Donald Smith Devices too good to be true

[tesluh]

Still nothing spectacular for results...


Has anyone here experimented with the "Plasma tube" device? 4 foot tall plasma tube with a sheet of plastic/copper/aluminum as a plate capacitor. looks easy to make and Don said capable of something like 65000 watts.
[VIDEO]https://youtu.be/-NbmheIGG_Y?t=965 [/VIDEO]

[tesluh]

None of my nst's would spark to earth like that so this should be fun to experiment with today. (or maybe I never set them up properly)

[tswift]

Yes, exactly like that. It will zap the coil/cap combo at the output frequency of the ZVS driver and the spark should be a continuous hiss. This is the configuration shown on many of Don's schematics and Bruce's original diagram.

In the new modified diagram from Bruce, all that the ZVS does is charge up a cap to a high constant voltage. The small current from the antenna charges the C2/L1 cap/coil combo in the configuration more like the standard Tesla coil primary, where the spark gap goes between the HV supply and ground, and the coil and cap are in series. If you did it this way using the ZVS it might spark 100 times a second and the spark will sound like rapid-fire pulses and can be quite loud. The oscillation frequency of the coil and cap combo is the same in either case and is determined by the capacitance and inductance, not the driver frequency.

[tesluh]

something like this?

[tesluh]

"The spark frequency might be 30 KHz for example, and the coil ring frequency will be determined by its quarter-wave resonance and is a function of its length and geometry."

So basically if I tune the coil to resonance peak at that same frequency (using its geometry and capacitors), the circuit will ring continually at resonance?

If you wouldn't mind, please tell me how you would use your pvm12 to configure the primary of a tesla coil? I don't need inductances and capacitances, I'm just looking for the schematic of how you would connect it all.

Sorry if this should seem obvious to me but I have looked at hundreds of schematics and have spent so many hours on the bench with so many nsts that I cant see straight now I am playing with flyback coils and learning more yet. So I am questioning the basics (since I cant seem to reproduce the basic effect of resonating a coil).

[tswift]

If you're using the output of any HV HF supply with a rectifier then it becomes essentially a high voltage DC supply and frequency isn't an issue. For instance, the standard spark-gap Tesla coil configuration on the primary side of the Tesla coil, where the HV charges the cap/coil in series, but then as soon as the voltage on the cap reaches the spark gap breakdown voltage, it arcs over and the cap and coil form a parallel resonant circuit that rings at the resonant frequency determined by the inductance and capacitance. Similarly, if you rectify the output of the HV HF supply, then through a spark gap, into the hot end of a coil whose other end is grounded, the coil will ring with every spark, see some of my posts from a few pages back where I did this with pictures and scope traces. The spark frequency might be 30 KHz for example, and the coil ring frequency will be determined by its quarter-wave resonance and is a function of its length and geometry. For a typical benchtop coil it might be 500 KHz to several megahertz. You can then add capacitance in parallel with the coil to bring this down to any desired frequency, although the capacitor better be able to withstand the voltage required.

Another consideration comes into play using a flyback instead of a function generator, and that is the power in watts. If you had a coil w/ parallel cap combination that resonated at say 50 KHz, and you adjusted your flyback/ZVS power supply to this exact frequency and drove the coil with it (with no spark gap) it would resonate but the voltage is going to depend on how much power your flyback can deliver, it would take significant power to oscillate it continuously with kilovolts of AC, and the capacitor (or cap bank) had better be able to handle the large circulating currents, it will get hot. The amount of drive power will be related to the Q (quality factor) of the parallel resonant circuit, which is defined as the reactance part (inductive or capacitive reactance at the resonant frequency) divided by the resistance in the tuned circuit. So a bigger coil with more inductance, which needs less parallel capacitance to tune it down to a given frequency, will ring proportionally better than a smaller coil with more parallel capacitance at the same frequency. The best situation is a coil driven at it's own quarter-wave resonant frequency with no added capacitance beyond the coil's own parasitic inter-turn capacitance. Then the bigger the coil and the bigger the wire, the higher the Q, which is how Eric Dollard was achieving tuned circuit Q values of 5,000 and more. If you've seen the pictures of Tesla's "extra coil" from his Colorado Springs Notes then you get the idea of what a more-optimized coil looks like. The modern "Tesla coil" type resonator coils which are intended to throw big sparks primarily for entertainment value are very far from optimum but work in that application.

Changing the frequency of a ZVS driver is not so easy, you would have to change out the resonant capacitor on the primary side of the HV transformer. The "zero voltage switching" design is self-resonant and automatically runs at its own resonant frequency, ensuring that the MOSFET's always switch on and off only when there is no voltage across them so that they switch efficiently and stay cool. But you can't easily tune it, you would have to change the resonant frequency of the circuit by changing the capacitance. You might be able to do this on the HV side (the flyback transformer secondary) with a tunable cap rated for enough voltage, probably someone out there has done it this way but I haven't personally tried it. This is one of the advantages of the PVM12 design, it has an IC chip driver that forces the MOSFET's to switch at a given frequency which you can then dial up or down with the knob on the back panel. Its tuning range is about 48-70 KHz. But the ZVS is simpler, more robust, and can deliver significantly more power especially when you ramp up the input voltage to 24V or more, depending on how much the components can take. So they each have their advantages.

[tesluh]

I saw inside the end of your pvm12 and saw what looked like a flyback, must be very similar. can you suggest how to drive a coil at its resonant frequency using a flyback? I have the feeling it will be the same as when I hook my function generator up except instead of 10 volts it will be kv! Would be nice to hear that from someone else before doing it though

This is tesla coil type info that you would think would be easy to find. not so much though.

