UCC37322/1 DRSSTC Gate Driver Problems / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

UCC37322/1 DRSSTC Gate Driver Problems

1 2 3 next

Move Thread

LAN_403

Daniel Kramnik

Sat Jul 23 2011, 10:39PM

Registered Member #3885 Joined: Sun May 15 2011, 12:47AM
Location: Newton, Massachusetts, United States
Posts: 94

Hello,

I've been having some problems getting my DRSSTC running. I first tried a primary feedback driver, but couldn't get it to work properly. Then, I switched to a secondary feedback oscillator with a flip-flop to synchronize switching, however, this failed too. Exasperated, I disassembled one of my old SSTCs and used antenna feedback to finally achieve first light. Unfortunately, though, the old SSTC driver has some sort of internal problem that is causing voltage spikes on one of the gates (I killed a couple IGBTs this way before I decided to scope the gates to see what was happening), so I decided to grab some spare parts and build a new driver with secondary feedback but no flip-flop switching synchronization.

The problem I am having is that the UCC gate drivers are overheating and exploding and that I cannot get the coil to oscillate. I know that my interrupter is good because it works with my SSTCs and I know that my GDT is fine because it briefly worked with my primary feedback driver (see link).

When I tried adding a 250kHz start up oscillator (my coil resonates somewhere between 100 and 200kHz, so I figured it was close enough) on pin 1 of my 74HC14 and disconnecting the GDT secondary windings from the gates of my IGBTs, the heating problem went away, but when I reconnected the GDT secondary windings, one chip immediately exploded and the other overheated and melted. At this point, I have probably already killed a couple dozen gate drivers with this coil.

Has anyone else encountered this problem before, and if so, do you have any suggestions? I do have a camera and an oscilloscope, so I can post pictures/signal waveforms if that would help.

Thanks!

Daniel

Edit: Just uploaded a YouTube video of first light: http://www.youtube.com/watch?v=Z851RUa_TMw

Dr. ISOTOP

Sun Jul 24 2011, 02:55PM

Registered Member #2919 Joined: Fri Jun 11 2010, 06:30PM
Location: Cambridge, MA
Posts: 652

What IGBTs and gate resistors are you using? UCC's get very angry if you try to drive large transistors with no gate resistor.

Daniel Kramnik

Sun Jul 24 2011, 05:14PM

Registered Member #3885 Joined: Sun May 15 2011, 12:47AM
Location: Newton, Massachusetts, United States
Posts: 94

bwang wrote ...

What IGBTs and gate resistors are you using? UCC's get very angry if you try to drive large transistors with no gate resistor.

I'm using STGW40N120KD TO-247 IGBTs with 5 ohms on the gates, I added the resistors in accordance with advice I received here. Also, for now, I'm not using some of the windings on the GDT so my ratio is only 1:1:1; this should easily give me enough voltage to drive my IGBTs to between 100 and 200Apk, which is enough to get started. I don't think that the problem comes after the GDT because I used that same bridge with my old SSTC antenna feedback controller (identical to the driver I am testing now except that my new driver has larger bypass capacitors across the UCCs and secondary feedback instead of an antenna) without problems.

Dr. ISOTOP

Sun Jul 24 2011, 08:34PM

Registered Member #2919 Joined: Fri Jun 11 2010, 06:30PM
Location: Cambridge, MA
Posts: 652

Do you get a decent waveform on the output of the unloaded GDT?
You *sure* you aren't shorting the outputs when you connect to the bridge?

Daniel Kramnik

Sun Jul 24 2011, 08:52PM

Registered Member #3885 Joined: Sun May 15 2011, 12:47AM
Location: Newton, Massachusetts, United States
Posts: 94

Some people in the chat room pointed out that I need clamping diodes on the outputs of the gate drivers. Steve Ward's older designs don't use clamping diodes, so I figured I might not need them, but then again, the designs I'm referencing aren't even DRSSTCs! Right now, I'm going to try adding 1N4148s (what I have on hand) and scoping the output of the GDT.

On a different note, could the huge difference in frequency between my startup oscillator and my actual resonant frequency be having an ill effect? I ran out of 100k resistors to limit the startup signal, so I grabbed a 33k resistor instead; if the signal from the oscillator isn't limited enough, could it be preventing the feedback signal from becoming dominant and taking over?

Also, how much current limiting should the feedback signal have? I'm using a 1k resistor for limiting and a 100nF capacitor for DC blocking, could the 1k resistor also be contributing to the problem? I'm afraid to take the resistor out because if I turn the bus voltage up, I think it might kill my 74HC14's input.

