fewer than ten turns on primary/secondary,
- Try 2 to 10 ohm resistors for lowering the Q of the circuit,
- Ensure the transformer is wound neatly (bifilar) and the input and output leads are twisted together.

Dr. Slack

Sun Oct 07 2007, 09:51PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

peeps use a turn or two of coax for making transformers in the HF and even VHF bands. Coax gives you good mutual coupling, and a very calculable (and modellable) mutual capacitiance and coupling.

Steve Conner

Mon Oct 08 2007, 08:47AM

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

We investigated that quite a lot last year:

Danielle

Mon Oct 08 2007, 05:10PM

Registered Member #632 Joined: Mon Apr 09 2007, 01:09AM
Location:
Posts: 267

well I was able to get past the problem by using a commercial HF GDT heres the one.

those should work past the specked 2Mhz

Experimentonomen

Mon Oct 08 2007, 06:30PM

Registered Member #941 Joined: Sun Aug 05 2007, 10:09AM
Location: in a swedish junk pile
Posts: 497

Uhh that link takes me back to forum main site.

uzzors2k

Mon Oct 08 2007, 07:12PM

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

I've fixed up the GDT and ruled it out as a source of error for now. The odd waveform seemed to be caused by ringing between the dummy capacitor and leakage inductance. With the IRFP450 gate capacitance the waveform improved greatly. There was still a bunch of ringing though, which I traced back to the driver. When the push-pull stage drives any one of the mosfets alone the waveform looks like this:

With both mosfets the waveform turns into this:

Nastee. Both scope shots were taken at the gates of the IRF630/IRF9530s, but the ringing is amplified all the way to the GDT load. At the load the ringing is much worse, like 10V p-p worse. I tried replacing the mosfets with TIP131 and 132 Darlington transistors, but the IRFP450 gate waveform becomes too sine. Not enough current I think. The ringing went away at least.

How should I drive the GDT driver gates properly? I haven't got any gate driver chips or I would have used them ages ago. The BD139 / 140 push-pull stage is good, but when the gates are put in series and driven things go wrong. Suggestions? Are darlingtons the way to go, just with higher current capacity than 5A?

Bjørn

Mon Oct 08 2007, 07:12PM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

Copy the URL and paste it into the address line in your browser, that should work.

http://www.coilcraft.com/da23xx_gate.cfm

1 2 3 4 next last

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