Help wiring a flyback transformer.

[Hiemal]

Good luck! If you need any help I'll be glad to give it.

 

[Atomicrox]

Thanks! I found a nice circuit simulator and tried to simulate this but I couldn't figure out how to add the IRF540 and the flyback properly. Take a look, it's pretty cool.

 

[Hiemal]

Circuit Simulator Applet

You just gotta change the windings ratio on the transformer.

And the mosfet is generic. It doesn't really have anything you need to change.

 

[Atomicrox]

It worked, I'll post a video tomorrow!
Thanks a lot man, this driver is indeed way better than the one with TV parts.

The only thing wrong is a flyback pin that sometimes shoots an arc (seemingly at the ferrite core). I didn't run it much because I was afraid that could fry the flyback.

Some sites advice sealing it with hot glue, what do you think?

 

[Hiemal]

That's odd. Are you using an older flyback transformer? It should only be at HV potential at one end, since it's supposed to be rectified.

You may have accidentally killed the diode inside of it though. Try measuring the secondary with a multimeter, to see if you get any readings on the resistance setting. If nothing, then the diode is still there. If something, your diode is shot, and you now own an AC flyback transformer.

Glad it worked a lot better though!

And a more ...suited form of HV potting would be to submerge it in some sort of veggie oil, or mineral oil.

Mineral oil is better because it doesn't go rancid.

 

[Atomicrox]

I don't think it's very old, it's from a CRT monitor and looks a lot like OP.

I got a reading of 1900 k but it vanished quickly, like it had a resistance but was above the DMM's limit (2000k).

I'd like to tune the amount of primary windings before submerging it in oil. Do you think it's gonna die if I keep running it? The MOSFET isn't even getting warm so I'm guessing the current is low.

 

[Hiemal]

I don't think it's very old, it's from a CRT monitor and looks a lot like OP.

I got a reading of 1900 k but it vanished quickly, like it had a resistance but was above the DMM's limit (2000k).

I'd like to tune the amount of primary windings before submerging it in oil. Do you think it's gonna die if I keep running it? The MOSFET isn't even getting warm so I'm guessing the current is low.

What do you mean "tune" the primary windings? Are you trying to find the resonant frequency?

What do the arcs look like, thin and purple, or fat and yellow?

Thin and purple are HV low current arcs. They'll bite you but won't kill you unless you have some sort of reason for them to (weak heart, wet skin, broken skin etc)

The fat and yellow arcs are much more dangerous and must be handled with care. A bite from one of those can be easily lethal, as the current can be anywhere from 20 mA to 100 mA, at 10-30 kV.

If your mosfet isn't getting warm at all, then I'm going to guess you have purple arcs. If you wanna make your arcs a bit ... better, decrease the number of primary windings a little. It decreases the winding ratio, and also allows more current to flow. Your mosfet will also consequently get hotter, so be careful!

I wouldn't worry too much about it dying if you keep running it. I'd just stick it in oil for good measure though.

 

[Sigurthr]

From what pin is the arcing occuring? Perhaps it is just a teriary medium voltage winding arcing over to the ground pin?

Also, glad you got it working!

If you did kill the diode don't feel bad, it is an awesome thing. Los of stuff you can do with an AC flyback, LOTS.

 

[Atomicrox]

What do you mean "tune" the primary windings? Are you trying to find the resonant frequency?

What do the arcs look like, thin and purple, or fat and yellow?

Thin and purple are HV low current arcs. They'll bite you but won't kill you unless you have some sort of reason for them to (weak heart, wet skin, broken skin etc)

The fat and yellow arcs are much more dangerous and must be handled with care. A bite from one of those can be easily lethal, as the current can be anywhere from 20 mA to 100 mA, at 10-30 kV.

If your mosfet isn't getting warm at all, then I'm going to guess you have purple arcs. If you wanna make your arcs a bit ... better, decrease the number of primary windings a little. It decreases the winding ratio, and also allows more current to flow. Your mosfet will also consequently get hotter, so be careful!

I wouldn't worry too much about it dying if you keep running it. I'd just stick it in oil for good measure though.

The arcs are thin and purple, about 2cm long.

Sorry, I didn't use the best word. I meant decreasing the amount of windings until I achieve a good balance between mosfet heat and nice arcs.
I used 20 turns, will try 15 next.

From what pin is the arcing occuring? Perhaps it is just a teriary medium voltage winding arcing over to the ground pin?

Also, glad you got it working!

If you did kill the diode don't feel bad, it is an awesome thing. Los of stuff you can do with an AC flyback, LOTS.

