Futterman Class C oscillator?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Disabled Account
Joined 2013
Here are of output impedance plots for //811a VTTC, the higher the current the low output Z.

Glad you like the Class C sim. Because the gain of 6c33c is only 3, is hard to get oscillation, maybe like you say have to use additional tube to boost the gain to 100-200v phase splitting, cap or transformer coupling is fine. I see if I can find suitable tubes, any suggestion?
 

Attachments

  • VTTC 811a output Z plot-1.png
    VTTC 811a output Z plot-1.png
    62.5 KB · Views: 60
  • VTTC 811a output Z plot-2.png
    VTTC 811a output Z plot-2.png
    71.7 KB · Views: 73
  • VTTC 811a output Z plot-3.png
    VTTC 811a output Z plot-3.png
    61.9 KB · Views: 64
Last edited:
Disabled Account
Joined 2013
Here are of output impedance plots for //811a VTTC, the higher the current the low output Z.

Glad you like the Class C sim. Because the gain of 6c33c is only 3, is hard to get oscillation, maybe like you say have to use additional tube to boost the gain to 100-200v phase splitting, cap or transformer coupling is fine. I see if I can find suitable tubes, any suggestion?

Wait.. the sch is wrong..
 
One of the reasons why I want to be able to input my own frequency vs be phase locked with the secondaries secondary coil (via tickler) is to be able to have more control over the different resonance modes. For instance, a grounded resonator can resonate at 1/4λ, 3/4λ, 5/4λ, 7/4λ... while a free resonator will resonate at 1/2λ, 1λ, 3/2λ, 2λ, etc.
Lately I have been finding it pretty insightful to have control over what mode of resonance a coil is in.

Manually adjusting the input frequency also comes in handy for experiments involving concatenated coils. (Two coils serial connected coils that are not magnetically coupled to one another) This is a subject matter that gets irritating with Tickler feedback circuits.

The 6c33c totem pole design sparks my curiosity because it can handle darn low impedances. Primary coils are most effective with the least number of turns possible, demanding for a low impedance transmission circuit to drive it.

Anyway, driving two tubes in parallel, thus abandoning the self oscillating idea, is much easier than in an OTL arrangement. And, btw, source impedance is quite the same, as in OTL both tubes actually are paralleled, AC wise.

Best regards!
 
Disabled Account
Joined 2013
Here is the completed VTTC amp. ECC88/6DJ8 is a high freq tube, should be good as driver. The square off waveform helps to limit the current in amp, as well the max. output. Transformer is chosen so balance drive the totem pole is easier to achieve, the coupling between 3 coils can be very tight and seal if necessary.
 

Attachments

  • 6c33c OTL Class C VTTC amp-1.png
    6c33c OTL Class C VTTC amp-1.png
    93.2 KB · Views: 45
Here are plots for corrected schematic.

Im curious what your interpretations of your 811a simulations of it are

You mention in your graphs things like 'no oscilations' with phase angle etc. I was wondering if you could elaborate on these graphs a bit because I have had a problem where oscillations would eventually cut-out the more voltage I give it from the Variac. (resulting in drawing of too much current) I believe it happens around 200 or 300 volts at the plate. The variac feeds a microwave oven transformer that is voltage doubled. I have a current probe monitoring the current going into ground so I can see the wave form when it decides to collapse after a certain point when turning up the VAC.

Not quite sure if its an impedance missmatch phenomenon or a damaged something.

But for the most part, I've been able to oscillate the coils reasonably when staying bellow that point on the variac.

cYrCWnc.jpg


Also in your last simulation with the 6c33c, I had a couple questions

(1) what is the amount of grid bias you set here?
(2) you mention that the square wave from the first stage limits the current, is that because the ECC88 is clipping? Im cool with that style of current limiting, but what if I want to push the tubes harder (1amp example), does that mean the driver tube choice needs to be able produce more voltage swing?

LSKg5oJ.jpg
 
Disabled Account
Joined 2013
"Also in your last simulation with the 6c33c, I had a couple questions

(1) what is the amount of grid bias you set here?
-150v each tube

(2) you mention that the square wave from the first stage limits the current, is that because the ECC88 is clipping? Im cool with that style of current limiting, but what if I want to push the tubes harder (1amp example), does that mean the driver tube choice needs to be able produce more voltage swing?"

Yes, yes.

Back the plots.

"Im curious what your interpretations of your 811a simulations of it are

You mention in your graphs things like 'no oscilations' with phase angle etc. I was wondering if you could elaborate on these graphs a bit because I have had a problem where oscillations would eventually cut-out the more voltage I give it from the Variac. (resulting in drawing of too much current) I believe it happens around 200 or 300 volts at the plate. The variac feeds a microwave oven transformer that is voltage doubled. I have a current probe monitoring the current going into ground so I can see the wave form when it decides to collapse after a certain point when turning up the VAC.