Update: I tried again with the clamping diodes and with a scope connected to the GDT. Unfortunately, the UCCs immediately overheated and started letting out smoke. I got some sort of signal on my scope, but it was just a blur, I didn't have enough time to change the settings to figure out what it was. My scope was set to 2uS per division so that 1 period at 250kHz would be exactly 2 divisions. The fact that the signal was blurred indicates to me that it was at a completely different frequency, what on EARTH is going on?

I only have 1 more set of UCCs to try, so I don't want to mess it up!

Is there any simple way to build a gate driver without UCCs that doesn't use a totem pole topology? I don't have any powerful PNP transistors at the moment.

Inducktion

Tue Jul 26 2011, 12:53AM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

UCC's need to have a DC blocking cap before the GDT, otherwise they will get very very hot and very very unhappy.

A resistor (like 10 ohms) before the GDT also helps too. Use around 1 uF for the DC block cap. It's worked for me.

Daniel Kramnik

Tue Jul 26 2011, 02:15AM

Registered Member #3885 Joined: Sun May 15 2011, 12:47AM
Location: Newton, Massachusetts, United States
Posts: 94

Inducktion wrote ...

UCC's need to have a DC blocking cap before the GDT, otherwise they will get very very hot and very very unhappy.

A resistor (like 10 ohms) before the GDT also helps too. Use around 1 uF for the DC block cap. It's worked for me.

I was using a 100nF DC blocking capacitor in some of my earlier experiments with this driver, but then I switched to a 1uF capacitor. That doesn't seem to be the problem, however, because the UCCs instantly overheat in either case. I'm also not too convinced that I need a resistor in series with the GDT primary because nothing overheats if the bridge is not connected.

On the other hand, perhaps when the GDT secondaries are connected to the bridge, this draws more current from the UCCs and causes them to overheat. I'm not sure if this is the case, but it might be worth a shot. I'll try adding a 10 ohm resistor before the GDT, lowering the frequency of the startup oscillator to around 100 to 150kHz, and scoping the output with my final pair of UCCs.

One quick question; I'm new to oscilloscopes and I haven't figured everything out yet, so I'm not sure how to correctly connect my probe to measure the GDT input and output: should I connect the ground clip to the ground point in my circuit and the test probe to one side of the GDT winding I want to scope, or should I connect the clip to one side of the GDT and the probe to the other? Does how I connect it depend on whether I'm measuring the GDT input or output, or do I do the same thing in both cases? I have a dual trace oscilloscope, so I'd like to connect one probe to the input and the other probe to an output. Thanks!

Dr. ISOTOP

Tue Jul 26 2011, 03:00AM

Registered Member #2919 Joined: Fri Jun 11 2010, 06:30PM
Location: Cambridge, MA
Posts: 652

With no bridge connected, the UCCs drive no load, hence they don't get hot.
Try a 1 uF DC blocking cap, a 6.8R resistor on the primary side, and no secondary resistors; this has always worked for me while driving small IGBTs.

Inducktion

Tue Jul 26 2011, 03:01PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

You can connect the measuring probe to one side of the GDT and the ground to the other side. This lets you see both parts of the signal (I.E. the true signal that's being output)

Daniel Kramnik

Tue Jul 26 2011, 10:24PM

Registered Member #3885 Joined: Sun May 15 2011, 12:47AM
Location: Newton, Massachusetts, United States
Posts: 94

bwang wrote ...

Try a 1 uF DC blocking cap, a 6.8R resistor on the primary side, and no secondary resistors; this has always worked for me while driving small IGBTs.

I just tried doing exactly that (1uF blocking, 6.8R primary resistance, no gate resistors) and lowering the startup frequency to 100kHz, but nothing happened. I scoped some sort of strange noise across the GDT input winding, but that was it. I tried reversing the phasing of the GDT, but that didn't help.

Also, now that I've added the primary resistor, only one UCC chip gets hot: the one that's not attached to the GDT primary winding through the resistor. I tried switching the position of the resistor and that caused the other chip to heat up. Should I replace the 6.8R resistor with a 3.4R resistor attached to each side for 6.8R total to get rid of the heating problem? Has this ever happened to anyone before?

Is it possible that I have something hooked up the wrong way around in my feedback loop? Even if the UCCs heat up, I don't think that should prevent the coil from oscillating. Am I right?

1 2 3 next

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