Take a look at the picture, the pin the arcs come from seems to have been cut at the factory. I tried to draw the path of the arc in red.
Note that it doesn't arc continuously, just shoots some arcs.

Is there any way to see if I'm getting AC output without frying my multimeter?



Edit: I just got bitten by it and it hasn't been turned on since yesterday, LOL! Is there a capacitor inside the flyback?

 

[Sigurthr]

The diode acts as a capacitor when negativel biased, so that is how it holds a charge - this is an indicator that the diode is fine.

The standard test for a flyback to determine if AC or DC is to try to charge a capacitor with it (it has to be rated for the full voltage of the flyback - not easy to find). If it charges it is DC, if it doesn't it is AC.

Based on your picture I would say my assessment is correct; a medium voltage pinout is arcing. Just pot it with high temp hot glue, make sure the glue gun is very hot because seams where cold glue meets hot glue are not as electrically insulative as a single "pour", so make sure you can fully pot the area in one go.

I had to pot one of my flybacks as well, there were two center-tapped/shared medium voltage windings which were floating at about 6kV each referenced to center.

Oh, btw, your primary has WAY too many turns on it. Less primary turns = less reactance = more current draw AND higher step-up ratio = higher output voltage. Do watch the heat of your transistors though as you are reducing the primary turns... the difference between too hot and just right can be as little as half a turn.

 

[Atomicrox]

I managed to pot that pin with hot glue and it stopped arcing.
At complete darkness I can see another pin arcing with the HV negative but it's a very faint arc. There's also some corona coming out of that white wire (the unused one that came with the flyback).
I'm thinking about cutting all pins, gluing everything and sticking a blob of glue on both unused wires. What do you think? I'd like to avoid sinking it in oil if possible

There's also another issue - I decreased to 15 windings and the PSU will shut down on all but the lowest potentiometer setting as soon as the arcs start. It's a Zalman PSU so it might be triggering the short-circuit protection. Any tricks to bypass that?

Last but not least - the videos!

 

[Sigurthr]

Okay, first off, a BIG no-no in the second video. NEVER EVER grab the high voltage lead with your hands or metallic objects not insulated and rated for >50kV. Right now your arcs are low current but that doesn't mean it can't injure/kill you on a bad day. ALWAYS bring the HV return (ground/negative) to the HV lead, never the other way around.

Secondly those arcs are still somewhat low current but I'm thinking you've got current sourcing issues more than anything else at this point. What are you using for a PSU? You said Zalman so I'll assume an ATX computer psu. Problem is, flyback drivers like high current at >10V and computer PSUs only deliver high current on the low voltage rails (3 or 5V). For example, my ZVS Royer IGBT Flyback driver likes 6 amps @ 12V minimum. My 555 driven single Mosfet Resonant driver will pull a little over 9A @ 12-20V (limiting factor is the Vgs of the Fet and Vmax of the 555) when drawing an arc. Your PSU's 12v rail likely won't source more than 3A before being too heavily loaded down or triggering a short circuit protection.

If you can source a small 12v SLA (or car battery) or two large 6v lantern cells which you can run in series (best would be four cells, two parallel pairs in series) and give it a shot you will likely have better output.

 

[Atomicrox]

I'll be more careful and use a stick next time, didn't really know I should move the negative and not the other way around. Lots of people are doing it wrong on YouTube :/

Yeah it's an ATX PSU (Le Quack's suggestion) - Zalman ZM460B-APS - according to the manual the maximum current is 16A/18A and peak is 20A/22A on +12V1/+12V2 respectively. There's no diagram explaining which pin is which but either way I don't think it's pulling over 16A - the mosfet would get really hot and it only gets warm.

I've tried an electric fence SLA battery but the poor thing is old and barely holds a charge anymore. I can get those big lantern cells but they're really expensive and I don't think they'll last long.
Other power source ideas are welcome! Unfortunately here this kind of stuff is expensive and hard to come by :/

 

[Sigurthr]

Wow, 16A on 12 rail, eh? That's pretty awesome. I haven't built PCs in about 7 years so... haha.

Make sure you're properly loading the lower voltage rails, they need to see some load otherwise the psu won't be fully turned on. There is a guide on how to do it somewhere in the non-lasers sections of this site.

 

[Atomicrox]

Yeah computers got power hungry XD

Oops, I didn't know I had to do that. The only thing I did was connect the green pin to ground to make it "boot up".
I'll try to find the thread, thanks!

 

[BShanahan14rulz]

I keep on hearing that too, but what does that mean? I've only had to pull 6A @ 12V, but if it can supply that without actually being "on" then why bother loading the other rails?