Pardon me if this is the wrong place to post. Still getting used to forum. A few days ago I posted about a 6L6 and the expected power for a PP design that I cobbled together. I came back and added a 'scope trace that shows some sort of small step voltage change near the peaks of the waveform. Perhaps the edit/addition went unnoticed? The edit is at the top of the thread, to my original post. The original thread is here:
https://www.diyaudio.com/community/threads/help-with-a-6l6-pp-design.415228/#post-7739172
It includes the circuit schematic, scope trace and explanation. Thx
https://www.diyaudio.com/community/threads/help-with-a-6l6-pp-design.415228/#post-7739172
It includes the circuit schematic, scope trace and explanation. Thx
What is the signal source you are using?
Some Function Generators have a glitch, at or near the crest of the signal.
My Denon Audio Technical test CD, plus my CD player do not have any such glitches.
Disconnect the negative feedback from the output transformer secondary (caution, the amplifier gain goes up).
So reduce the amplitude of your signal source, so that you can re-test (this time with the negative feedback unconnected).
Some Function Generators have a glitch, at or near the crest of the signal.
My Denon Audio Technical test CD, plus my CD player do not have any such glitches.
Disconnect the negative feedback from the output transformer secondary (caution, the amplifier gain goes up).
So reduce the amplitude of your signal source, so that you can re-test (this time with the negative feedback unconnected).
Floating paraphase requires the feedback connection. In addition to FB, I think it is used to provide a signal path to ground. I do know for sure you can't remove the FB altogether, the circuit will not work. I have checked the sig generator. It's actually quite nice. And lastly, the step in the waveform goes away at higher frequencies, but I can see a slightly distorted sin wave starting where the step was. I'm thinking the problem is mismatched output tubes. The discussion of that is happening at the original thread.
I forgot to look at the schematic, to remember the negative feedback detail.
Disconnect the feedback resistor from the output transformer secondary, and connect that end of the resistor to ground.
Now, the paraphase circuit will still work.
Of course, I tried paraphase, but was not impressed with it.
I use a CCS, coupled cathodes, and very well matched plate loads to do my phase splitting.
The gain is 1/2 of the paraphase splitter, not a disadvantage if you have plenty of signal going to the CCS coupled cathode phase splitter.
Also, the peak to peak swing is 1/2 as well, but if there is plenty of swing for the output stage, that is not a problem either.
All of the amplifier designs that I am using now are:
Cathode coupled CCS splitter, and output push pull
Balanced push pull with balanced input
Self inverting push pull, with a CCS in the coupled cathode push pull output stage (single ended drive to one output tube; other output tube grid is tied to a stopper resistor to ground.
Disconnect the feedback resistor from the output transformer secondary, and connect that end of the resistor to ground.
Now, the paraphase circuit will still work.
Of course, I tried paraphase, but was not impressed with it.
I use a CCS, coupled cathodes, and very well matched plate loads to do my phase splitting.
The gain is 1/2 of the paraphase splitter, not a disadvantage if you have plenty of signal going to the CCS coupled cathode phase splitter.
Also, the peak to peak swing is 1/2 as well, but if there is plenty of swing for the output stage, that is not a problem either.
All of the amplifier designs that I am using now are:
Cathode coupled CCS splitter, and output push pull
Balanced push pull with balanced input
Self inverting push pull, with a CCS in the coupled cathode push pull output stage (single ended drive to one output tube; other output tube grid is tied to a stopper resistor to ground.
I'd add that your open loop gain is very low, compared to the similar Dynaco schematic, which uses 12AX7 or 6SL7 instead of 6SN7. Those tubes have a much higher mu, resulting in higher open loop gain, allowing the negative feedback to better do its job.
Hi @wstein25 you have done well!
As regards the step, it may not be the output tubes making this?
Its possible that it is getting in from somewhere, have you checked the input from your signal generator on the scope? Also have you had a look at the Grids (pin 5) of the two 6BG6 tubes?
It could also be some sort of high frequency oscillation that gets going for the 6BG6 tubes? (unlikely). High frequency not seen by your scope.
It could be a breakdown at high voltage?
I don't know the sense of which tube is which let's say the worst "kink" get's transformer coupled into the other tube so alternately tubes are producing the kink when Plate voltage gets high? Some sort of bias problem?
So the kink only occurs once per cycle per tube but it appears to be twice per cycle per tube but its not - its transformer coupling. So the kink is voltage dependent.
Find out if the full kink happens when the Plate is reaching maximum voltage or when the Plate is reaching minimum voltage. Maybe the Plate is getting too low and hitting Ground?
Have you checked what's happening on the 440 Volt and 268 Volt lines. Is the screen voltage stable? Try triggering the scope on a 6BG6 Plate and looking at the power supply lines with the other probe, look for that kink.
PS the HT B+ power supply is a voltage doubler and the capacitive decoupling is minimalist - wouldn't be surprised if you are getting 50Hz and 100Hz hum.
The 268 Volt line only as a 40uF capacitor to hold it up, the 16K resistor can only provide 10mA - its probably OK but worth a look.
Be careful stay safe with those high voltages.
As regards the step, it may not be the output tubes making this?