Not quite sure if its an impedance missmatch phenomenon or a damaged something.

But for the most part, I've been able to oscillate the coils reasonably when staying bellow that point on the variac."

811a is a triode, so I think the output Z is much less than that of the pentode, so the output Z looks reasonable. When it say "no oscillation" simply means the coupling in simulation is disable, sec coil is isolated, so measuring combined Z of the pri tank coil (exclude sec), cap and tube. The peak value of Z also is where the gain is high, I see that the peak for single tube is about 1MHz, but the peak Z for // is about 0.7Mhz, so I think there is a mismatch there. You should see that the peak is very near to freq you operating. Second, regard the relation of power supply voltage in relation to gain and output Z, the more current draws, the less is gain, so if you increase the voltage there is a point where the gain drops until oscillation stop. I attach a plot of el509 plot from 250 to 1.4kv. Can you also lebel on your schematic exactly the component value to see if I can plot again more accurately, and hopefully like you said it's mismatched.
 

Attachments

  • EL509 VTTC resonate voltages -1.png
    EL509 VTTC resonate voltages -1.png
    74.8 KB · Views: 46
Disabled Account
Joined 2013
Here are the plot as I now tuned the primary using a single tube, then plot again with 2 tubes, by connecting C2=2.1n // with pri coil, the peak freq does not vary much except with HT is above 1kv but the response width is wider and Z (max peak gain drop form 96db to 82db=14db HT=500V) drops with 2 tubes. It appears to me Z=5k, HT=1kv is the lowest you can get, you can compare see how.
 

Attachments

  • VTTC single 811a output Z plot-8.png
    VTTC single 811a output Z plot-8.png
    76.8 KB · Views: 34
  • VTTC double 811a output Z plot-9.png
    VTTC double 811a output Z plot-9.png
    85.5 KB · Views: 28
Last edited:
Disabled Account
Joined 2013
Here are plots of fully sim tuned plots. The pri and sec are fully tuned (to 1.26Mhz), then pri is coupled to grid winding, sec winding is excluded, found that the freq peak shifted to the left below 1Mhz. However when all winding are engaged, the freq peaks at correct freq. There are very sharp spike on left on impedance plot, they have to do with amount of couplings between the winding, too high coupling has a severe impact of output impedance, and has to try many value before optimal result is obtained.
 

Attachments

  • VTTC double 811a output Z plot-10.png
    VTTC double 811a output Z plot-10.png
    76.4 KB · Views: 37
  • VTTC double 811a output Z plot-11.png
    VTTC double 811a output Z plot-11.png
    77.4 KB · Views: 34
811a is a triode, so I think the output Z is much less than that of the pentode, so the output Z looks reasonable. When it say "no oscillation" simply means the coupling in simulation is disable, sec coil is isolated, so measuring combined Z of the pri tank coil (exclude sec), cap and tube. The peak value of Z also is where the gain is high, I see that the peak for single tube is about 1MHz, but the peak Z for // is about 0.7Mhz, so I think there is a mismatch there. You should see that the peak is very near to freq you operating.

Interesting to note that the peak Z value of the Pri Tank Circuit is greatly constrained by the Tube choice. This sounds like it leaves a narrow window of effective Frequencies I should design the coils to based on the tube of choice. I remember reading on Steve Ward's site of this concept on getting the sweet spot for reactive power in these designs.

The 6c33c designs you have been modeling, do you think that by having the low gain & low impedance properties, this design would perform better over a broader/more agnostic range of frequencies? (less prone to impedance sweet-spots for efficient usage?)

Second, regard the relation of power supply voltage in relation to gain and output Z, the more current draws, the less is gain, so if you increase the voltage there is a point where the gain drops until oscillation stop. I attach a plot of el509 plot from 250 to 1.4kv. Can you also lebel on your schematic exactly the component value to see if I can plot again more accurately, and hopefully like you said it's mismatched.

This sounds really important, but im not sure I have a full grasp around why more current produces less gain to the point of oscillation stop.

Here are more detailed numbers of the components I used in my setup:

LTQWCTi.jpg


And here is the setup in which im tuning the coils with.

NTbAdzm.jpg


Interesting note on the issue of high coupling too.
Im finding that a tickler here seems to perform best with a fewer turns & further away. Raising the grid resistance beyond 500R -1000R apears to hurt the oscillations..
 
Last edited:
Disabled Account
Joined 2013
Hi,

This is what you gave me:
.08mH primary coil
fine, can be tuned to 1.26Mhz by variable cap.

.28mH Secondary or tickler coil (2Mhz) 21 turns
fine

.72mH Secondary coil (1.26Mhz) aprox 100 turns

The turn ratio is 5, isn't too low. Do you intend to do double tuned resonant coil? Now I see you sec is not tuned (?), so you're using the energy in reflected wave VSWR to generate spark?

