GDT problems / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

GDT problems

previous 1 2

Move Thread

LAN_403

Andrew H

Wed Jan 09 2008, 04:46PM

Registered Member #1198 Joined: Sat Dec 29 2007, 05:39AM
Location: Seattle, WA
Posts: 21

I do have a DC blocking cap in series with the GDT, so that isn't the problem. I adjusted the duty cycle to about 50/50 and that did not fix the problem either (it actually made it more apparent). I'll try and rewind the GDT without the outer sheath so I can fit more turns on it. After reading Richie's site and some of the comments here I think I don't have enough turns.

The scope was set to .2 mS/DIV in the picture I posted.

Marko

Wed Jan 09 2008, 06:16PM

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

Single-turn GDTs are acceptable at insane frequencies. It thereby goes that two-turn GDTs are appropriate at insane/2 frequencies, and a four-turn GDT is only appropriate at insane/4 frequencies.

Lol, Matt, I hope that isn't a professional opinion

Where do you just get that cool copper foil tape?

If what Andrew is saying is right he's running at a bit over 1kHz.

You need to increase the frequency by a like factor of 100 to get into ballpark.

The best would be if you tested the GDT with your coil's operating frequency.

Wolfram

Wed Jan 09 2008, 06:54PM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

Andrew H wrote ...

I'll try and rewind the GDT without the outer sheath so I can fit more turns on it.

With the CAT5 wire you're using, you already have 8x 4-turn windings on it. Why not just series two of them? Also remember to series the secondaries too, if you don't the gate voltage is just half of the UCC supply voltage. However, if you are running a fullbridge with it, you don't have enough windings. In that case, removing the outer sheath, as you suggested, is a good way to get space for more turns. My favorite way of making GDTs is to use some CAT5 cable without the sheath, wind an appropriate number of turns on the core, parallel all the white pairs, and use each of the colored pairs to drive a FET. This should keep the leakage inductance down. Also, twisting each pair from the GDT is a good idea.

The frequency you're driving it with is way off for TC duty. Make an oscillator that gives you 100kHz-1MHz at 50% duty cycle. Use a load on the GDT that is similar to what you are going to drive. If there are still spikes, use gate resistors, they can help greatly.

Marko's suggestion, using a 7474, is good. This will ensure almost* perfect 50% duty cycle. Notice that the output frequency will only be half that of the oscillator you're using. BTW, the 7474 is a flipflop, not a monostable.

* usually not quite perfect, because of jitter from the oscillator. Close enough for this application though, probably by several orders of magnitude.

Edit: Here's

a picture of my SSTC gate waveform before and after I added gate resistors. 5V/div, 500nS/div

Anders M.

Andrew H

Thu Jan 10 2008, 06:23AM

Registered Member #1198 Joined: Sat Dec 29 2007, 05:39AM
Location: Seattle, WA
Posts: 21

I changed the frequency of the 555 to something around 40 kHz (I used the smallest cap I have on hand for the timing cap). I also put a 3900pF cap across one of the GDT outputs. The wave form improved dramatically. There now is no consistent droop. However, the output oscillates at the top of the pulses. According to Richie's site that means I need to reduce leakage inductance and add a dampening resistor.

I'll post pictures after my camera battery charges.

EDIT:
I added a 3ohm resistor in series with the capacitor across the output. This made the output wave even worse!

EDIT2: Here is a picture of the output with a 3900pF + 3ohm load

1199982394 1198 FT1630 Gdt1 Out 3900pf 3r Load

I think I need to lower leakage inductance to get a better output. I'm going to try removing the outer sheath and rewinding the GDT.

Steve Ward

Thu Jan 10 2008, 07:07PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

You are measuring *after* the 3 ohms, right (that is, directly across the capacitor).

Andrew H

Fri Jan 11 2008, 06:41AM

Registered Member #1198 Joined: Sat Dec 29 2007, 05:39AM
Location: Seattle, WA
Posts: 21

Steve Ward wrote ...

You are measuring *after* the 3 ohms, right (that is, directly across the capacitor).

I was measuring after the dummy load.

I'm pretty sure my problems were related to either excessive leakage inductance or not enough turns. I removed the sheath of the CAT5 cable and made a second GDT:

The output of the new GTD is much better:

1200033464 1198 FT1630 Gdt2 Out 3900pf 3r Load

uzzors2k

Fri Jan 11 2008, 01:37PM

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

It still seems slightly wavy, are you running it at 40kHz? Really there's no point in testing that far away from your TC's operating frequency as the waveform may be substantially different. I suspect it'll look just as bad as your previous GDT at 200kHz or so. I put the resistor on the primary side, this way you only need one resistor and it won't burn if the gate shorts.

previous 1 2

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