adventures in homemade flyback bobbins.

[oneKone]

Having no option but to put my drsstc build on hold i still wanted to work on a high voltage project, but remain cheap at the same time. I've had a few flyback cores sitting for a long time so i decided to make use of one of them. i was looking online when i saw the internals of a flyback (how it was wound) so i thought i'd give it a go.

for a driver i'm using mads circuit except rather than using max4420 i'm using a ucc37325 and 2 pairs of irf9540/540 to handle the switching. The igbt of choice is what i had laying around..... a cm300 

 the OD of the bobbin is 40mm with a winding depth of 7.5mm, this should give 225 turns per section but it was wound by spinning a drill and getting the wire to go from left to right within the section. So as a pure guess i'm going to say the amount of turns would be in the 1500 range. The first bobbin i tried didn't have enough of a wall between the core and the windings as well as i didn't give the oil long enough time to penetrate, this resulted in a very short life. The next bobbin wall thickness was increased by 2.5mm and the oil was given 8hrs to penetrate. this is the current bobbin i am running now. The oil i'm using is synthetic engine oil.

I think the next step would be to get a cheap vacuum pump and epoxy the next bobbin rather than oil. 

/>

[Mads Barnkob]

That is one heck of a killer-inverter for a small high voltage transformer!

Is the bobbin 3D printed, kind of looks like it. If it is, isn't a part of the failure then that the material is not solid, but full of holes and air inside?

If you want to whole-cast a transformer in epoxy, be sure to mix it quite thin so that it can get in-between everything a little easier.

[the_anomaly]

Nice! I wound a few of my own as well but chose a different approach for bobbin:





I started with a piece of PVC pipe.  I wound a layer ~180turns of 26awg wire and then wrapped an insulating sheet around the layer before winding the next layer.  Each layer added a larger radius.  IIRC I managed ~12 layers before calling it done. 

Wish I had pics of them running.  I made two, one produced something around 14KV which I fed into a Cockcroft–Walton multiplier and the other about 20KV with no multiplier.

I put them under mineral oil and used a vacuum pump to remove the air.  10 minutes under 500 microns and I was able to run them for hours with no problems.

If you decide to epoxy pot, you might need to include ultrasound to help shake the bubbles out.  I like https://www.epoxies.com/products/potting-encapsulating-and-casting/ for potting compounds.  Choose as low a cP as possible; above a certain point, ultrasound is required in addition to a vacuum.

[oneKone]

@Mads,
haha yeah... its extremely overkill, i should have some smaller igbts waiting for me at work. The bobbin is 3d printed but the infill was 100% so it's "solid". When i make my newer ones i'll machine the bobbin from acetal or teflon if i have some.

@The_anomaly,
That secondary is clean! did you wind left to right, insulate then wind right to left? Also thank you for the information about potting.

[the_anomaly]

Yes I did.  I kept the winding going as I wrapped a layer of insulation and when the previous winding was completely covered by a small margin, I continued winding in the other direction. 

Once you get a vacuum pump and desiccator jar, you will be impressed how much air it removes!

[oneKone]

after having luck with me second bobbin i thought i'd wind a smaller one using 0.14mm wire. It had 6 segments with a rough estimation of 2500 turns. After leaving it in warm oil overnight and gradually running hotter arcs through it (my choice of making sure any trapped air bubbles will rise to the surface) as a bed in process all seemed fine. I then moved the 2 electrodes further apart and that's when failure happened. On the third segment it arced over... My best guess is that using such thin wire with many turns had too high resistance, or that i was simply generating too high voltage per segment.


My second attempt i used 0.3mm wire with an approximate of 1200 turns. Having the same bed in process led to a working core.



at the moment i'm feeding  45v dc into the half bridge which then is connected to an isolation transformer with 5 turns primary side and 10 turns secondary side on a flyback transformer. this ups the voltage to about 60v (i think it may have been higher but can't remember what the voltage was under full load). I have also made a new driver that's still based on mads flyback driver except has 2 irf9540/540 output stage, the reasoning for this is that i had the parts laying around and i can driver pretty much whatever brick i want.


Knowing this method of winding flybacks actually works i'm going to order a cheap vacuum pump and try potting in some cheap resin and paraffin wax. If all fails with using cheap mediums i'll continue submerging the bobbins in oil and use the pump to accelerate the process. 

[oneKone]

After having a bit of a break from this project I stumbled on some cheap ferrite U cores from AliExpress. So experimentation is back on!