In the 6c33c there is no tuning cap in pri, only sec is tuned. The pri coil inductance is chosen so that the voltage is peak at plate, the more the more effective to release the energy from pri coil, but it's no so selective, and so small value can be used together with low output Z to achieve a high ratio, and hence high output in sec. The down side is max. voltage swing (+HT) on the plate the tube can withstand, in this case only < 1kv for 6c33c, for 811A, it can be a few kV more.
 
Last edited:
Oh nice, my mistake, I realized that I confused that 100Turn number from an another extra coil. :E

That secondary is really 42 Turns 72mH.

The reasoning behind my low turn count secondary coil is because Im interested in the resonant rise of the voltage with high Q coils (vs the classical tightly wound high Z secondary coils). The proportions of the coil to wire diameter are ideal proportions to maximize the Q of the coil (minimizing parasitic effects). The typical VTTC usage is for making sparks and swords but I am interested in the not-sparks part of resonant rise and strong ground currents. Yea, I supose VSWR is the idea. Reflecting standing waves.

I have been re-arranging the orientation of these coils quite a bit in hopes for a more effective setup so it may look untuned in the photo. :p
Currently the extra coil is not in the photo. The extra coil gives the secondary something for it to "work into". I'll be putting it back on the set up soon.

In the 6c33c there is no tuning cap in pri, only sec is tuned. The pri coil inductance is chosen so that the voltage is peak at plate, the more the more effective to release the energy from pri coil, but it's no so selective, and so small value can be used together with low output Z to achieve a high ratio, and hence high output in sec. The down side is max. voltage swing (+HT) on the plate the tube can withstand, in this case only < 1kv for 6c33c, for 811A, it can be a few kV more.

Oh.. You're saying that if all things go well, there is a potential danger that the reflected waves from the coils oscillations can be reflected back to the plates of the 6c33c and build up enough pressure to flash over?
 
Disabled Account
Joined 2013
"Oh.. You're saying that if all things go well, there is a potential danger that the reflected waves from the coils oscillations can be reflected back to the plates of the 6c33c and build up enough pressure to flash over?"

Yes that is the danger of flashing. At time the pri can swing voltage higher than the HT+, just like output transformer, more if mismatched.

I'm now try to replace 6c33c with 811a, run HT 1kv or so. It works with bias at -180. With the same coil, the pri is more than 1.2kv and 70kv on sec, freq 200kc. I'm still working on it. Unlike 6c33c, 811a is normally bias at 0 to +100V, if bias at -180v, require at least 200V drive to make it conducts. I raise the HT supply for ecc88 so it can output 250V now.
 

Attachments

  • 811a VTTC totem pole-1.png
    811a VTTC totem pole-1.png
    101.8 KB · Views: 67
I'm now try to replace 6c33c with 811a, run HT 1kv or so. It works with bias at -180. With the same coil, the pri is more than 1.2kv and 70kv on sec, freq 200kc. I'm still working on it. Unlike 6c33c, 811a is normally bias at 0 to +100V, if bias at -180v, require at least 200V drive to make it conducts. I raise the HT supply for ecc88 so it can output 250V now.

Interesting, the totem pole with the 811a.
Pros:
can handle higher voltage swings inside the primary
No need for primary capacitor & tuning?
Similar wattage performance to 6c33c?

What are other advantages that this SSRP like arangement gives us

Any cons?
like the tube impedance frequency mentioned earlier?
Will that hamper the ability to parallel the 811a stages?

> tube ....oscillators for tesla coils.

Tesla didn't have tubes.

He had lived through their invention. If anything
He was fond of Crookes and his tube work most of his life

Tube based oscillators for me proposes a clean, robust and quiet means to creating impulses into coils. Although the rotary spark gap is all performance, the noise bugs me and peoples interpretation of my studies. :p
And solid state equipment is expensive and fragile. I hate the short lived joy of success.
 
Last edited:
Disabled Account
Joined 2013
Interesting, the totem pole with the 811a.
Pros:
can handle higher voltage swings inside the primary
No need for primary capacitor & tuning?
Similar wattage performance to 6c33c?




What are other advantages that this SSRP like arangement gives us

Any cons?
like the tube impedance frequency mentioned earlier?
Will that hamper the ability to parallel the 811a stages?

I just realise the output is just driving a tuned circuit which is of very high impedance, and therefore the output is low. Should use a transformer of 600 ohms and 75ohms for 811a/6c33c, and because in class c and by adding a bank of tubes for more power, With //tubes, the current may not be even among the tubes and impedance drop too low if too many //tubes. I think both should use matching transformer rather than drive a tuned circuit directly. The impedance of the tuned sec of TC is more than 100M, and about 45k after step down which still very high load to drive. So I think better if you step down to 75 or 600 ohms or whatever impedance matching the output. But this load maybe very good for some tube topology such as //811a you get 170W?
 
