Odd waveforms from my 1.25kW SSTC - please analyze / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Odd waveforms from my 1.25kW SSTC - please analyze

1 2 next

Move Thread

LAN_403

Sigurthr

Sat Aug 11 2012, 03:27PM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Hello everyone!

Let me first preface this by saying that I am very much a novice when it comes to the oscilloscope. I just got my first one mere weeks ago so I know I have a lot to learn when it comes to understanding waveforms. That being said, I know my way around my scope pretty decently as far as basic use goes.

I just received a x100 probe I ordered from eBay, so I am just now comfortable scoping my TCs as I would not even power on the oscope while the TC was on with a x10 probe attached. The way I set this measurement session up was to support my x100 probe on a PVC tube about 8 to 9 feet from my TC, and I powered my TC via my variac to 30Vrms input. Current draw was approximately 1Arms. Moving the probe mere centimeters closer greatly increased the amplitude above levels I was comfortable with.

With my novice eyes I can see 120Hz AM imposed on the f0 "carrier", as well as what looks like a second frequency much much higher making the trace appear so thick. I also noticed (immediately) that there is an additional wave function added to f0, and I'm not sure exactly what it is. I think it could either be another f0 wave out of phase that is of a lesser amplitude, or perhaps another frequency added into the mix. I've no experience looking at anything but harmonic-free sine and square waves, so I don't know what it looks like when various harmonics are added in, or when various phases are in the mix either.

What would the other waves be from? I was very surprised to see the mystery wave function added to the expected sine at ~180KHz, but I was also surprised to see that MHz band wave in there as well and have no idea what would be causing either of them.

Pics:
SSTC-4 immediately after completion:

Low frequency (120Hz) Amplitude Modulation of the envelope due to DC bus ripple:

2uS/div @ 2V/div (20mV + x100 probe):

1uS/div @ 2V/div (20mV + x100 probe):

My scope is only 40MHz analog so the fastest timebase I have is 0.2uS/div, at this timebase the edges of the very high frequency which is thickening the trace can just barely be seen. Many cycles fit in one of the 0.2 ticks, so this frequency looks to be well in to the MHz range. I didn't take pics for anything above 1uS/div because I can only move the waveform about a half a screen in each direction from where it shows, and each cycle of the main frequency is so long at that point that all you see is a barely curved line, so nothing really to see there.

I plan to scope my other SSTC in the near future. I'd do it today but I'm afraid I am out of free time already.

Ben Solon

Sat Aug 11 2012, 04:46PM

Registered Member #3900 Joined: Thu May 19 2011, 08:28PM
Location:
Posts: 600

the thickness of the trace is standard for any nonuniform signal on an analog scope. every pass gets a slightly different reading. i would say that you're not switching quite in tune... i dont know though what are you using as a feedback system?

Sigurthr

Sat Aug 11 2012, 11:17PM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

It uses an antenna + 74HC14 Schmitt Trigger for feedback.

Dr. Dark Current

Sun Aug 12 2012, 08:28AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

I think the Tesla transformer, especially with higher coupling, can also transfer higher harmonics from the bridge output. If it was a DR with near-sine wave on the primary, I think you would see a pretty clean sine from the secondary as well.

Steve Conner

Sun Aug 12 2012, 09:34AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I've never seen anything like that! Any time I've tried this experiment the result was a sine wave as you would expect.

The higher resonant modes of a Tesla transformer aren't simple multiples of the fundamental frequency, so they aren't excited by harmonics from the inverter. Maybe with tight enough coupling they could pass through anyway.

Mr Pink

Sun Aug 12 2012, 11:57AM

Registered Member #5174 Joined: Tue Jun 05 2012, 06:28PM
Location:
Posts: 4

I once saw similar waveforms on my SSTC.
It was very strange, almost no activity at the top terminal but when waving at it with a chicken stick I noticed there was plenty of high voltage at the middle of the secondary.

Don't know why it started behaving like that, there was no change in drive behavior (same frequency as always) and I hadn't changed the coil in any way, at least to my knowledge.

After a bit of trial and error it disappeared with lower coupling.

