Wifi SSTC attempt 1...

[ldm314]

Been doing lots of reading, watching videos, etc. These inductive heaters seem to be nice circuits that resonate a significant amount of power. I've done lots of things with micro controllers, and even had one running an old ignition coil in my car with an igbt.

In all my reading I've only seen a couple examples of Tesla coils using a zvs circuit like this. Most used it to drive a flyback as an input to an sgtc. Here I'm going to attempt to use the zvs and build a secondary that matches the frequency it runs at. Controlled with an esp-01 microcontroller that has wifi built in.

 

[ldm314]

Step on, make it resonate in a more useful frequency range by changing the caps around. Went from 100khz to 225khz. My 'toy' pocket scope is struggling to keep up at this speed.

[ldm314]

Today the replacement IGBTs have arrived. The cheap inductive heater circuit came with IRFP260N mosfets. These ones are rated 200V, 50A. This will limit the input voltage to around 60 volts.

The new IGBTs are IRGP4063D. The specs look promising to me and they were only a couple bucks each on ebay. Rated at 600V, 96A continuous they should be able to directly replace the mosfets and allow me to run off rectified mains.

[Mads Barnkob]

I am not completely sure how you are going to put all these features together, but here are some pit falls you should investigate in your experiments

The Royer oscillator is a self-resonating circuit, you do not have any control over it as such, its really one of those crash-and-burn circuits when pushed either too little or too hard.

A fixed frequency primary circuit is not a great way to produce sparks, as sparks load the secondary LC circuit, the resonant frequency of the secondary changes,

I assume you are talking about 120VAC mains, but even here you will have around 540V across the switch, add switching spikes to that and you are running at the edge of failure. Offline ZVS/Royer oscillators is something many people have tried to make and many has also failed. Larger gate charges and slower high voltage switches simply makes it hard to make a offline version of it.

What were you going to use the microcontroller and wifi for?

[ldm314]

Thanks for the great advice. I'm not hugely skilled in analog electronics, mostly digital side things. The idea in the end is to run on rectified mains, slowly stepping up to that. Currently testing with a 12v supply that is barely enough for it. I have a 42V 27amp supply to run it from after I get things sorted, and I have a transformer here that has 60 volt and 90 volt windings from the 120mains.

My thought is to use the microcontroller/wifi to act as the interrupter. Tuning would be like an old SGTC with a primary tap. I am hoping to be able to trigger the Royer for short bursts with the microcontroller and some extra IGBTs I have from another project that are too slow for the oscillator but may be fast enough for this.

In the end, if it doesn't work I'll have all these components from the royer to work with and build a proven design.

[ldm314]

So far things are looking promising. Attached is the waveform from the primary coil and the connections to the microcontroller.

Microcontroller is powered by 5v usb charger, regulated to 3.3v. GPIO output is at 3.3V  goes into a 2n3904, which goes to a 2n2907 connected to the 5v usb power. The output of the 2n2907 runs the IGBT I had lying around. The IGBT is an ISL9V5036P3 that is rated up to 390V, 46A continuous and is switched with 5 volts. I had this IGBT from an automotive ignition controller project I did some time ago.

This scope capture was with an output pulse of 80us from the microcontroller. I can generate pulses down to about 8-10us with this microcontroller.

Currently waiting until I get paid at the end of the month so I can get some magnet wire and build the secondary.

[ldm314]

Also looking for advice on capacitors.  If this idea doesn't work I want to reuse these in a MMC. They are the 'MKPH' capacitors found in many induction cookers. The specs seem like they would be ok, but the dirt cheap price of them makes me have second thoughts. This link here was the only specs I could find.

http://www.cgegd.com/en/displayproduct.html?proID=1223865

[ldm314]

I've decided to test the capabilities of this microcontroller. It can output pulses of custom duration, the input is just an array of delay in microseconds.

It worked better than I thought. I set it up to repeat 2us on, 20us off and the result was the following image. I did add some resistors to make sure the transistors turn off all the way. Not sure how useful this may be, but at minimum it will be a nice programmable interrupter.

[Mads Barnkob]

Also looking for advice on capacitors.  If this idea doesn't work I want to reuse these in a MMC. They are the 'MKPH' capacitors found in many induction cookers. The specs seem like they would be ok, but the dirt cheap price of them makes me have second thoughts. This link here was the only specs I could find.

http://www.cgegd.com/en/displayproduct.html?proID=1223865

I would not trust them to be anything superior when there is no datasheet, but I think you can assume they will perform as MKP X2 capacitors.

I have some different capacitors in my MMC calculator here: http://kaizerpowerelectronics.dk/calculators/mmc-calculator/ and also a list of some more capacitors here: http://kaizerpowerelectronics.dk/theory/good-mmc-capacitors/

[ldm314]

I would not trust them to be anything superior when there is no datasheet, but I think you can assume they will perform as MKP X2 capacitors.

I have some different capacitors in my MMC calculator here: http://kaizerpowerelectronics.dk/calculators/mmc-calculator/ and also a list of some more capacitors here: http://kaizerpowerelectronics.dk/theory/good-mmc-capacitors/

These cheap chinese pieces definitely are hit or miss. Thanks for the list, I've seen the calculator before. Probably have read through your pages a dozen times.

