Wifi SSTC attempt 1...

[ldm314]

Primary coil is formed and I will be testing it's frequency response on the Royer. I mainly want to determine the minimum number of turns it will still run at. JavaTC is showing that this should be within the range I need, I might have to add a couple more capacitors in series to keep from having too few turns.

I designed it with Sketchup and printed it on my 3d printer. It is a combination end cap and primary form, once the coupling is tuned I will be gluing it in place.

The design is available for download here: https://www.thingiverse.com/thing:2470288

[ldm314]

I've set up the primary with the 12v supply. I'm using this to test things until I am sure it is working well enough before I connect up the larger supply. I added 2 more capacitors in each of the strings to raise the frequency as well. With this setup at the lowest setting I tried, 1.75 turns, the royer ran perfectly around 333khz. At 5.75 turns it runs at 200khz. I'm fairly satisfied with this result. The secondary without a top load will be near 320khz.

It is not drawing much current in this configuration. I don't have the tools to measure this but I judge it by the response of the power supply. At high currents the fan comes on and the led on the supply dims. At up to half a second of on time there is no response from the power supply yet. The IGBT that I am switching the royer with is cool to the touch and does not have a heatsink on it. I put an aluminum soda can in the middle of the coil to attempt to load it but it didn't have any effect. So far in my tests it takes a minimum of 40ms for the primary to ring up to its peak voltage.

Next I'll be working on a torroid to use up the rest of this copper tubing. I've cheated a bit on the secondary and ordered one made for me. It is currently being shipped and should be here in a few days. It comes partially finished, without endcaps. I'll be 3d printing a top cap for it that will also be used for the torroid mount.

[Mads Barnkob]

You are making some fast progress, you are moving way faster than I can manage these days You know, first there is 2 hours spend to clean the work desk before the latest projects gets into the daylight again!

Neat primary support and thanks for sharing the files, is it stiff enough to not need any cross beams between the supports? When using zip ties to hold the copper tube in place, be sure not to run the primary coil hot, the zip ties will loosen up and in worst case you have a shorted primary coil and exploding bridge.

[ldm314]

You are making some fast progress, you are moving way faster than I can manage these days You know, first there is 2 hours spend to clean the work desk before the latest projects gets into the daylight again!

Neat primary support and thanks for sharing the files, is it stiff enough to not need any cross beams between the supports? When using zip ties to hold the copper tube in place, be sure not to run the primary coil hot, the zip ties will loosen up and in worst case you have a shorted primary coil and exploding bridge.

I'm outgrowing the space here quickly, right now all of this is sitting just next to me. Once I get the secondary I definitely will need a bit of room to work around. I've got another desk to move in here today that will be dedicated to this project for a while.

The form is somewhat stiff. No bracing is needed, but I had to be careful at the top. With the secondary in, it should be a fairly tight fit. I formed the coil by hand into the general shape of it and then zip tied starting at the top down. The form was printed mostly hollow, the walls are 1.6mm thick. If I printed it solid it would be quite stiff. The plastic I used also has a fairly low melting point as well so it may deform if it gets too hot. I may have to print another one out of a different material if heat becomes too much of an issue.

All of the testing I did in this configuration worked well. I got it to draw a bit of load putting a fruit can instead of aluminum soda can. Without the micro controller the royer ran fine for longer times. I only tested up to 30 seconds or so, but that is more than enough. I am concerned that this ignition IGBT used to turn the royer on/off might not work well at the full supply, but now I have a the gate driver and a bunch of the 600V IGBTs as a backup.

[ldm314]

Built my first topload today. Used the leftover tubing I had, not as large as I wanted so I may go and buy some more tubing for it. I want more of a toroid shape.

I've seen many made from copper tubing but not much detail on how it is connected electrically. A coil is definitely not what I want so I've connected them together with some stainless steel safety wire. Stainless doesn't conduct too well so I am going to follow it up with some copper wire between coils.

I've posted this model up as well. It needed more support as it was taller and won't be supported on the inside like the primary is. https://www.thingiverse.com/thing:2472952

[Mads Barnkob]

I agree, you need some more turns closer together when using think copper tubing like that, else the electrical field does not quite resemble that of a toroid made from a sheet of metal.

It is also correct that you short circuited the coil, as you need one large surface to store electrons on, and not just a thicker extension of the secondary coil

[ldm314]

Was bored today waiting for the secondary so I moved the micro controller to a bigger breadboard and updated it a bit. Got rid of the usb power adapter in favor for a 7805. The micro has a 3.3v regulator on board so its just a simple 5v power source. Instead of the transistor setup I had, I added an inverter that does the same thing. Added a feedback antenna type thing to the IN A and connected the micro's inverted output to the EN A on the mosfet driver.

If anyone remembers Kynar wire wrap wire, It's silver plated copper with a nice insulator around it. I'm going to wind a bunch around a paper towel tube and see if I can get something to happen.

Edit: Results were amusing, but according to JavaTC my resonance frequency would be in the 2mhz range. It somewhat worked, but the mosfet heated up pretty quick and the voltage wasn't great. With 12v in, it arced maybe a couple of mm to the tip of my finger. My scope reads no where near 2 mhz, so it was difficult to tell how well it worked.

[Mads Barnkob]

Before your edit I was about to comment that it sounded like a very high frequency coil that you wanted to try there, you properly used all the power for switching losses as the poor little MOSFET driver can not supply enough current to drive it at 2 MHz.

What supply voltage are you giving the MOSFET driver? I hope its more than 5VDC The driver is actually rated at up to 2 MHz, so that should not be the problem, but if its only 5VDC you will spend all the time turning the IGBT on/off

[ldm314]

Before your edit I was about to comment that it sounded like a very high frequency coil that you wanted to try there, you properly used all the power for switching losses as the poor little MOSFET driver can not supply enough current to drive it at 2 MHz.