Sigurthr

Sun Aug 12 2012, 01:13PM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

@Steve Conner
I think you may be right about the tight coupling allowing harmonics into the output. I scoped my much smaller 500VA coil today and the normal sinusoidal waveform was displayed, just as expected. The primaries of the two coils are identical in size (10 turns 10ga on 4.5" form), the only difference is that the small coil has a 3.25" secondary and the large coil has a 4.5" dia secondary, so coupling is less on the smaller coil. I even have the small coil's primary raised to the point that highest output and current draw is observed, where as on the large coil it is starting just above the first turns of the secondary. When I built the large coil I saw the 11A draw and could not afford to risk destruction which may have happened if I were to have looked for an input/output maximum by changing the primary coupling. As is the 11A draw drops my line voltage significantly and risks heating of the wiring in the walls. I try not to draw more than 10A on any one circuit in this house.

Here's the oscillograph of my small 500VA SSTC with identical test procedure/setup (note that the time/div is different as this coil operates at 500KHz).

What are (and how do you determine) the higher resonant modes for a tesla transformer? Do you know anywhere (online) I can read about it more? I wasn't aware there were higher resonant modes, I thought a tuned circuit of high Q could only oscillate at a fundamental frequency and no other. Is there anyway I can extrapolate the frequency of the harmonics from my oscillographs?

@Mr Pink:
This coil is operating rather well though with no noticeable HV at the middle of the secondary. At full power (120Vrms @ ~11Arms) I get 6" wavy spikes (from the 120Hz) from the big fat CW corona discharge. Mosfet heating is well within normal parameters too.

The only issue I have with this coil is that the Primary gets very warm. It is 10ga solid multi-strand copper in 600V rated hi-temp PVC insulation, and after about 45sec at max power the primary is over 110 degrees F.

Steve Conner

Sun Aug 12 2012, 01:34PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

A Tesla coil secondary has a whole series of resonant modes. The one we normally use is the 1/4 wave mode. The 1/2 wave would produce zero voltage at the top and a voltage maximum in the middle of the secondary. The 3/4 would produce two maxima and so on.

As a radio ham you surely know that a straight piece of wire can resonate in multiple modes. Coiling it up doesn't change that fact, but it shifts the mode frequencies away from a simple integer relationship. Adding a topload to the free end shifts them even further: the fundamental frequency goes down and the others go up.

Paul Nicholson's TSSP papers describe all of the math in agonising detail. He wrote software that would calculate the first 10 resonant modes of a coil given the dimensions and wire gauge. The engine of JavaTC is based on his work, but it only calculates the fundamental mode.

The Seibt coil is a very tall and skinny Tesla coil designed to demonstrate the modes.

Sigurthr

Mon Aug 13 2012, 02:15PM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Hi Steve,
Your reply has sparked a great deal of questions in my head! Please do not think that I am questioning your answers' validity with this post, no; I am just starving to learn now that I find many things I believed I understood have turned out to be misunderstood. Text cannot convey what speech can, so; this post should be read with a tone of "very excited student" and not "arrogant thinks-he-is-right s.o.b.", haha.

Steve Conner wrote ...

A Tesla coil secondary has a whole series of resonant modes. The one we normally use is the 1/4 wave mode. The 1/2 wave would produce zero voltage at the top and a voltage maximum in the middle of the secondary. The 3/4 would produce two maxima and so on.

Ahh, okay, that clears it up a bit. When you said "modes" I was thinking maybe something analogous to laser transverse modes where multiple wavefront phases clash to produce interference and got thrown off. In antenna theory, an end-fed 1/4wave antenna is really just a center-fed half-wave antenna where the ground comprises the other "missing" leg. Your explanation about placement of voltage maxima sounds more like feedpoint location than wire length. A center-fed half-wave wire has two voltage maxima; one on each end, but an end-fed half-wave wire has one maxima in the center and a minima at the end (iirc). If a tesla transformer is end-fed by definition (grounded secondary end) and you forcefully pump the transformer with double the frequency the tesla transformer normally operates at you would get what you describe (because you would be going from end-fed 1/4wave to end-fed 1/2-wave), but I don't see how a self-resonant design where the output frequency sets the input frequency can operate at any frequency other than the fundamental. If that were the case would we not see lots of coils operating at frequencies far from the predicted (fundamental) value?

Steve Conner wrote ...

As a radio ham you surely know that a straight piece of wire can resonate in multiple modes.