Its possible that it is getting in from somewhere, have you checked the input from your signal generator on the scope? Also have you had a look at the Grids (pin 5) of the two 6BG6 tubes?
It could also be some sort of high frequency oscillation that gets going for the 6BG6 tubes? (unlikely). High frequency not seen by your scope.
It could be a breakdown at high voltage?
I don't know the sense of which tube is which let's say the worst "kink" get's transformer coupled into the other tube so alternately tubes are producing the kink when Plate voltage gets high? Some sort of bias problem?
So the kink only occurs once per cycle per tube but it appears to be twice per cycle per tube but its not - its transformer coupling. So the kink is voltage dependent.
Find out if the full kink happens when the Plate is reaching maximum voltage or when the Plate is reaching minimum voltage. Maybe the Plate is getting too low and hitting Ground?
Have you checked what's happening on the 440 Volt and 268 Volt lines. Is the screen voltage stable? Try triggering the scope on a 6BG6 Plate and looking at the power supply lines with the other probe, look for that kink.
PS the HT B+ power supply is a voltage doubler and the capacitive decoupling is minimalist - wouldn't be surprised if you are getting 50Hz and 100Hz hum.
The 268 Volt line only as a 40uF capacitor to hold it up, the 16K resistor can only provide 10mA - its probably OK but worth a look.
Be careful stay safe with those high voltages.
I find starting with NO negative feedback in testing and getting the amplifier working well with no feedback is much easier.
Then add the negative feedback.
That way feedback issues and basic amplifier design issues can more easily be separated.
I also find first fully testing the input stage(s) with the output tubes removed and getting that working really well very helpful.
Clipping or odd input stage behavior mixed with output stage issues can be pretty confusing.
Divide and conquer.
Taking a complete amplifier design with all feedback enabled and trying to debug it has proved unproductive for me.
Then add the negative feedback.
That way feedback issues and basic amplifier design issues can more easily be separated.
I also find first fully testing the input stage(s) with the output tubes removed and getting that working really well very helpful.
Clipping or odd input stage behavior mixed with output stage issues can be pretty confusing.
Divide and conquer.
Taking a complete amplifier design with all feedback enabled and trying to debug it has proved unproductive for me.
The 6BG6 is very much a 807, cost reduced for the TV industry in the 1950s
It was also know to be prone to snivets like the 807 is.
Snivets are a oscillation caused by an area of negative plate resistance in the tube anode.
I found with a 807 they show up often at about the 75% power level just as one one tube is being fully cut off in current while the other tube reaches toward maximum current. I admit my understanding of the exact nature of snivets is a bit fuzzy.
See
https://dpnwritings.nfshost.com/ej/beam_power_tube_parasitics/
You have two of the common cures for snivets a grid stopper and a screen resistance so it may not be your issue.
However a 50 ohm plate resistance right at the cap was also commonly advised for snivets.
The physical layout of the amplifier strongly affects snivets so it is hard to predict when they will happen.
Happy hunting!
It was also know to be prone to snivets like the 807 is.
Snivets are a oscillation caused by an area of negative plate resistance in the tube anode.
I found with a 807 they show up often at about the 75% power level just as one one tube is being fully cut off in current while the other tube reaches toward maximum current. I admit my understanding of the exact nature of snivets is a bit fuzzy.
See
https://dpnwritings.nfshost.com/ej/beam_power_tube_parasitics/
You have two of the common cures for snivets a grid stopper and a screen resistance so it may not be your issue.
However a 50 ohm plate resistance right at the cap was also commonly advised for snivets.
The physical layout of the amplifier strongly affects snivets so it is hard to predict when they will happen.
Happy hunting!
Last edited:
Thanks for your suggestions. I figured it out last week, and the discussion of the solution is in the original thread. By adding a small amount of high frequency roll off to the feedback loop everything cleared up. I had tried snubber capacitors on the output transformer, but that was not particularly stable. Operation would change just by attaching a scope probe. The link to this is at the top of this discussion. BTW, I was under the impression that a 6BG6 was a top connected 6L6.
wstein25,
Capacitors to ground from the primary's plate leads creates a high frequency, high Q resonator.
The capacitor resonates with the primary to the secondary's Leakage Inductance.
That is why you will see snubbers in plate circuits, and snubbers in B+ HV secondary circuits, are most often Series RC . . . which allows the Q of the resonance to be reduced.
And resonance, gain, and negative feedback (negative feedback which changes phase whenever a plate snubber capacitor is used) . . .
Yes, especially when a capacitor-only [non RC] snubber is used.
Unstable indeed, and perhaps it even causes a power oscillator instead of an amplifier.
Rules are meant to be broken, but often the results are not what is desired.
Capacitors to ground from the primary's plate leads creates a high frequency, high Q resonator.
The capacitor resonates with the primary to the secondary's Leakage Inductance.
That is why you will see snubbers in plate circuits, and snubbers in B+ HV secondary circuits, are most often Series RC . . . which allows the Q of the resonance to be reduced.
And resonance, gain, and negative feedback (negative feedback which changes phase whenever a plate snubber capacitor is used) . . .
Yes, especially when a capacitor-only [non RC] snubber is used.
Unstable indeed, and perhaps it even causes a power oscillator instead of an amplifier.
Rules are meant to be broken, but often the results are not what is desired.