Killed ZVS. Any ideas why? / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Killed ZVS. Any ideas why?

Move Thread

LAN_403

kiat

Mon Dec 07 2009, 07:25AM

Registered Member #2115 Joined: Fri May 08 2009, 01:17PM
Location: Singapore
Posts: 46

Hi all,

Today I constructed a zvs, using a dedicated 12v supply for the igbt switches. My main supply was a 35vac 5.5a (per tap) centre tapped transformer - i used one tap.

I rectified the output with a silec BD36 931 diode bridge, smoothing was done with a 63vdc 10000uf electrolytic cap. When rectified and smoothed, output voltage was just under 50vdc.

The 12v goes through 2 100 ohm 2W resistors to get to the gates of the igbts.
I use UF4007 fast recovery diodes, IGBTs are HGTG11N120CND.
Inductor is 30 turns round a yellow/white iron powder core.
Caps are 4 pcs 1.2uf 450vac mmc. 2 in series 2 in parallel making a 1.2uf @ 900vac cap bank.
FBT is a big DC diode split type, the seller said its rated 50-60kv (i tried-makes 5.5cm starting arcs)
FBT has 5+5 turns for the primary.
I did not use a variac, just on-off with mains switch.

At first i switched on gate power (12v supply) and it made small sparks (from gate power). then i turned on the main supply and got fat white arcs.

But for my 2nd try, i switched off everything, then switched on the main supply and finally the gate supply. Then i saw smoke coming off something...couldn't see clearly, and the wires coming from the main supply and on the inductor heated up real hot and smoked too.

Anybody, know what failed and why, what i should do next time?

Thanks,
kiat

----------------------------- -----------------------
UPDATE:I deduced that only the 2 UF4007 diodes or the 2 HGTG11N120CND IGBT's failed. The power transformer diode bridge is ok, electrolytic cap's fine, how can a 900v cap bank fail with Vin of only 50? so they must be fine. The inductor can't fail and the resistors were measured -- still 100ohm each.
-------------------------------------------- --------
UPDATE: I replaced the diodes- still no spark with gate power (i didnt dare to use full power)
i tested IGBT's they conducted from the collecter to the emmiter
------------------------------------------ ----------
UPDATE: I replaced igbts. guess what -- works again!

heheheh... that defeats the whole purpose of this thread... but i still want to know why it failed with main power 1st followed by gate power.
------------------------------------------------- ---
UPDATE: SIGH! i killed it again...after a few secs of using...perhaps my igbts are crappy... but im sure HGTG11N120's can do the job... perhaps its because i bought them from a cheapo china seller on ebay...anyways gonna try IRFP460's now...

Antonio

Mon Dec 07 2009, 01:59PM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

Try a larger series inductor, and make it sure that it does not saturate. Verify also if just 12 V at the gates is enough to saturate the IGBTs. This oscillator has another mode of oscillation where it oscillates slowly and draws a much higher current. A larger series inductor helps in avoiding this mode. Drawing arcs without current limitation may also cause the irregular oscillation. Look at this simulation, that shows the two modes:

Firefox

Mon Dec 07 2009, 03:44PM

Registered Member #1389 Joined: Thu Mar 13 2008, 12:50AM
Location: Pittsburgh, PA
Posts: 346

IGBTs are relatively fragile compared to MOSFETs in this duty, it seems (I've never killed a MOSFET), so extra care needs to be taken to make sure it does not fall out of oscillation. On this note, my experimentation with the ZVS flyback driver seems to indicate that a larger primary inductance and a smaller tank capacitor can help reduce the likelyhood of oscillations. A larger input inductance helps as well.

Also, your IGBTs are overkill. You may want to consider switching to a 200V mosfet rather than those 1200V IGBTs, which would be better used elsewhere.

Dr. Dark Current

Mon Dec 07 2009, 06:44PM

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

I had the same problem too, with 100ohm gate resistors and 12V supply for them.
The answer probably is "parasitic oscillation" which has been associated with this circuit ever since.
Try to put some small caps (?few nF to 10nF) between collector and emitter of each transistor. This could get rid of the oscillation, but I have not tried this so do at your own risk

kiat

Tue Dec 08 2009, 04:39AM

Registered Member #2115 Joined: Fri May 08 2009, 01:17PM
Location: Singapore
Posts: 46

Fixed the zvs. used american igbts and the problem completely goes away.sparks are extremely hot; i accidently blistered my finger touching a screwdriver that was arced on. I will start a new thread on experimentation with super-high value tank caps (here in singapore caps are very cheap -- i got an 8uf 250vac cap for 4.50 sgd)

Colin 99

Tue Dec 08 2009, 11:19PM

Registered Member #192 Joined: Fri Feb 17 2006, 03:08AM
Location: Canada
Posts: 44

Kiat,

I've blown my IGBTs too and what I did wrong was to turn down the gate power supply. So what I learned was first you turn on your gate supply, then turn on the main (high power) supply. Turn off in reverse order only, or have them both running off the same supply.

50 volts is a pretty high voltage to run at, I'm using 36 volts from a battery supply for main power and now I run my gate supply off of the same 36 volts. My system will draw 13 amps off of the batteries when pulling a long arc off of the flyback. You might want to use a lower voltage to save from burning out your flyback or use more turns (not 5 + 5).

For your iron powder toroid, I calculated about 42 turns on a yellow / white torroid, and I ordered larger torroids to avoid saturation, not sure where that starts though. Use a thick length of magnet wire like 14 or 16 guage.

happy experimenting,
Shaun

kiat

Wed Dec 09 2009, 07:27AM

Registered Member #2115 Joined: Fri May 08 2009, 01:17PM
Location: Singapore
Posts: 46

Shaun,

30 turns on a powdered iron core works fine for me, the driver does not have parasitic oscillations. 50 volts so far seems ok (current draw is 10 amps when arcing) if runs are kept short - caps barely get warm, only the igbts heat up, but still not alot. The 5+5 primary and the ferrite core gets real hot, the flyback secondary gets touch-warm after prolonged arcing.(i use 2 lengths of 1.25mm2 multi strand copper cable in parallel)

Yes, now i always turn on the gate supply followed by the main supply, and switch the main supply off first followed by the gate supply (if not the transistors will lock in the on state and get massive current flow through - my supply transformer for the circuit has 80 amps short circuit current)

Thanks,
kiat

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