This is my first try at putting together something like this, most of my projects before have been lower powered/digital circuits. At some point I am going to try and learn how to etch a pcb and make all that from scratch. Currently I am looking for a better scope since this pocket thing only goes up to 200khz, 1Msps.

My end goal for this first project isn't to make something that works. It would be nice if that happens, and I expect I'll get something, but my main goal here is to learn enough to make a good medium sized one in the future.

[ldm314]

One thing I thought of today. Now that I have this signal generator with my microcontroller, I can try winding my own GDTs and check it with the scope. I recall coming across someone using ethernet cable to wind a GDT. I've got a box of unshielded cat 5 sitting right here and a bunch of yellow iron cores I salvaged from PC power supplies.

[Justin]

One thing I thought of today. Now that I have this signal generator with my microcontroller, I can try winding my own GDTs and check it with the scope. I recall coming across someone using ethernet cable to wind a GDT. I've got a box of unshielded cat 5 sitting right here and a bunch of yellow iron cores I salvaged from PC power supplies.

http://thedatastream.4hv.org/gdt_index.htm

You don't want to use iron cores.

[ldm314]

http://thedatastream.4hv.org/gdt_index.htm

You don't want to use iron cores.

Thanks for that link, that would have been a hard to find mistake! I have the kind that it specifically says not to use.

[ldm314]

First attempt at trying to wind something. I didn't have a good core to wind on, but I had one that was "ok" according to the guide. Obviously didn't work but was interesting to try. This is the output of the transformer, using an EI33 core. 15 turn primary, 17 turn secondaries. Just spending time tinkering until more parts come.

I've gone through with javatc and simulated the secondary. Going with 1200 turns of 26awg on a 4.5" pvc. Around 20" tall total, with a toroid should resonate around 250khz. The primary will be 5" diameter, 2-3 turns of 1/4" copper tubing. Within that 2-3 turns I should be able to tune for a variety of top loads. And if the Royer doesn't work out, it will be great for a more tried and true design.

As a backup, I am trying to put together a schematic for a base feedback sstc.

edit: My first guess is that the transistor setup I have isn't right for this. I've got UCC27423 drivers arriving in the mail soon to try out.

[ldm314]

Today I worked on the power supply. I picked up a used Cisco power supply rated at 1300W output for only $17.  It has power factor correction to .99 and will stay regulated for 20ms after input power is removed. Should be good enough to power everything.

The outputs are rated:

• 15A @ +3.3 VDC
• 5A @ +5 VDC
• 6A @ +12 VDC
• 27A @ +42 VDC

[Mads Barnkob]

Very nice, the "27A @ +42 VDC" is perfect for a Royer, its around 1200 Watt and is properly also in that ball park that a standard Royer fed with 40VDC will be around, the higher voltage you feed it, the more current it will also try to consume, even pulling down the supply voltage from the heavy load.

So if you load this too heavily, it will properly go into some kind of over-current/load mode and shut down.

[ldm314]

It has a whole bunch of protection circuitry, EMI shielding, it is way overkill. At the cheap price though it was worth it just for parts. I've got it all put back together with some terminals for easy connection. The proper power cord came in for it and it is ready to go. All the voltages tested fine.

[ldm314]

More back up plan components have arrived. Gate driver IC, UCC27423 rated at 4A. 12V input. Photos are of the circuit, the low side drive coil output and the high side drive coil output. The signal from the microcontroller goes to INA, as well as a 2n3904. The inverted signal from the 2n3904 is sent to INB. OUTA and OUTB are connected in series with a 1uf capacitor to the primary.

The drive transformer is a salvage EI33 core from a power supply, with the enamel wire being unwound and reused. Primary is 15 turns, secondaries are 17 turns each wound in opposite directions. The resulting waveforms look like they are reasonable now.

[Mads Barnkob]

Good progress, if you want to put more work into the GDT waveform or it begins to act worse when loaded, you can use Richie Burnett's excellent guide on GDT waveforms, causes and fixes. http://www.richieburnett.co.uk/temp/gdt/gdt2.html

[ldm314]

That's a great guide, helped me identify what needs to be fixed.

For testing I took this old circuit apart that I blew a mosfet on earlier. It is one of the cheap "5-12v" royer induction heaters. I removed the feedback diodes and replaced the blown mosfets with similar ones. The blown ones were 30n06, now replaced with 70n06. The voltage droop on the top was amplified significantly when connected to the circuit. Based on the guide I either need to wind more on the primary or use a better core.


One thing I have noticed with the Royer is that the coil is fed positive voltage in on both sides, with the mosfets alternating the ground on the coil. SSTC designs I have looked at keep one end of the coil at the mid point and switch in either the positive or negative side. I am not sure what effect this will have, but I believe it will keep this arrangement from running if the input voltage is too high.

Edit: After reading some more it looks like the bottom part of an H bridge instead of the left or right. Top section is tied to positive with the two inductors to block ac from going back into the supply. The problem I can think of is having too large on time for either mosfet as the inductor will eventually allow too much current through.