Tuning and Frequency Splitting within a DRSSTC (or two coil) system / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Tuning and Frequency Splitting within a DRSSTC (or two coil) system

1 2 next

Move Thread

LAN_403

HV Enthusiast

Thu Apr 13 2006, 02:12PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Just a question regarding frequency splitting within a two coil coupled system.

With two coupled coils, there is a frequency split which occurs that is depending on the coupling factor. The higher the coupling, the greater the frequency split. That we know. Now, with DRSSTCs, it has been stated that tuning the primary circuit to the lower frequency pole of the frequency split is beneficial for DRSSTC systems.

The question is, and its not as easy or straightforward a question as it sounds, is what exactly is occuring when you "tune" the primary to the lower pole of the split frequency. Firstly, as you decrease the natural resonant frequency of the primary circuit, the lower pole frequency will move as well, so you will never match the natural primary frequency with the lower pole frequency.

Does tuning the primary circuit lower in frequency (closer to the lower frequency pole) force the feedback to operate the DRSSTC at the lower pole frequency? To basically kind of fall in place?

What are your thoughts?

Dan

Steve Conner

Thu Apr 13 2006, 02:53PM

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

I'm not telling, but you can buy my book that explains it for only $39.99

Ha-ha only kidding.

It's true that you will never match any of the pole frequencies with the primary frequency.

Now, as you change the tuning of the primary relative to the secondary, both pole frequencies move, and (this is the important bit) their relative magnitudes also change, as follows:

If you tune the primary to a lower frequency than the secondary, both pole frequencies move downwards. The magnitude of the lower pole increases and that of the upper pole decreases. (assuming the magnitude you're measuring is primary current: for secondary voltage the opposite is true.)

If you tune the primary to a higher frequency than the secondary, exactly the opposite happens. Both pole frequencies move upwards, and the primary current is heavier at the upper pole than the lower one.

Now, a simple feedback DRSSTC will operate at whichever pole has the highest loop gain. So the pole you end up operating on depends on whether you tune the primary above or below the secondary, whether you use primary or secondary base feedback, and whether streamer loading ever pulls the secondary below the primary, causing it to swap poles mid-burst. (I like to call that mode-hopping.)

In contrast to that, my PLL driver operates at whichever pole I tell it to, which made the whole puzzle that bit easier to unravel. I figured it out while designing and testing the driver, and tried explaining it to a bunch of other people, but I don't think any of them ever understood. I found that with this driver, I got best results operating at the upper pole frequency with the primary tuned lower than the secondary. That isn't possible with an ordinary feedback driver, unless you use secondary base current feedback.

Under extremely heavy streamer loading, both poles flatten out and the response looks like that of a 4-pole bandpass filter. It doesn't really matter which "pole" you drive at under those conditions.

HV Enthusiast

Thu Apr 13 2006, 05:15PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Thanks. Okay, that makes more sense now. I did simulate it (see attached), so the upper pole's magnitude does increase when primary frequency is tuned higher than the natural frequency.

Thanks again!

Dan

]1144948550_15_FT7499_freq_split01.pdf[/file]

JimmyH

Fri Apr 14 2006, 12:38AM

Registered Member #358 Joined: Sat Apr 01 2006, 06:13AM
Location: UCSB
Posts: 28

Now, a simple feedback DRSSTC will operate at whichever pole has the highest loop gain. So the pole you end up operating on depends on whether you tune the primary above or below the secondary, whether you use primary or secondary base feedback, and whether streamer loading ever pulls the secondary below the primary, causing it to swap poles mid-burst. (I like to call that mode-hopping.)

I'd agree with all that, if we were running steady state, but the whole point of the DRSSTC is to not be anything like CW :P

If you use primary feedback (which, I would argue is the best and safest way, since it is the only way to ensure you switch after the 0 crossing), you switch when the primary current crosses 0.

If you look at the primary current, it is not a single sine wave, there is some beating going on. If you do a Fourier transform, you'll find that it is oscillating simultaneously at both frequencies, and therefore, you are driving it at both frequencies at once (although they aren't driven equally).

If you tune the primary to a lower frequency than the secondary, both pole frequencies move downwards. The magnitude of the lower pole increases and that of the upper pole decreases. (assuming the magnitude you're measuring is primary current: for secondary voltage the opposite is true.)

Right, and the since the lower pole is bigger, the primary drives the lower pole more. It 'beats' less, but in the transient response, it is still exciting the upper pole.

I would call that "driving the lower pole", and I think this is Dan was looking for.

Say you had it set up so that the poles were nearly equal in magnitude (ie forming full notches) , but streamer loading pushed the balance a bit, it would just drive the other pole slightly more. There is generally not enough time for a full "mode hop" in a DRSSTC.

teravolt

Fri Apr 14 2006, 02:19AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

In my SSTC I can controle what ever frequency I chose. I noticed that if the tesla is sligtly detuned, eather pole becomes dominate. If it is detuned will the will the stremers be afected in amplitude if the frequency is set to the pole with the greatest amplitude? If the frequency is set to the center frequncy and boath poles are equal isn't more eficiant at F resonant because coupling is gratest?

HV Enthusiast

Fri Apr 14 2006, 02:49AM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Cool! Thanks for the comments again everyone. I did manage to take some measurements with a spectrum analyzer this evening on one of my DRSSTCs and was able to get a nice double-hump response and I did indeed verify that i could reverse the magnitudes off the upper and lower pole by adjusting primary tuning.

