A big-ish diy xformer

[oneKone]

For a while now I've been absent from the forum, the ease of FB has driven my lazyness not to post (most importantly to share).

I've been working on creating a reliable secondary for a while now. I pursued 3d printed bobbins for the longest of times, but unfortunately they're not reliable enough as there was always a failure mode of the secondary arcing through the print. To beat this I started to make the bobbins from acetal with almost flawless results (so far).

Some specs;
Core: UF86A

Primary: 6+2+2+2 turns of litz wire made from 35 strands 0.4mm wire, finally wrapped in Teflon tape,

Secondary: the bobbin is machined from acetal/delrin, OD: 70mm, ID: 31.7mm, length 60mm, segment depth is 9.5mm/2mm wide with I think* 2.5mm walls. Each wound segment has 500 turns of 0.14mm wire. The "free" segments are simply for the wire to transition from the top of one layer to be perfectly ath the bottom of the next. 

The enclosure is made from PVC pipe with an acrylic top plate, brass M6 studs/nuts, 3d printed core support and PVC electrodes with acrylic caps/press fit brass terminals.

/>
The driver itself has been another can of worms... The occasional brick failure is annoying, but solving the problem is a not so bad challenge. Atm it's running a standard TL494 with 35% duty cycle, A3120 gate drives (each with an isolated 18v supply) driving a skm200gb. I've recently added over current set to 180a primary current, in an over current event it draws the driver down to 10% duty cycle and needs a power cycle to reset. I'm currently working on a new series inductor and a OCD that doesn't need to be power cycled.

After I get the driver sorted I'll be starting on a few other sizes high voltage transformers, just to see what I can come up with.

Cheers

[davekni]

The "free" segments are simply for the wire to transition from the top of one layer to be perfectly ath the bottom of the next.

Interesting solution!  I'd thought about slanted gaps between segments, but don't have any method to create such.  Seems like it would require a milling machine with a rotational axis for the part holder.  If slanted gaps could be created, would allow use of all segments.

[oneKone]

We have the equipment at work for the manual mill, but the setup time would be a killer to get it done before work/lunch break. Or could even be someill looking into on the CNC router at home 🤔

[Mads Barnkob]

Very nice project, good looking final enclosure as well! Not just that regular oil filled IKEA container from the kitchen department

Which tapping are you running it at on the primary coil? Since you say it can pop bricks and limited to 180A?

Your driver reset, driving down to 10% duty cycle is fine, but power cycling to reset, why not just do something like in the DRSSTC drivers, that it just resets on the next interrupter pulse? (could just be a 555 timer giving 100 BPS reset pulses, as there is no interrupter here)

Your inverter box needs some safer terminals

[FonziDaytona]

I’m pretty jealous of you guys with access/skills for machining…This thing would make an excellent base for a fusor power supply.

-Matt

[oneKone]

Very nice project, good looking final enclosure as well! Not just that regular oil filled IKEA container from the kitchen department

Which tapping are you running it at on the primary coil? Since you say it can pop bricks and limited to 180A?

Your driver reset, driving down to 10% duty cycle is fine, but power cycling to reset, why not just do something like in the DRSSTC drivers, that it just resets on the next interrupter pulse? (could just be a 555 timer giving 100 BPS reset pulses, as there is no interrupter here)

Your inverter box needs some safer terminals

Cheers,
Originally I had made a fibreglass enclosure, but I wasn't completely happy with the outcome (more fibreglass practice is needed 😅)

For the video it's running on 8 primary turns. Atm for the OCD, series inductor and enclosure. I'm treating it all as the first revision. I do plan on making a more robustly housed driver that's also somewhat safer from wandering fingers.

[oneKone]

I’m pretty jealous of you guys with access/skills for machining…This thing would make an excellent base for a fusor power supply.

-Matt

These bobbins would be something easily made on a small Chinese lathe, but there's also the upfront cost of a lathe and tooling, sometimes it's also the luck of knowing a machinist 😂. I had thought about selling bobbins, but shipping from Australia is just too expensive for what it is.

[FonziDaytona]

Can’t be that much to ship…I’d be interested. I just smoked by transformer and looking for an alternative. 

[abstruse1]

Can you explain why you left every other segment of the bobbin empty?

Have you tried other alternatives to Litz wire for the primary?  I played with Litz primaries once but didn't have adequate wire and melted the wire.  I also compared the typical fine stranded 12 and 14AWG to solid magnet wire but couldn't tell much/any difference, so I'm using the solid wire.  Heresay suggests that Litz would have some benefit for the primary.

[oneKone]

@abstruse1

I was having problems with the winding not starting at the very bottom of the new segment, it's just a simple (but wasteful) transition. As for the litz wire, I did try some standard car audio power cable, it it quickly melted. I think it was also slightly larger diameter (conductor) than the litz wire I wound.