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

I found it's the value of grid resistor and cap that cause the distortion and large current pulse. Using 30k/50p seem to be ok, much the same as if separate grid leak circuit. So I think you can try to change the value only not the circuit for now.

"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?"

To same extend but not as serve as the component values in use, separated one is better as one does not influence the other, and it's combined value is not the same as separate one. A separate grid leak has fixed value free from interference.

"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."

The final load presented to the output is important as maximum power can be only transfer with the output impedance matches or lower than the load applies. So if the pri has 1kv rms on the plate, to out 1A rms, the load is 1k ohms, and if Rp, internal resistance of tubes combined is also 1k ohms, the max. power is transferred, i.e the power on the load is the same power dissipated in the tubes, this the basic consideration. It's easy to get an oscillation but the power out is really unknown, unless you dealt as if design of Class C amplifier in mind. I have a telecommunication background, there are many theories we learn and forgotten because of lack of practice. So if you have a BE and don't practice in 5 years it's all returned to the teachers :). This forum is a great place for practice and test of your knowledge.
 

Attachments

  • 811a  350W Class C VTTC amp-4.png
    811a 350W Class C VTTC amp-4.png
    73.1 KB · Views: 65
  • 811a  350W Class C VTTC amp-6.png
    811a 350W Class C VTTC amp-6.png
    80.6 KB · Views: 65
  • 811a  350W Class C VTTC amp-3.png
    811a 350W Class C VTTC amp-3.png
    84 KB · Views: 53
Last edited:
OK! Great Success!

I found the area to why I couldnt get past 15v on the variac. It was a bad diode in the multiplier. I also confirmed that 1k & 1nF gridleak works well with each tube independently ran, and than together. 30 to 35 volts is the next limit before I start tripping breakers. Perhaps a 30k sized resistor is necessary
The tickler circuit is best when all windings used as close as possible to the secondary now.

it is definitely alot more lively than before now c:

The final load presented to the output is important as maximum power can be only transfer with the output impedance matches or lower than the load applies. So if the pri has 1kv rms on the plate, to out 1A rms, the load is 1k ohms, and if Rp, internal resistance of tubes combined is also 1k ohms, the max. power is transferred, i.e the power on the load is the same power dissipated in the tubes, this the basic consideration. It's easy to get an oscillation but the power out is really unknown, unless you dealt as if design of Class C amplifier in mind. I have a telecommunication background, there are many theories we learn and forgotten because of lack of practice. So if you have a BE and don't practice in 5 years it's all returned to the teachers . This forum is a great place for practice and test of your knowledge.

Ah yes, kind of had to visualize a voltage divider analogy in my mind for a moment. :p Making sense.

I definitely don't have a telecommunications background, but its been a heavy research/hobby area of mine me lately so its nice to catch up on all the practice i've been missing out on. c:
I need to build the things I see to comprehend it. So its nice to walk through these problems here for sure.

In my last successes with my set up, I noticed best performance with all windings used in the primary coil. lower current draw and stronger results than minimal turns tapped. Will this totem-pole circuit you have been simulating bring out any benefits to enabling less turns?
 
Disabled Account
Joined 2013
"I found the area to why I couldnt get past 15v on the variac. It was a bad diode in the multiplier. I also confirmed that 1k & 1nF gridleak works well with each tube independently ran, and than together. 30 to 35 volts is the next limit before I start tripping breakers. Perhaps a 30k sized resistor is necessary
The tickler circuit is best when all windings used as close as possible to the secondary now."

Increase the value of grid resistor reduces the drive width, level as well as the current in the cathode, it also reduces the conduction angle so the current pulse is narrower, which help to reduce overall current drawn. The other current limit can be done by add resistor in cathode, the current can be adjusted to same level and to desire gain level so the dissipation become acceptable.

Not sure you can pick up the drive level at the grid, you can try to measure it, maybe few hundred to few kv, the actual level that matters is the level above the bias level, i.e. < 600V.

"In my last successes with my set up, I noticed best performance with all windings used in the primary coil. lower current draw and stronger results than minimal turns tapped. Will this totem-pole circuit you have been simulating bring out any benefits to enabling less turns? "

This is because you only have primary tuned, it works better with all windings in the pri coil. Try to tune the second also by a torrid cap, you may then find tap works better. You should be able to use less turn (impedance) with totem-pole as it output impedance is less than the // tubes.
 

Attachments

  • Grid leak sim-1.png
    Grid leak sim-1.png
    35.8 KB · Views: 44
Last edited:
Increase the value of grid resistor reduces the drive width, level as well as the current in the cathode, it also reduces the conduction angle so the current pulse is narrower, which help to reduce overall current drawn. The other current limit can be done by add resistor in cathode, the current can be adjusted to same level and to desire gain level so the dissipation become acceptable.