What supply voltage are you giving the MOSFET driver? I hope its more than 5VDC The driver is actually rated at up to 2 MHz, so that should not be the problem, but if its only 5VDC you will spend all the time turning the IGBT on/off

Running at ~14VDC input. It was an interesting test, learned quite a bit about feedback to get it to oscillate at the secondary resonance frequency. Before I started it, I put a decent size heatsink on the mosfet. Even at short runtimes it got quite hot. Looking at the datasheet the mostfet was spending likely 50% of the time switching.

Once I get the secondary and finish with the Royer circuit I am going to move on to a proper DRSSTC. The secondary is at the local post office, will be picking it up in the next few days. I also picked up enough enameled wire to wind another similar sized secondary. I have 10 of these bigger IGBTs, and a handful of the driver chips now too.

[ldm314]

Got the secondary today, will be working with it later this evening. Bought this from eBay and it looks great. The 3d printed form is a little on the tight side so I may have to reprint it scaled up slightly. First testing will still be on ~14VDC, will be how the Royer behaves with the coil in place at the various tuning points without the top load in place. Then I will be adding the top at 14VDC.

If everything is good, I'll be testing with the 42VDC supply but it requires a larger workspace first. I doubt my computer or 3d printers would be happy with a coil nearby.

Update: It does something. No sparks but I can light up bulbs from about a foot away. http://i.imgur.com/KMIHSHF.mp4

Update 2: I checked the input voltage. It quickly drops from 12v to 9v on my 30A power supply. With the lightbulb on the top the brightest is around 2 turns on the primary.

[ldm314]

Running on the 42vdc we have first light! I have not tried how long it runs, but easily 30+ seconds. It isn't quite impressive but I can control the sound with the microcontroller. Full on is about 2-3cm continuous spark that hisses.

Definitely some power here, the ignition IGBT I am switching the Royer with gets noticeably warm. I am guessing the small size of the sparks has to do with the input voltage of the Royer. More testing is needed for higher power, but first I am going to write software to make it do music or something. Probably the first tesla coil with a built in web server....

[Mads Barnkob]

Congratulations on first light! That is always a moment of joy and a good sense of achievement

The heating and small sparks could indicate that its far from being in resonance, so little energy is transferred and the higher current in the primary circuit just results in higher losses.

I think you are right about it being among the first with a built-in webserver

[ldm314]

Insomnia has kept me up and coding. I guess now it is the first wifi musical tesla coil? While the voltage is low the current appears to be amazingly high. I am going to need to save up for some test equipment to get better measurements on this.

Next steps will be finishing the power supply for higher primary voltages. I expect past 50 volts these mosfets that came with the Royer circuit will fail. I plan to replace them with my IGBT's and see if I can get it to run at a much higher bus voltage. I've scavenged a transformer out of a dead AV receiver that is supposed to be "35 KVA" and has some high current taps at ~ 60 and ~90VRMS. Eventually I hope to go to directly rectified mains (120 or 240 single phase here). With a voltage doubler and some capacitors I could get maybe 10X the input voltage I have now.

Tuning wise, the best I can get is in this video here. Higher or lower results in reduced spark length and reduced distance to light up this CFL bulb. After playing with the wrench and music here my breakout point has melted into a ball and doesn't work as well anymore....

Edit: the heating I was referring to was the IGBT that is turning on/off the entire Royer circuit. If I remove the top load I have to re-tune the primary at a different point to get sparks again.

[ldm314]

After as much tuning as I could do, this is the best result. About 3-4cm spark tells me I am at nearly 100kv. The flame tells me I am pushing lots of current, as well as the fact that the 6ga primary wires get pretty warm.

The mosfet heatsinks on the royer stay cool to the touch. The capacitors warm slightly and the filter inductors warm up as well. I've been playing midi files for an hour or so and the only issue I have is the breakout wires melting at the tip. I am thinking I just need more input voltage or a larger topload.

Thoughts going forward:

• Swap mosfets for higher rated IGBTs
• Lower primary capacitance and add more turns for higher coupling
• Convert Royer to half bridge and make a DRSSTC
• Put together a rectifier/voltage doubler to run the primary at nearly 10x higher voltage

[ldm314]

An update on this project, things are still in the works. At 42VDC input this circuit will run pretty much continuously. The current limitation is in the interrupter IGBT that is hard switching the 42vdc as an input to the royer. It gets fairly warm playing music, does not heat up if left on. The weakest link is the steel wire I am using as electrodes. They spark and melt if left on for any significant time.

The Royer mosfet heatsinks remain cool to the touch throughout operation.  Without interruption it creates a pretty solid arc/plasma flame that hisses slightly. Estimated at 100kv, I have not measured the current.

With school starting for my children things have been too busy to make much forward progress. Over the next few weeks as things have stabilized I hope to improve things dramatically.

Next steps are to start cranking up the input voltage. I have easy access to 120vac mains. Using a voltage doubler I could get a ~300V differential instead of the 42V I have now. This would result in a much more impressive spark length. If this isn't good enough I can wire up a circuit here providing 240VAC at up to 100 amps...

Time to start exploding some mosfets and IGBTs!

[Mads Barnkob]

About the voltage estimation from Tesla coil spark length, the most recent measurements done by Hydron, Uspring suggests that the output voltage is more in the range of 15kV/cm for a high frequency Tesla coil generated spark.

It is good that everything runs cool, that is always a good sign of low losses/ZVS. How about primary coil and secondary coil heating? When pushing so much current, estimated from the flame like discharge.

Friends, family and job certainly is known to ruin many good hobbies! You are not the only one struggling to find time for it all.