Last edited:
A matching transformer at the output? Saturation? Core losses at higher frequencies?
Also, too high of an impedance to drive is not ideal for the totem pole design?
I always thought that there was a rule of thumb that driving a higher impedance load is always a better choice (perhaps in standard setups, not totem pole)

I uploaded a video of me walking through my 811a performance dilemma. Im not sure if this is normal behaviour, but with the initial 811a setup, I was never able to open up the variac more than 15v with soso performance. Anything higher and it seems to draw too much current and sags out.

Does this look like normal behavior for the impedances my coils are at?
 
Disabled Account
Joined 2013
"A matching transformer at the output?"

I should rephased that as you need multiple tap on pri coil rather than just one, this makes thing so much easier, to check if particular tap happen to work better. From the plot you see why 100 ohms is a good load for 6c33c at 400V, rather 50k. The plots for 811a is much higher it's similar to a pentode.

I always thought that there was a rule of thumb that driving a higher impedance load is always a better choice (perhaps in standard setups, not totem pole)

A 50k is better load for pentode or transmitting tube like 811, 813 etc, again multiple tap will help to select the best match.

I seen the video. How about you tap the pri coil by connecting the clip a couple of turn less, not full. It's also appear to me the coupling is too low such that the sec is not fully excited, try to experiment with distance of 2 coils after you change the pri tap.
 

Attachments

  • 6c33c totem pole output Z HT 150V.png
    6c33c totem pole output Z HT 150V.png
    94.6 KB · Views: 49
  • 6c33c totem pole output Z HT 300V.png
    6c33c totem pole output Z HT 300V.png
    80.4 KB · Views: 51
  • 6c33c totem pole output Z HT 400V.png
    6c33c totem pole output Z HT 400V.png
    80.8 KB · Views: 50
  • 811a totem pole output Z-1.png
    811a totem pole output Z-1.png
    82.8 KB · Views: 51
Last edited:
Disabled Account
Joined 2013
Using the values of components supplied, I put them in the original sch. Found the waveform is no where near sin wave, more like very distorted short impulse of very high voltage. Something wrong with the grid leak circuit. So I redesign almost completely, separate grid leak and cathode resistor which limit the max. current of each tube. Added the tap in primary coil by splitting into 2 coils. The plates now has load about 1.3k, depends on secondary load. The result looks good, the waveform look good. Let me know what you think.
 

Attachments

  • 811a  350W Class C VTTC amp-1.png
    811a 350W Class C VTTC amp-1.png
    74.4 KB · Views: 34
  • 811a  350W Class C VTTC amp-2.png
    811a 350W Class C VTTC amp-2.png
    71.1 KB · Views: 34
https://youtu.be/uybIo2A6a34

I put the Primary back closer to the Secondary coil along with the tickler coil.
Retuned and experimented with more and less turns. I was able to get stronger results when my 9 turn primary was taped to 5 turns. going lower than 5 turns didn’t produce much results for me… perhaps I need to add a little more in my variable capacitor setup.

At one point I noticed some sputtering sounds & erratic waveforms, I was not able to identify if it was a short somewhere like in my variable capacitor… but upon trying to isolate the sound, the diode in the voltage doubler circuit started to smoke. I can’t seem to reproduce that instability, but I was able to have stronger oscillations than my last setup.

But overall with the new orientation, I was unable to surpass the 15volt mark on my Variac. :( the waveforms just collapse like before. Im starting to think that it may be a component or stray lead somewhere killing the oscillations. There is no way that the mismatched impedance is that far off in all my tests with this 811a setup.


Using the values of components supplied, I put them in the original sch. Found the waveform is no where near sin wave, more like very distorted short impulse of very high voltage. Something wrong with the grid leak circuit.

Im curious to see this bad waveform you are seeing in your simulations… could this possibly be the area for my dilemma?

So I redesign almost completely, separate grid leak and cathode resistor which limit the max. current of each tube. Added the tap in primary coil by splitting into 2 coils. The plates now has load about 1.3k, depends on secondary load. The result looks good, the waveform look good. Let me know what you think.

I also noticed that in your simulations that you put a smaller 50p bypass capacitor at the grid leak resistor as well. Do you think that the shared grids were causing asymmetric loading in your simulations?


A 50k is better load for pentode or transmitting tube like 811, 813 etc, again multiple tap will help to select the best match.

In your post here, you show some different tapped turns in this impedance graph for the totem pole setup. Im having a bit of a hard time comprehending how the lower impedance tap will benefit the current draw… Maybe I’m conflating my short circuit like results (in my videos) with what I’m looking at here with the totem pole setup.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.