Not sure you can pick up the drive level at the grid, you can try to measure it, maybe few hundred to few kv, the actual level that matters is the level above the bias level, i.e. < 600V.

To see what adding more resistance does to the grid leak circuit, I found a 1k 10W potentiometer laying around to add in series with the existing 1k & 1nF setup. It did help reduce the overall current drawn as you said. but at around (+) 400ohms on the potentiometer, the whole system failed to oscillate. I would tap on the coil with a stick and crank up the voltage to much higher levels to trigger oscillations. When it did decide to oscillate (at 80v on the variac), it hummed loudly and popped some circuit breakers. I also couldnt tell if its my variable resistor was causing the lack of oscillations, I had a minor difficulty measuring the values with a multimeter... made me suspicious for the look of its age.

100pf & 1k didn't produce any oscillations. I had to try it out.


But I need to buy some more resistors/pots to get up into the 30k range to properly test. Which brings me to the question of identifying the Wattage of the resistor. The origional 811a schematic I used had some interesting wattage ratings for when 'increasing' the resistance here. How do you tell what kind of wattage rating you need for this kind of tube?

This is because you only have primary tuned, it works better with all windings in the pri coil. Try to tune the second also by a torrid cap, you may then find tap works better. You should be able to use less turn (impedance) with totem-pole as it output impedance is less than the // tubes.

Raising and lowering or adding capacitance to the secondary / extra coil is nowhere as dramatic in change of results as being able to adjust taps / impedance with the large gang variable cap. I think the 811A needs a few more grid leak experiments before I seal its fate by going to a totem pole design. Im so curious to see how much of a difference it is physically. It just so happens I have two spare 811a tubes. :p
 
Disabled Account
Joined 2013
"100pf & 1k didn't produce any oscillations. I had to try it out."

If you try a large value grid resistor Rg, the grid cap Cg value should also be reduced, since there is a reduced current or voltage to charge the cap and so the bias voltage is reduced the tube conduct more again if RC constant is too low. So you need to use RC constant calculator, RC constant of 1k &1n is same as 30k &30p.

"How do you tell what kind of wattage rating you need for this kind of tube?"

The power is dissipated in both Rg and internal grid of the tube as well. So if the voltage across the grid coil is 150V, and grid current is 50mA with Rg=1k, the total dissipation is 7.5W, I*I*R=(50*50)/1000=2.5W in grid resistor. Grid current will be reduced as grid resistance of Rg is increased, hence it's wattage is less than that of 1k, 2.5W. Need to consult the original data for more accurate grid current flows, or you can measure it or not? Also the current only flow more when it is above 0v not the entire cycle, so the actual power dissipation is less than this, but grid dissipation wattage is a good indication.
 
Last edited:
If you try a large value grid resistor Rg, the grid cap Cg value should also be reduced, since there is a reduced current or voltage to charge the cap and so the bias voltage is reduced the tube conduct more again if RC constant is too low. So you need to use RC constant calculator, RC constant of 1k &1n is same as 30k &30p.

The power is dissipated in both Rg and internal grid of the tube as well. So if the voltage across the grid coil is 150V, and grid current is 50mA with Rg=1k, the total dissipation is 7.5W, I*I*R=(50*50)/1000=2.5W in grid resistor. Grid current will be reduced as grid resistance of Rg is increased, hence it's wattage is less than that of 1k, 2.5W. Need to consult the original data for more accurate grid current flows, or you can measure it or not? Also the current only flow more when it is above 0v not the entire cycle, so the actual power dissipation is less than this, but grid dissipation wattage is a good indication.


RC constant! Thats what that is, Im used to thinking of it being under the cathode typically. I just put an order in for a few variety of resistor sizes, they should be coming in the next couple of business days. Been feeling pretty good about the direction & results now, finally :p

Now gaining back the confidence, I wanna take the learnings from this set up & bring it back to the totem pole design.
x9d7FL4.jpg

With the bipolar supply, I think I may have just realized that in order for the stack to work, both tubes have to be conducting at the same time, I.E. the half wave rectification needs to be smoothed to work. Is this right? or is there a simple step around.

If I have to live with a filtered DC supply, I guess this amp will naturally be able to play music as well.

