Winding small flyback transformers / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

Winding small flyback transformers

Move Thread

LAN_403

The Wumpus

Tue May 02 2006, 12:01AM

Registered Member #325 Joined: Fri Mar 17 2006, 12:42AM
Location: Turku, Finland
Posts: 55

Hi,

I've lately been constructing some small flyback transformers based on scavenged ferrite cores and awg #38 (about 0.1 mm) magnet wire secondaries with 500-1000 turns. The first goal is to get a reliable working ~12kV transformer for charging a special purpose photo flash capacitor. It should be powered from pulsed 12V driven by power mosfet (IRF730) and provide at least a few mA.

This has first and foremost been an exercise in frustration--and learning that corona and arcing between windings and to the core WILL occur unless one is meticulous in the construction process. Potting with some kind of compound seems to be absolutely mandatory for any kind of "long term reliability" (read: doesn't stop working after running for 5 minutes).

Here's how I did it:

5x6cm ferrite core from broken computer PSU. I'm winding 5 layers of ~150 turns per layer. The layers are insulated with pieces of recycled OH sheet (mylar I think). A battery powered drill which runs at very low RPM takes some of the tedium out of the winding (still takes a good hour to do it properly though).

Blobs of hot melt glue is used to fix the high voltage and ground terminal wires as well as to provide a first layer of insulation between the layers and the core.

First test with the primary in place (20 turns of 0.5mm magnet wire)--it's working. To detect possible corona losses I run it unloaded for a minute and then disconnect the power and take a good sniff at the transformer, any produced ozone is likely to be noticed (more scientific method wanted

)

This is a smaller transformer I made earlier and potted in silicone. It is currently being used as a trigger transformer for a spark gap. Very ugly looking but seems to do the job. Even though I am using these strictly for DC, I do not pot the rectifier diode (I'm using BY 8114 type) because I want to be able to replace it if (when) it burns out. It also adds the possibility of adding CW multipler stages to the output.

Any comments? I'm aware that this technique is crude and that these transformers will break down sooner or later (probably sooner). I'm mainly looking for tips to guarantuee long life--it can be frustrating to spend hours on a transformer and then have it break down quickly. Any suggestions on potting compound, some kind of epoxy maybe? I've used silicone (the bathroom type) and it seems to do its job as an insulator, but it is difficult to the air bubbles out of it.

The flybacks I built so far (5) all seems to work best at rather low frequencies (10kHz) or less. To lessen the noise I'd like to run the at higher frequencies (like TV/monitor flyback), but at these frequencies they just draw a lot of current and provide little or no output power. This could possibly be a problem with my driver (I'm using a 555 driving a mosfet via a totem pole like here

). What's the best driver to use?

Also, anybody know of a good resource for the theory behind flyback transformer design, e.g. how turns ratio, core dimensions and material, air gap, etc. affect the operating frequency, output voltage and current?

Thanks in advance for any suggestions.

...

Tue May 02 2006, 01:57AM

Registered Member #56 Joined: Thu Feb 09 2006, 05:02AM
Location: Southern Califorina, USA
Posts: 2445

I have built some (larger) flybacks...
I was using a zvs to drive them, using a decent sized core I pulled out of a flyback (.5' in diameter, 2" long and 1" inside the U) and 32awg wire (because I had it handy). I had ~500turns, and was driving it with a 4+4 turn primary with up to 75v (I am too chicken to turn it up any higher than that, and the core waveform was starting to look pretty ugly) in giving about 10kv out that drew ~10a in (so I guess 70-80ma out) and was extremely bright, so I put on a level shifter with a 2.2nf cap and now it gives ~20kv (my best guess, it starts arking at about 2cm), and at 50v it can be stretched out to 6cm easy. I am scared to get anywhere near it at 75v in, judging by the fact that between the 2cm gap I got an ark that was about 8-10cm tall, that is like 20cm of spark...

Enough gloating, and some of the construction details...

I used a pair of flyback C cores, (the legs were .5" in diameter, it was 1.7" tall on the inside dimension, and .8" wide on the inside)the primary was wound around the leg that was square (the one that was outside of the flyback) 10awg stranded copper wire (gets hot under extended operation) and the secondary was made by...
1. cut a few 1.7"x11" strips of transparency and wrap one around a .5" former. Tape it to make a tube
2. cut 10 1.5" circles of transparency with a .5" hole in them (stack 10 pieces up and cut them all at once)
3. Use hot glue to attach them to make a former. Add the return wire inside the joint between the side/tube before gluing
4. wind ~100 turns of 32awg wire inside of it leaving about .3" on either side open
5. Add a nice thick layer of silicone (RTV) around the the winding
6. Wrap another strip of transparency around it, making sure to get out all air bubbles and to get an even layer of silicone
7. Wrap a layer of cellophane tape to keep it in place
8. Wind another layer like the first one, silicone, wrap, etc and keep going until you have enough layers. Try to stagger where the wires cross layers (at the ends of the coil) to make sure it does not ark over there.