After looking at what teslaexplorer was doing with his diy flyback bobbins I had to give it a go! The new coil is just two cores side by side and an oval shaped bobbin. The total turns are still around the same estimate as my original bobbin. The driver is the same as before but now with a dual h-bridge of irfp460.. Anyways pictures speak loader than words!

[Mads Barnkob]

Great results for a compact HV transformer like that, home printed bobbin as well?

How much power is it consuming here? Some 2-3 kW?

[oneKone]

Great results for a compact HV transformer like that, home printed bobbin as well?

How much power is it consuming here? Some 2-3 kW?

Cheers,
Yeah still 3d printed. I had thought of machinists the bobbin from acetal or Teflon but it's just finding the time.

I don't have any official numbers but the variac hums around 5a so I think it's about 1-1.5kw. I think I'll be continuing making these bobbins (mainly larger). I understand not everyone has a 3d printer but it seems the only affordable way to make them.

I can also upload the 3d files if anyone wants them. They're extremely basic and have been drawn in tinkercad.

[profdc9]

I made a bobbin using a sheet of high density polyethylene a while back.  I should take a picture of it, but I would have to clean the oil off of it.

The sheet was about 6 mm thick.  I cut four 80 mm wide, 150 mm long pieces and made a rectangular long tube with the edges of the pieces glued to each other.  To adhere the polyethylene to itself, I used a butane torch to flame treat the polyethylene so that I could join the four sides together with epoxy.  Then I field off the corners of the rectangle to smooth them somewhat.  I was able to wind 500 turns of 0.25 mm magnet wire around it.  I had some large U-shaped ferrites which then were inserted into the bobbin.

It was one of the first things I did when getting started in high voltage and it was very difficult to get working figuring it out by myself.  After I was done I wrote a document on it.

https://drive.google.com/file/d/0BwFicJLV0O4jYlQ0ejJ6NE9XVE0/view

Dan

[oneKone]

Cheers for the link dan. I'll have to have a proper read of it, quickly looking over it seems to be extremely detailed.

[dexter]

This thread motivated me to finally make a HV transformer.
The core is the largest TV flyback core i had.
The coil form is made from 2mm thick acrylic glued on a 18mm PVC pipe.
Wound about 2000 turns of 0.25mm wire on six 2.5mm slots by 9.5mm thick and 16 turns primary with multiple taps.
Vacuumed it under oil for a few hours using 2 fridge motors.
Connected it to a full bridge of skm200gb123d bricks running at 30kHz and 130V

/>Here is running interrupted to give it mains humm

[Bambinz]

very very good
Have you tryed to drive it with a royer ZVS driver?   

[oneKone]

Nice work @dexter. I like your multi tap primary. That 50hz interrupter is an awesome idea, gives a very iconic sound

[Mads Barnkob]

Nice work @dexter. I like your multi tap primary. That 50hz interrupter is an awesome idea, gives a very iconic sound

Nice result and good use of sewer pipe for a oil container, I usually go scout out the kitchen department of Ikea for transformer containers

Its a nice idea, but 50 Hz is not giving the right sound is it? Since its not going positive and negative, but just on/off, maybe 100 Hz is a better emulation of full wave rectified, non-smoothed mains sound.

[dexter]

Its a nice idea, but 50 Hz is not giving the right sound is it? Since its not going positive and negative, but just on/off, maybe 100 Hz is a better emulation of full wave rectified, non-smoothed mains sound.

i used 100Hz because 50 didn't sound right

[Patrick]

what material did you use to print your bobbin ?

[ElectroXa]

Hi everybody 
I recently made a flyback secondary.
This secondary consists of 7 layers of 0.2mm wire, insulated by oiled paper, and there are around 1000 turns.
I tested the flyback with 20V @ 9 amps (zvs driver)
the secondary was hot to touch after 5 minutes of arcing.

[AlexanderHun]

Hello! Who is wondering how much voltage has been achieved with these solutions ?! Thanks!

[davekni]

Great information here - I'm just starting to experiment with DIY plasma balls.  The highest I've achieved with home-wound transformers is +-11KV at 30kHz.

Since your coil form is 3D printed, could you angle the slots between sections so that the wire transitions from the outside of one section to the inside of the next without resting directly against the wire of either section?  A steep (close to circumferential) angle would be ideal.  Were the arc-over failures at a point where the end (top layer) of wire for a section was adjacent where the wire came into that section from the previous?  That would be my guess - where there's the maximum voltage across the thin wire enamel insulation.