[tswift]

Also a quick progress update, I'm doing a little engineering trying to deal with the issue of too much voltage possibly reaching the supercaps. If there really turns out to be a power gain coming through the L1/L2 coils, then the L2 voltage will be something like 100V at low impedance, and charging a 12V cap bank from it will cause some problems, to put it mildly. The small bridge rectifier diodes probably would blow first from excessive current, but I really don't want to be buying too many $40 cap banks, especially when I have to wait two weeks for them from overseas.

So my current plan is a shunt regulator with a zener diode driving the base of a beefy shunt transistor. Using a 2N3055 on a good heatsink it should be possible to dissipate 50-100W continuously and much more for a short duration. As I said before, I'm not really expecting those kinds of power levels but I don't want to take chances. I'm also thinking about a relay that will cut out the supply current to the HV supply in the event the voltage climbs to danger levels for the caps (around 16V) but how to design the circuit is a bit tricky. Any kind of digital logic running next to high voltage circuits and spark gaps is an issue, you'll never eliminate glitches and problems. Pretty much only simple robust analog circuitry is likely to work well. Normally you could ground and shield the digital parts of the circuit, but on the L2 side of the coil it needs to be deliberately NOT grounded and shielded, so it's exposed to the ambient environment.

Another idea I had was to use either a tap coming from the L2 coil just a few turns from one side, where the voltage should be only somewhat greater than 12V, or a totally separate tertiary coil with just a few turns as a pickup coil. I am envisioning this coil wouldn't be self-resonant so no parallel cap on it. However, at that point I'm really changing the circuit and departing from Bruce's design so I don't want to go that direction if there is some better option. Probably the simplest option is to just run the circuit open-loop from a 12V battery to begin with, so I can measure the power output and find out how much needs to be dealt with.

[tswift]

Hey, good find! I may pick one of those up, right now I'm building this rig around my only PVM12, so I either won't have a driver for other experiments or I have to find something else that will work for the experiment. If you look at what's inside the PVM12 it's quite similar with a 2-MOSFET driver board but it's a frequency-adjustable driven inverter and not self-resonant, so it's not zero voltage switching. The ZVS probably runs cooler and more efficiently, but you can't tune it. For some configurations this might matter and for some it won't. For Bruce's current schematic that I'm building to it won't matter.

[tesluh]

W6 12v-36V ZVS Tesla coil flyback driver/Marx generator/Jacob's' ladder | eBay

flyback and driver delivered for less than $19. shipped from Philadelphia took total of 6 days to arrive instead of weeks from china. now to try to get it to drive this this circuit!

[tswift]

Right, so now the L1 primary is essentially excited just like a spark-gap Tesla coil, charging in series and discharging in parallel when the PPV breaks down. Since the C2 capacitance is so small this should happen very often, many many times per second. If you look closely there are some other significant differences too: the antenna will now be charged negative, and the HV/HF power supply is essentially just supplying a bias voltage across C1. It might be even be possible to use a static machine like a Van de Graaf or Wimshurst to charge C1, although obviously the voltage rating of C1 would have to be much greater. Essentially none of the power from the HV power supply will be delivered through the L1/L2 step-down transformer in this configuration. The only power input is what is supplied by the ion current through the antenna. Also, the L2 side of the circuit will now be fully floating and not ground referenced at all, so it will be very difficult to measure while the circuit is in operation. Also, the required isolation voltage between L1 and L2 is greater, so care should be taken where wires are close to prevent flashovers and accidental damage, especially to the sensitive supercaps. And finally, lightning protection could be an issue with an elevated wire charged to several kilovolts negative and generating ionization, still thinking about the best way to handle this.

This was all news to me too: I spent a good half an hour staring at the schematic until I felt like I understood it. As far as the build goes it doesn't really change very much, just rearranging a couple of connections. I feel like we're getting closer, step by step. Again, thanks to Bruce for sharing his hard work with the rest of us.

[serendipitor]

For those that didn't look closely at Bruce's v5.4.7 disclosure, it contains a slightly revised/improved schematic.

Changing L1/C2 from parallel to series resonant is big change. From a conventional view point (which I know has some pitfalls) it makes more sense to have the series connection, given that the antenna presents such a low impedance. In fact, I wonder how it could have worked the other way?

[tswift]

OK, more progress today. I cut and drilled the PVC poles for the long wire antenna. I had some old scrap pieces of 2" PVC that are no good for plumbing any more but perfect for this. I should be able to rig it 12 to 20 feet high, I have to route around some obstacles and use existing posts for reinforcement, PVC isn't very rigid. Fiberglass poles would be much better but this is what I had on hand.

Also more progress on the device, more components are now mounted on the board. For those that didn't look closely at Bruce's v5.4.7 disclosure, it contains a slightly revised/improved schematic. I have revised the layout to this configuration.

[soundiceuk]

The first antenna is up!

[tswift]

Bruce, is that really you posting here? If so, welcome! And thanks for joining us. Thank you for sharing all your hard work and research with the rest of us. I know that I personally have been following in the same footsteps and it's time consuming, expensive, and discouraging a lot of the time. Only someone truly committed can keep at it long enough to produce results, and even then mostly it's a thankless job. This is the most important technology in the world, and together we can all help it see the light of day.

Also, I'm not fond of Paypal because of their tendency to censor things they don't like. There are numerous examples, Wikileaks being the most notable one. To truly bring this technology forward I believe it will be necessary not to work within the existing systems because they will use that to shut it down. Instead I use and promote bitcoin. Much like the IPFS file system I've been using to post, it is distributed, peer-to-peer, and censorship resistant. It can't really be stopped short of turning off the internet. So if you would like to either post a bitcoin address or send it to me privately I will happily send you some financial support.