Triple Flyback Stack advice wanted / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Triple Flyback Stack advice wanted

1 2 next

Move Thread

LAN_403

Kiwihvguy

Tue Sept 08 2015, 01:21AM

Registered Member #3395 Joined: Thu Nov 04 2010, 08:42AM
Location: Christchurch, New Zealand
Posts: 193

Hi all,

I have made several posts about troubleshooting my ZVS Flyback driver which I have made several threads about in the last few years or so and I'm aware the ZVS is very popular albeit can be unreliable. I think this is because the simple design of the circuit only controls a few variables which influence the circuit's performance, and a lot of other variables are left uncontrolled which can cause confusion for people trying to troubleshoot it.

However, I have had more reliable success having rebuilt it for the 3rd time. Note by success I mean running the circuit multiple times for longer durations without failure. I consulted my electronics mentor, who is a power engineer working on the transmission lines in New Zealand, and he gave me perhaps the tip that changed everything. As I have been consistently using SLAs to power my ZVS, I had also been having multiple failures immediately upon switch-on. I am now using an inline resistor when switching on initially to limit any inrush current, which is then shorted out to allow full operating current to flow. So far I have had no failures at switch on using IRFP250 MOSFETs.

I also placed the resonant capacitor closer to the flyback's primary as per one suggestion I saw on Youtube, took care not to run the IRFP250s over a supply voltage of 36V (48v is more likely over 50V when SLAs fully charged, 4 x 50V = 200V, which exceeds IRFP250's 200V limit, as the circuit diagram mentions)

In order to be able to run it safely on 48V, I replaced the 250s with HGTG30N60B3D IGBTs (600V 60A, very fast) as I needed a higher voltage rating as well as being fast (also tried GT60N321 - 1000V 60A -, far too slow and heated up considerably at 12V)
So far works reliably at 48V, IGBTs are mounted on a large heatsink with good quality thermal grease and are force air cooled. They don't get warm after 2 minutes arcing operation, but needs to be run for a duration of 10-15 minutes to test the reliability.

I intend to separate the gate drive and flyback power supply to increase the flyback voltage without having to modify the gate driving circuit. Now that my circuit is so far proving to be reliable enough, I want to increase the HV side current output. I have thought about running 3 flybacks in parallel (both primaries and secondaries in parallel).

Can I please have some advice about the following points:

- phasing of the primary windings i.e: do the 3 separate primary windings have to turn in the same direction or do some have to be wound in the opposite direction?

- if the primaries are wound in parallel, then the effective inductance will be roughly 1/3, so the capacitance needs to be tripled in order to maintain the resonating frequency? I may parallel some IGBTs to increase the current capability

- although the ZVS self-resonates depending on the inductance and capacitance of the tank circuit, will a flyback perform better and have a higher power output (voltage and/or current) if the capacitance is changed so that the ZVS resonates at the frequency the flyback was designed to run at?

Many regards and thanks in advance,
Cheers

EDIT: I have only learnt about series resonance at high school at the moment, but I'm aware parallel resonances operates differently.

hen918

Tue Sept 08 2015, 06:28PM

Registered Member #11591 Joined: Wed Mar 20 2013, 08:20PM
Location: UK
Posts: 556

With regard to your first two points:
If the secondaries are connected "nose to tail" then the primaries will have to be to maintain the right phasing. Doing this, however, will mean that the insulation between the primary and secondary of the flyback will have to withstand the total voltage of the stack (30kV+). This will probably be a problem.

Yes, to keep the frequency the same, the capacitance will have to be changed to make up for the drop in inductance, however this will affect more than just frequency: the quality factor of the coil will change dramatically, and with the finicky ZVS driver, anything could happen... It may change reliability, efficiency, output power, whether it actually oscillates or not, etc...
Driving flybacks, changing values like that don't tend to cause catastrophic failures as often as with induction heaters or wireless power transmission (which can both be used with a standard ZVS driver (I tried and killed too many MOSFETs)) but Royer / Mazilli ZVS still tend to be a little unpredictable as well as unreliable (in my experience).