The it would be nice to let it dry for a while (but not really necessary), and fire it up. At first I got a lot of corona between the secondary/primary (since the whole length of the secondary was in contact with the primary) I stuffed about 10 layers of transparency in between them (as many as I could fit). Now I can run 75v+ into it and produce 10kv+ out.

A little bigger than yours, but the technique should be pretty easy to scale down. Silicone is your friend, hot glue works pretty well too, but it doesn't seem to insulate as well.

I find corona on mind by listening to it (from a meter away), if I hear a loud hissing... Other than that I don't care

It probably has 10 minutes of run time on it and I haven't noticed any signs of damage.

I run mine around 100khz so that I can use the highest possible v/turn. The way I think is the higher v/turn the less turns I need--which is good in many ways. It reduce resistance losses, although they are pretty negligible wen running a mere 100ma through 32awg wire. It also reduces the required number of turns, so I need less layers, so the coil is more compact (a big plus with the restricted core I am using), so I can add more insulation.

I use the gaps that came with the core, they are like 5mil thick. I have found that the gap isn't critical, so long as you have it. My cores work fine (well they draw more current for a given output) with gaps as large as 3mm

I am going to venture to say that 12kv out of that small core (I am guessing that you have ~1cm of winding space on either side of the middle leg?), although if you are running at a mere 10khz insulation will be quite a bit easier. I would aim for 6kv out, and use a simple level shifter to get 12kv out. The cap needed is very reasonable, only a few nf... You could also go for a full wave doubler and use 3kv, but I am going to venture to say you can get 6kv out...

Good luck!

edit:
wow, this thread is headed nowhere fast. Yes hot glue is pretty crappy, but it seems to work in a pinch where you need to stop something from arking over; and it dies a pretty good job of supressing corona. I have yet to find glue with any considerable conductivity, but your results will vary. I really hate oil as it can make a huge mess, and you can never get rid of it--but it does have it's purposes.

Part Scavenger

Tue May 02 2006, 02:51AM

Registered Member #79 Joined: Thu Feb 09 2006, 11:35AM
Location: Arkansas
Posts: 673

hot glue works pretty well too, but it doesn't seem to insulate as well.

Be careful using hot glue, I've had awful luck with it. I've gotten to the point where I consider it a conductor. In some cases, air is a better dielectric(really). Not to say ...'s wrong, he's an awfully smart dude, but I'd at least test it before I relied on it. I buy the cheapest hot glue I can find too, so that might be a factor.

Just my $.02.

Self Defenestrate

Tue May 02 2006, 03:14AM

Registered Member #87 Joined: Thu Feb 09 2006, 01:36PM
Location: San Jose
Posts: 191

Ditch the hot glue, and introduce it to the land of oil, where it can frollock free of worries that corona will eat its insides. Something I've done to soak oil into the windings of tiny coils is to drop it into a large syringe barrel, and draw up some oil. The air tends to get sucked out of the coil and replaced with whatever oil or potting compound you want. Another thing that tends to help it some polyurethane spray, give the core a few coats, should reduce some corona as well.

The Wumpus

Tue May 02 2006, 11:20PM

Registered Member #325 Joined: Fri Mar 17 2006, 12:42AM
Location: Turku, Finland
Posts: 55

... wrote ...

I used a pair of flyback C cores, (the legs were .5" in diameter, it was 1.7" tall on the inside dimension, and .8" wide on the inside)the primary was wound around the leg that was square (the one that was outside of the flyback) 10awg stranded copper wire (gets hot under extended operation) and the secondary was made by...
1. cut a few 1.7"x11" strips of transparency and wrap one around a .5" former. Tape it to make a tube
2. cut 10 1.5" circles of transparency with a .5" hole in them (stack 10 pieces up and cut them all at once)
3. Use hot glue to attach them to make a former. Add the return wire inside the joint between the side/tube before gluing
4. wind ~100 turns of 32awg wire inside of it leaving about .3" on either side open
5. Add a nice thick layer of silicone (RTV) around the the winding
6. Wrap another strip of transparency around it, making sure to get out all air bubbles and to get an even layer of silicone
7. Wrap a layer of cellophane tape to keep it in place
8. Wind another layer like the first one, silicone, wrap, etc and keep going until you have enough layers. Try to stagger where the wires cross layers (at the ends of the coil) to make sure it does not ark over there.

Thanks for the detailed description, sounds very similar to the procedure I used, though I find it eaiser to use thinner wire such as 38 awg so you can fit more turns per layer. Even 40 awg is rather easy to wind when you get the hang of it (a magnifiying glass/lupe definately helps here).

... wrote ...

I am going to venture to say that 12kv out of that small core (I am guessing that you have ~1cm of winding space on either side of the middle leg?), although if you are running at a mere 10khz insulation will be quite a bit easier. I would aim for 6kv out, and use a simple level shifter to get 12kv out. The cap needed is very reasonable, only a few nf... You could also go for a full wave doubler and use 3kv, but I am going to venture to say you can get 6kv out...

Good luck!

I currently get about 1 cm DC arcs between pointy electrodes, so I think the the voltage is ok but not quite there. It draws 6-8 amps from a SLA battery at 12V when arcing. I will try adding a level shifter and adjust the duty cycle to get the right voltage.

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