Thanks again

Steve Conner

Fri Apr 14 2006, 10:51AM

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

Woo! That's cool that you were able to get experimental results that agree with the theory. It's even cooler that you have a spectrum analyser

JimmyH: I like to think that burst lengths in a DRSSTC are long enough for it to be considered as "sort of steady-state-ish". To deal with the non-steady-stateness of things, I imagine that the drive from the inverter isn't a single frequency, but a band of frequencies, on account of the burst not being infinitely long. (The fewer cycles in the burst, the wider the band of frequencies.) Some of the energy in that band ends up exciting the other pole from the one you're trying to drive, and this is what causes any ripples in the envelope you might see, as the two modes beat against each other.

In the coil I have just now, with a high operating frequency, longish burst, and tight coupling, the unwanted mode hardly gets excited at all. There are no visible ripples or beats in the RF envelope. But I imagine as I build a bigger coil it'll need looser coupling to avoid flashovers, and a shorter burst length in terms of cycles, both of which will get the unwanted mode excited more. So I may need to learn how to drive both modes together, or it may turn out that my PLL driver can't do that and I need to use Steve Ward's design

It's interesting to note that Jimmy's original DRSSTC used a drive frequency midway between the two poles to excite both modes equally, and used the beating between them to his advantage, to get complete energy transfer like in a classical Tesla coil. But nowadays we seem to try and drive one mode much more than the other.

Steve Ward

Fri Apr 14 2006, 06:14PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

It's interesting to note that Jimmy's original DRSSTC used a drive frequency midway between the two poles to excite both modes equally, and used the beating between them to his advantage, to get complete energy transfer like in a classical Tesla coil. But nowadays we seem to try and drive one mode much more than the other.

Actually, on my larger coils, im not detuning much at all (just enough to compensate for streamers)... in fact, now im using beats as the shut off point for my drive. The only reason i detuned was to eliminate notches so that i could constantly increase primary energy in order to get a bigger bang. But with the big coils, i found i can get plenty of bang energy in just a few cycles (maybe 7-9 cycles). So now i just tune it to notch and turn off the drive at that point. At high power, the notching isnt so apparent, but there definately is a drop off in primary current at the end. I gained much of this insight from my pspice simulations. As mentioned before, a notch indicates that secondary voltage (energy) is max, making it the most effective time to shut down the inverter. This tuning method has gotten me to 12' sparks now.

I think the optimal tuning depends a lot on the system itself. As ive said many times before, i think only small coils should need to use the "detuning trick" to get a bigger bang before notching.

JimmyH

Fri Apr 14 2006, 06:36PM

Registered Member #358 Joined: Sat Apr 01 2006, 06:13AM
Location: UCSB
Posts: 28

[JimmyH: I like to think that burst lengths in a DRSSTC are long enough for it to be considered as "sort of steady-state-ish".

Well, it certainly depends on the coil. In my two DRSSTCs, there is a very clear beating going on, although for some coils that drive for many cycles, it may not be as apparent. Even those coils I would have a hard time believing are going to chose one pole, and then 'hop' to the next. Either way, ignoring the transient response is ignoring a huge part of it.

To deal with the non-steady-stateness of things, I imagine that the drive from the inverter isn't a single frequency, but a band of frequencies, on account of the burst not being infinitely long. (The fewer cycles in the burst, the wider the band of frequencies.) Some of the energy in that band ends up exciting the other pole from the one you're trying to drive, and this is what causes any ripples in the envelope you might see, as the two modes beat against each other.

It's not the fact that you have a short pulse that happens to have some energy in the other pole, it's due to the drive frequency being synched to the primary frequency(ies), so when the primary has energy in more than one frequency, it will drive both.

In the coil I have just now, with a high operating frequency, longish burst, and tight coupling, the unwanted mode hardly gets excited at all. There are no visible ripples or beats in the RF envelope. But I imagine as I build a bigger coil it'll need looser coupling to avoid flashovers, and a shorter burst length in terms of cycles, both of which will get the unwanted mode excited more. So I may need to learn how to drive both modes together, or it may turn out that my PLL driver can't do that and I need to use Steve Ward's design

Your PLL also puts out a fixed frequency, so you're going to see less beating for that reason also.

It's interesting to note that Jimmy's original DRSSTC used a drive frequency midway between the two poles to excite both modes equally, and used the beating between them to his advantage, to get complete energy transfer like in a classical Tesla coil. But nowadays we seem to try and drive one mode much more than the other.

What I think is interesting is that you get the same beating if you drive both poles equally as you do when you drive at the zero between them.

It's also worth noting that this approach is more difficult to transfer enough energy with. When my CT broke, I started tuning for max spark length (which happened to be the lower pole), rather than for the notch, and I got a big jump in max spark length. It went from ~50" or something to 74" just by changing the drive frequency, and eventually got up to 97" after playing with it more.

Steve Conner

Fri Apr 14 2006, 08:00PM

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

I think driving both poles equally is pretty much the same thing as driving the zero midway between them, for short transients at least? I guess it's not quite the same, since driving at the zero gives practically no output in the steady state.

I suppose it is possible for the feedback driver to respond to a pattern of zero crossings that actually contains two frequencies and reproduce the same two frequencies in its output, so both poles get driven. But I always saw that as a disadvantage, since the feedback might go funny at the notches.

1 2 next

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