Kiwihvguy

Tue Sept 08 2015, 08:38PM

Registered Member #3395 Joined: Thu Nov 04 2010, 08:42AM
Location: Christchurch, New Zealand
Posts: 193

hen918 wrote ...

With regard to your first two points:
If the secondaries are connected "nose to tail" then the primaries will have to be to maintain the right phasing. Doing this, however, will mean that the insulation between the primary and secondary of the flyback will have to withstand the total voltage of the stack (30kV+). This will probably be a problem.

I'm not quite sure what you mean by "nose to tail", but I intend to parallel the high voltage outputs and not connect them in series. Nose to tail sounds like connecting the flyback's HV output lead to the ground/cold side of another flyback?

I think you might have missed a word when mentioning the phasing, you didn't specify what direction

dexter

Tue Sept 08 2015, 10:43PM

Registered Member #42796 Joined: Mon Jan 13 2014, 06:34PM
Location:
Posts: 195

to bad you need 3 flybacks...
because one primary coil can be used with 2 flybacks

Kiwihvguy

Tue Sept 08 2015, 10:48PM

Registered Member #3395 Joined: Thu Nov 04 2010, 08:42AM
Location: Christchurch, New Zealand
Posts: 193

dexter, that's a good point though. It's just I had 3 similar flybacks and I'd like to get them all going in one setup. So this will mean I wind a 5 + 5 primary on each flyback, and make sure they are in parallel and are wound in the same direction?

dexter

Wed Sept 09 2015, 05:24AM

Registered Member #42796 Joined: Mon Jan 13 2014, 06:34PM
Location:
Posts: 195

Kiwihvguy wrote ...

dexter, that's a good point though. It's just I had 3 similar flybacks and I'd like to get them all going in one setup. So this will mean I wind a 5 + 5 primary on each flyback, and make sure they are in parallel and are wound in the same direction?

no
for 2 flybacks you just put the cores together and wind one single 5+5 primary on both
like this:

Ash Small

Wed Sept 09 2015, 09:35AM

Registered Member #3414 Joined: Sun Nov 14 2010, 05:05PM
Location: UK
Posts: 4245

dexter wrote ...

for 2 flybacks you just put the cores together and wind one single 5+5 primary on both
like this:

Should work with three as well, it certainly works with four.

Why wouldn't the same principle work with three?

dexter

Wed Sept 09 2015, 12:15PM

Registered Member #42796 Joined: Mon Jan 13 2014, 06:34PM
Location:
Posts: 195

Ash Small wrote ...

Should work with three as well, it certainly works with four.

Why wouldn't the same principle work with three?

with 3 the space inside the coil is not completely used
how much impact this would have i don't know

hen918

Wed Sept 09 2015, 05:43PM

Registered Member #11591 Joined: Wed Mar 20 2013, 08:20PM
Location: UK
Posts: 556

Kiwihvguy wrote ...

I meant series as you thought, as this isn't a problem. As for the phasing, yes, the primary has to be wound so that all of the flybacks are the same. It might be easier to wire the primaries in series. which shouldn't present a problem with the ZVS driver.

Kiwihvguy

Thu Sept 10 2015, 08:22PM

Registered Member #3395 Joined: Thu Nov 04 2010, 08:42AM
Location: Christchurch, New Zealand
Posts: 193

Hey guys,

Have managed to construct and get the triple flyback stack running so far. I had to make a few modifications such as paralleling 2 IGBTs in place of one, increasing the size of the resonant capacitor to handle the greater resonant currents etc.

It draws 26 amps at 36v from the SLAs, the arc is noticeably fatter and hotter than my single flyback setup before.

One concern I have before I am confident to test it at higher voltages for longer durations, I am worried that if I start drawing a lot of current, the inductors may saturate and cease to act as inductors, which wouldn't be too good. I have a feeling I might need to put some more in series and parallel to increase the current capability.

Does anyone have any advice or formulae I can use to ensure my inductors are big enough?

1 2 next

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