With that in mind, what do you think of a phase splitting driving stage with a 6SN7? simplify some parts, nix a line transformer? Found it here. Or is the ECC88 the same thing here
 
Last edited:
Disabled Account
Joined 2013
"With that in mind, what do you think of a phase splitting driving stage with a 6SN7? simplify some parts, nix a line transformer? Found it here. Or is the ECC88 the same thing her"

If you're not using transformer you need a balanced driver circuit, so the bottom end of 33k in the cathode of phase spilter should be connected to output, not ground as shown in the referred sch. The introduces some negative feedback, which is fine for audio amp. I have yet to verify it with simulation. Otherwise you need to change the upper 33k to 47k or more as the upper output tube needs more drive level for same output level, yes, there are push pull amp so there are 2 pulses, 90 deg and 270 deg rather than just one at 90 deg. Yes you'll need quite smooth bipolar supply or a single supply with DC blocking cap at the output.
 
Last edited:
Ok drew it up again with a question on the power supply:

I found some cheap selections for capacitors online, and was wondering if this would be good enough to smooth out the ripples enough.
Another question is the idea of using another M.O.T as a massive choke. Ive heard people suggesting this idea, but never heard conclusions on why not to do this. It is a big clump of iron thats mass produced. Seems cheap enough

fvJrJME.jpg


Also, as im drawing this up, I continue enjoying the thought that this can multi purpose as a music amp as well. The only thing that would change is the RC bias values at the grids of each 811a right? But because of the way this totem-pole design draws current, I shouldn't have to push the bias into class C like behavior (conducting at a fraction of the cycle) to reach reasonable efficiencies while preserving the ability to drive speakers for audio?

oh and perhaps the fact that the impedance output with the 811a tubes might be too high to direct drive right? :c

The audio feature would be novel, but if it begins to overcomplicate things, I can do without it. I like the idea of flicking a bias switch to get a different behavior out of this design :)
 
Disabled Account
Joined 2013
"I found some cheap selections for capacitors online, and was wondering if this would be good enough to smooth out the ripples enough.

Whatever electrolytic cap you used make sure they're tested before put them in use. You need to match the cap by their leakage current very closely so the voltage are equal across each cap. This is done by measure the voltage in resistor series with cap connected to power supply. As you remove the supply, the capacitance should be retained for days, the one gone down too fast is not good.

Another question is the idea of using another M.O.T as a massive choke. Ive heard people suggesting this idea, but never heard conclusions on why not to do this. It is a big clump of iron thats mass produced. Seems cheap enough"

Choke needs lot of lamination to produce high inductance at low freq, lump of iron makes a low efficiency choke, and heat since it is not as efficient, IMO. Better to take apart a PSU transformer and rearrange the EI lamination.
 
Last edited:
Disabled Account
Joined 2013
Also, as im drawing this up, I continue enjoying the thought that this can multi purpose as a music amp as well. The only thing that would change is the RC bias values at the grids of each 811a right? But because of the way this totem-pole design draws current, I shouldn't have to push the bias into class C like behavior (conducting at a fraction of the cycle) to reach reasonable efficiencies while preserving the ability to drive speakers for audio?

oh and perhaps the fact that the impedance output with the 811a tubes might be too high to direct drive right? :c

The audio feature would be novel, but if it begins to over complicate things, I can do without it. I like the idea of flicking a bias switch to get a different behavior out of this design.


The optimal load would be about 1K. So you can't connect 8 ohms speaker directly without a step down transformer. Extra bias needs not be used as zero bias is fine for 811a (not for 6c33c though). There is a 200W audio amp in RCA 811a datasheet, which shows how it's done.
 
is 5H on each side too skimpy for my application?
What about at 150ma.
Looking for areas to cut cost when I can :p Found a good array of choices for 5 to 10H sizes.

I may just get a bunch of new capacitors from mouser, there are some cost effective choices there at those voltages.

I may end up making a separate 6c33c amp after this totem pole 811a design
 
I got a couple of 10k 10w resistors in the mail today so I tried it out and got some improved results. Improved results as in being able to go up to the 60v - 80v line on my variac before I started getting some light arcing in my tuning capacitor. Moved to my home made cup capacitors and saw improvement in stability when not arcing over.
The grid resistors get pretty darn hot after some good 5 min running Though :eek:
 
Disabled Account
Joined 2013
is 5H on each side too skimpy for my application?
What about at 150ma.
Looking for areas to cut cost when I can :p Found a good array of choices for 5 to 10H sizes.

I may just get a bunch of new capacitors from mouser, there are some cost effective choices there at those voltages.

I may end up making a separate 6c33c amp after this totem pole 811a design

If you want to simulate the PSU, try to download PSU2, that will help to see if the components meet the criteria.
 
Thats a good resource. I had to get out my old windows partition out to see the software but its working well now.
I realized that my last drawing has capacitor placement designed to create inrush current.

Im still playing with it and other downloadable samples to get familiar with it
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.