Half-bridge flyback driver not working... / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Half-bridge flyback driver not working...

1 2 3 next

Move Thread

LAN_403

Dr. Shark

Sun Feb 19 2006, 07:09PM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

This thing has allready been in the "Flyback driver contest" thread, but it is giving me more and more headaches, so I am asking for help here.

It is a 555 oscillator driving a push-pull arangement of 4421/4422 gate drivers into a half-bridge of 600V 27A fast IGBTs. They are protected by 300V TVS's, antiseries 12V zeners and have 4.7Ohm gate resistors. Mid-point voltage is provided by two 3u3 250V caps. I am driving it of current-limmited, rectified mains. When I set the frequency of the 555 to a value the transfomer does not like, (range is 10-50kHz) it kind of works, giving sparks of about an inch. When I tune the frequency, the IGBTs die. I have killed four sets now, so getting help with this would actually save me money

Oh, one other thing that might help helpers: When the thing works, it does not give countinous arcs, but more like sparks with a couple of Hz repetition rate. I guess this is due to my filter capacitor taking a while to charge up enough to allow the output voltage to rise enough to break down the air between the electrodes.

Sulaiman

Sun Feb 19 2006, 07:35PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

1)
can you post a circuit diagram?
2)
Is your 555 producing a 50% duty squarewave?
3)
Assuming you have a Gate Drive Transformer,
is there a capacitor in series with the primary?
(there should be)
4)
2 x 3.3 uF will be resonant with the flyback primary
somewhere in the high audio range,
to prevent disaster put a diode across each 3.3 uF capacitor
to 'kill' any resonant operation.

(IF you want to operate in primary resonant mode
(MASSIVE output)
then use suitable foil/film capacitors, NOT electrolytics)

Marko

Sun Feb 19 2006, 07:37PM

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

This ''repetiting' of output seems bad, maybe because of current limiting, maybe you should (once it works well) give more current and it should stop...

12V gate zener seems low, are you sure IGBTS are opening fully??

IGBTS may blow like that if oscillator stops for (any) reason leaving outputs high, thats main reason why gate drive transformers are used (DC conponent cannot pass and leave transistors shorted trough few turns of wire)

Maybe (depending on driver) GDT may be inserted if this is a real problem.

You must make sure that igbts never get on and non-oscilating as it means quick death.

Yerterday I made simplest possible flyback driver, 2N3055 + feedback

and ran on 40V it gave fat and very hot spark that started at 1cm and could be dragged over an inch (that flyback was very large)
Arc was more yellow, almost mot-like, realy good for that kind of drive...

After long torture and overheating (I was having fun melting steel wool iwth incredibly hot

I plan to get bunch of some more powerful bipolar transistors and use them there.
If you 'need it now' for your marx that may be solution, maybe even IGBT could be used with some adjustmens in feedback offset, then you could give it far more power than I did for 2N3055...

Dr. Shark

Mon Feb 20 2006, 11:18AM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

I'll draw up a circuit diagram later, but it is really just the standart half-bridge arrangement. The only difference to steves ciruit here

is that I am using a push-pull gate driver stage and 1:1 windings on the GDT. The 555 is giving a nice D=50% square wave, and I am using a GDT with capacitor in series that gives me a very nice square wave on the IGBTs gates.

The point about 12V not being enough drive voltage is very good, in fact the data sheet states it should be 15V (I am using IRg4pc50 IGBTs). But since the total power in to the system is limited by a small 40VA isolation transformer that has about 1A of short circuit current, it should not be possible for an IGBT to die even if it was in the linear region all the time. I think it can easily dissipate more than 100W before it takes thermal damage.

The point about the caps in the halfbridge resonating with the leakage inductance of the flyback is something I did not consider at all. I would think that this could only happen if a cap was directly placed in series with the primary. I always thought of the caps in a halfbridge as just providing a midpoint voltage.

Marko

Mon Feb 20 2006, 01:54PM

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

Putting diodes may be a try but the most obivous problem is low gate voltage - with 1:1 ratio gdt output doesnt reach even supply voltage (when you consider losses).

Igbts simply turn partially on and get pulverized because most of dissipation is on them (not the flyback).
Use 1:2 transformer and 17 volt zener diodes. Thay will block spikes that could damage the gate and transistors will turn on surely.

You can also put some zener diodes across ce, around 400V in order that IGBT doesnt blow because of overvoltage.

Dr. Shark

Mon Feb 20 2006, 02:19PM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

I'll try that tonight. I was under the impression that the IGBTs would saturate at 6V or so, where there is the V_ge vs. Q_g curve, and the recommended 15V are only to speed up switching.
My GDTs are all 1:1 because I originally designed them for high speed operation so I had to keep leakage inductance to a minimum. But for this application I think I can get away with much more "sloppy" GDTs.

Marko

Mon Feb 20 2006, 02:44PM

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

Also good idea is to increase total number of turns on gdt, keeping 1:2 ratio, gate driver ICs should never get hot that amy be because of too few windings...

Dr. Shark

Mon Feb 20 2006, 08:22PM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

I redid the GDT for 24V output, but I decided _not_ to test it: The data sheet of the IGBTs clearly states that the gate threshold voltage is 3-6V, and at 10V it can already pass 100A pulses with less than 10V drop. So I am pretty sure that 12V will saturate it, and 15V or 20V will just do the same a bit more quickly, or keep it saturated at high current pulses. I am pretty sure a higher gate voltage will not keep the IGBTs from dying, so I better think of something else before I risk the last few IGBTs I have left.
Shit, I am really running out of ideas...

Dr. Dark Current

Mon Feb 20 2006, 08:31PM

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

How many turns do you have on primary?

more than 30 for 120V mains or more than 60 for 220 mains is usually OK
I also see that your flyback is generating quite high voltages even when not in tune, indicating you have too few turns on primary... In tune the output voltage is about 5x the out-of-tune output voltage.

Dr. Shark

Mon Feb 20 2006, 08:47PM

Registered Member #75 Joined: Thu Feb 09 2006, 09:30AM
Location: Montana, USA
Posts: 711

Umm, I was using 5 turns

I was getting up to 4cm arcs when the IGBTs were dying, and they were also quite yellow/white and not purple/blue, indicating that a few watts of power must have been there.

1 2 3 next

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