ZVS driver modification / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

ZVS driver modification

Move Thread

LAN_403

Inducktion

Sat Jan 09 2016, 10:15PM

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

Hello everyone.

I've recently been fiddling around with the ZVS driver we all know and love in an attempt to fix some of its better known drawbacks.

I have successfully made the driver so it doesn't seem to require an external power supply for gate driving; instead it uses direct feedback from the resonant tank.

However upon looking at the waveforms for the driver (and comparing it to a normal ZVS driver) it seems to have a slight "hitch" on the gate drive.

Here's the schematic for it and two oscilloscope captures.

All of the resistors are 470 ohms, D1 and D2 are ultra fast rectifiers and D3 and D4 are 12 volt zeners.

,11WPw67,QiusvL6

The second image is of the modified ZVS; Yellow is the resonant tank, and the blue is the gate. They're both referenced by the same ground, which is held at one side of the resonant tank.

The 3rd image is of a normal ZVS, yellow is once again the resonant tank, and blue is of the gate. And again, referenced to the same area.

I'm guessing the "hitch" I'm looking at is ringing from the gate.

It seems to work as well as the standard ZVS. I've tested it with a flyback transformer and it works well, though the only thing is that the FET's seem to get a little toasty. Drawing an arc makes it draw over 10 amps while only inputting about 10-12 volts, and because 10 amps is the limit for my supply... well I can't test it further! An alternative idea I had, instead of using another resistor to draw the gate back to low maybe swap in a PNP transistor to pull the gate to ground when it needs to be off?

Does anyone have any suggestions, thoughts or comments on the design of this? I'd be happy to take any criticism of this.

Sigurthr

Sat Jan 09 2016, 11:28PM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

I've used parallel LC dual-choke ZVS builds before quite successfully. I actually prefer them over the traditional center-tapped version.

I don't think it is ringing, but rather a bit of overshoot. It's possible the zeners you are using are of low tolerance, or are just too slow. You might consider trying TVS style diode pairing with zeners, similar to like what we use in SSTCs to better protect the gates. Just watch the current levels through the zeners unless you have high wattage ones.

It doesn't look like you used the same tank configuration for both, but even so looking at rise/fall time on the gate as a ratio of rise/fall:cycle it looks considerably slower than the traditional zvs. I would postulate that the gate resistance values are too high, and likely the gate-source resistors are too low.

Also regarding the fet heating I think quite possibly your input chokes are too low impedance. Remember that since this tank resonates so low frequency the chokes present less of an impedance, so there will be more current flowing through the fets.

Here you can see Marko's schematic of this type of ZVS circuit, he published this years and years ago:

By the way, it's good to see you active again. I haven't seen ya on here or LPF in ages. =)

DerAlbi

Sun Jan 10 2016, 12:54AM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

It could also be caused by bad measurement. Make sure when you probe a mosfet gate you connect the ground lead directly to the soruce pin of the same mofet and not to some ground somewhere in the circuit. Also make sure your layout avoids the same problem with the zeners. do not allow any inductive lines at the ground. Try to analyse your ground signal integrity with the Oscilloscope... just hook it up to ground and some other ground across the PCB (if you have one). If you find any signal... bad sign.

Inducktion

Sun Jan 10 2016, 01:31AM

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

I did some more playing around with the circuit; It seems like I can seriously reduce the current through the system by adding in two PNP transistors across the gate-source resistor and have their base's connected to the opposite mosfet drain. Like it goes from 1.5 amp draw to 0.8 amp... but then the gate drive voltage drops significantly, to 3-4 volts.

Also reducing the resistance as you suggested Sig for the gate resistor helped a lot i think.

Also I think they might be getting so hot because I was using such small heatsinks. I forgot how much power these types of drivers can draw (read somewhere that it can draw over 25 amps at 36 volts for a normal ZVS driver)...

I stopped playing around with it for now because it melted my breadboard holes. Lol.

I'll probably remake it on an actual protoboard and continue to play around with values and things.

Oh, and I also just realized I only had 4 turns of wire on the flyback. Most normal ones have 4 + 4 center tapped. Woops. That probably wasn't helping either. Granted comparing what other people got out of their flyback at 12 volts... I was getting the type of arcs you'd expect from 16 volts I think!

Sigurthr

Sun Jan 10 2016, 03:09AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Haha, melted breadboards... been there!

Glad the resistor change helped.

Yeah the nice thing, other than not needing a center tap, is of course that you need less turns for the same coupling/current/ratio.

I stopped making my own zvs drivers when eBay Chinese ones dropped in price to the point that simply ordering tank capacitors became more expensive than a finished driver. Just watch out if you keep using small heatsinks, remember zvs drivers are a shrapnel hazard!

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