GDT problems / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

GDT problems

1 2 next

Move Thread

LAN_403

Andrew H

Wed Jan 09 2008, 07:01AM

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

I constructed a GDT for my first SSTC last night. I made it by wrapping CAT5 cable around a ferrite torrid (in the fashion that Steve Ward used on his DRSSTC1). Here is a picture of the GDT:

I know the leads are a bit long and that is probably negatively affecting performance but I don't want to chop them down until I build the H-bridge and know exactly how long they have to be. I connected it to a set of TI driver chips (UCC37321 and UCC37322). A 555 timer provided the input to the driver chips. Here is the output relative to ground:

1199861342 1198 FT1630 Gdt Output First Test

The output has a huge over/under shoot (when not even connected to a MOSFET!) and the output droops substantially. I'm going to try removing the outer case of the CAT5 and winding the wires in a polyfilar bundle. Hopefully that will lower leakage inducance and help the output.

What else might help my GDT produce a cleaner output?

...

Wed Jan 09 2008, 07:22AM

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

You need to at leat attach a dummy load to the output (like a 1nf cap or so, make it about the same size as the gate on your transistors) to see what is happening. Also, to get rid of the spikes the traditional route is to add back to back zener diode across the output.

Steve Ward

Wed Jan 09 2008, 07:30AM

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

Be sure to twist the primary and secondary leads together to lower the wiring inductance. Be sure you arent saturating the thing! The droopy-ness of that waveform looks highly suspicious.

If you really followed the same technique i used, the coupling is very high. But it can be difficult to fit enough turns on a smaller core to avoid saturation, so sometimes its best to only use the number of wire strands that you absolutely need.

Andrew H

Wed Jan 09 2008, 08:32AM

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

For testing I am using one wire for the primary and one for the secondary (1:1 ratio). Other than those two wires all the rest of the wires in the CAT5 cable are not connected to anything. I don't think that I could be saturating the core that easily. How would I be able to tell if I was saturating it?

I tried attaching a small ceramic capacitor across the output but it didn't change anything.

Kizmo

Wed Jan 09 2008, 10:40AM

Registered Member #599 Joined: Thu Mar 22 2007, 07:40PM
Location: Northern Finland, Rovaniemi
Posts: 624

Take a look here:

ragnar

Wed Jan 09 2008, 10:41AM

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

wrote ...
I don't think that I could be saturating the core that easily. How would I be able to tell if I was saturating it?

Actually, with only four or five turns, I'm pretty sure you'd be close to saturating the core at this considerably low frequency. The "drooping" which Steve Ward refers to is a good indicator that something's wrong; double your turncount.

Wolfram

Wed Jan 09 2008, 11:54AM

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

Am I the only one that notices that his duty cycle is way off? Having a duty cycle other than 50% is an easy way to saturate the core if you don't have a DC blocking cap before the GDT. Either way, get the duty cycle as close to 50% as you can, this is going to improve the waveform too. I agree with many of the others in this thread, you seem to have too few turns on the GDT. Try to connect two windings in series for 8 turns.

HV Enthusiast

Wed Jan 09 2008, 12:02PM

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

You need more turns. Double your turn count and go from there.

Marko

Wed Jan 09 2008, 02:00PM

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

You guys all just read andrew's mind for the frequency (which he didn't mention anywhere)?

And yes, nobody notices anything wrong about duty cycle?

Andrew, you need closely 50% duty cycle input to GDT, and use a DC blocking capacitor in series with your driver output, otherwise small DC caused by uneven duty cycle saturates the core.

7474 bistable can give you good 50% duty cycle output.

I don't know why are you guys panicked for saturation so much, unless we are in audio frequencies here.

I used 6 turns on about 2x smaller core for my DRSSTC at 100kHz, and it worked very well.

If in doubt just use datastream's site.

ragnar

Wed Jan 09 2008, 02:42PM

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

wrote ...
You guys all just read andrew's mind for the frequency (which he didn't mention anywhere)?

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.

Since this is Andrew's first coil, I safely assume that the frequency is not more than quarter-insane. Therefore, he should use more than four turns on his gate-driver transformer.

Hence the summarized advice:
double your turn-count.
tweak the 555 until the duty is much closer to 50%.
use DC-blocking caps in series with the GDT primary if not already doing so.

1 2 next

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