My understanding is that any length of straight wire can radiate (not "resonate") at any frequency, and what makes the normal length to frequency ratios we commonly use useful is the impedances found at them, not that they permit any better emission. For example, a "half-wave" wire is ~70ohms impedance to the frequency which has a wavelength twice its length, and 70 ohms is close to both commonly used coaxial transmission line's characteristic impedances. If you used a theoretical perfect impedance matching network any length of wire would work equally as well for the same frequency as the integer based lengths we commonly use as long as the nonreactive impedance was negligible. A resonant antenna has zero reactance, leaving the total impedance as low as possible, so maximum current draw/delivery is achievable.

Also, most of the HAM material/papers/knowledge/operators I've seen incorrectly use(s) the term resonance/resonate to mean impedance matching. Resonance is the state at which inductive and capacitive reactances cancel out (equal in absolute value), nothing more, nothing less (as far as I know, anyway). So how can a piece of wire (or a coil) be resonant at more than one frequency? If you were to plot frequency, capacitive reactance Xc, and inductive reactance Xl on a line graph resonance is the point where Xc line and Xl line meet. The Xl and Xc lines never meet at more than one point.

(my graph is probably highly inaccurate as far as the slopes of the lines go as I think you can never have 0Xl or 0Xc, but the general idea should be right)
Fresw

X (Y-axis) = the absolute amplitude of reactance, not raw reactance.

Steve Conner wrote ...

Coiling it up doesn't change that fact, but it shifts the mode frequencies away from a simple integer relationship. Adding a topload to the free end shifts them even further: the fundamental frequency goes down and the others go up.

This I did not know! I thought the resonant frequency of a coil was completely determined by the inductance (and associated capacitance, be it self or external) and had nothing to do with the wire length. I also thought:
1)the integer relationship between fundamental frequencies and their harmonics are "hard wired" and a function of the wave nature of EM
2)a top load and its capacitance (which is to the earth and not self-contained) affect the feedback loop's frequency responce in the same direction no matter the frequency being applied. I.e. adding more capacitance to a LC filter changes the attenuation factor for both low frequencies and high frequencies in the same direction.
3)a resonant circuit has/produces/allows through ONLY the fundamental frequency and no harmonics (because harmonics are so far from the fundamental that an incredibly low Q would be needed to allow them through, and I don't even think Q's that low are realistically achievable).

I realize that real world circuits do not have infinite Q and thus strong non resonant (or harmonic) frequencies can propagate through reactive filters but:

4)I thought oscillators can only oscillate at one frequency at a time; you can't have an oscillator oscillating at both 60Hz and 120Hz, it physically can only do one or the other, even if the feedback network can allow both 60Hz and 120Hz through one of the two frequencies will have higher gain and win out in the end.

So I thought in a SSTC the tesla transformer is both a load and a LC filter in the feedback loop which allows only the resonant frequency to achieve gain greater than unity. Any multiple/harmonic of the resonant frequency or any other frequency would not achieve enough gain due to attenuation by the tesla transformer to maintain oscillation. if you pumped the transformer with anything but the resonant frequency you would get negligible output as the transformer effectively acts as a load with a series high-impedance.

Steve Conner wrote ...

Paul Nicholson's TSSP papers describe all of the math in agonising detail. He wrote software that would calculate the first 10 resonant modes of a coil given the dimensions and wire gauge. The engine of JavaTC is based on his work, but it only calculates the fundamental mode.

The Seibt coil is a very tall and skinny Tesla coil designed to demonstrate the modes.

Is there a link to that software? I'd love to get my hands on a copy!

I've seen Seibt coils once or twice before but thought they were force-fed nonresonant designs intended to show the effects of being off tuning / untuned.

Steve Conner

Mon Aug 13 2012, 03:11PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

You are right to compare the Tesla resonator to a LC filter, but all real inductors have unwanted self-resonances. This is a big problem in design of plate feed chokes for RF amplifiers, for instance. The Tesla resonator is no different, it has the lumped resonance due to the coil inductance and the topload capacitance, plus a whole series of transmission-line resonances.

A wire antenna will radiate at any frequency that you can drive a current in it. The resonant frequencies are those at which it's easiest (by far) to drive RF current.

Hopefully Paul Nicholson's papers will be able to answer more of your questions:

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob