Question about common anode mixer

[Craftyjam]

Hi there, experimenting with 6g13a circuit.

I'm feeding a common anode mixer with two identical 180 degree out of phase sine waves, and they are not canceling each other out on the output.

Instead of having seperate plate load resistors and mixing resistors, I'm using one solitary plate load resistor.

I'm having trouble understanding why the two sine waves aren't summing to "0" at the output coupling capacitor and are instead one slightly larger wave than at the inputs.

Is it because each plate sees a varying resistance at the peaks and valleys of the wave?

Attached is the schematic. I'm monitoring all waves on my scope.

[dickjonesify]

You’re feeding both grids so both outputs are in phase with each other… right?

EDIT* oh you said you’re feeding two out of phase signals. Hmm…

[passaloutre]

Have you tried rolling a few different tubes? Could be your triodes aren’t perfectly matched.

[ghostinthescope]

Hi, I have some questions for your question for further fun and exploration. an oscilliscope would make fast work of this if you have one handy.

I've never made this circuit but if I was experimenting with it I'd be wondering some of the following things:

can you verify that the voltage peak to peak is the same for both sinewave signals?
would they cancel eachother out with just two mixing resistors (without any amplification)?
do you think the triodes in the tube are balanced? can you test this?
are Sin1 and Sin2 where your sine signals are connected?
if so do the sinewaves pass through coupling capacitors before the points on your schematic?
do the all resistors in one channel have the same reading in ohms as the other channel?
which signal is present at the output? Sine1 or Sine2
how loud is that sinewave signal if you remove the quieter or non existent signal?
does it increase in volume when second wave is removed?
have you tried in1 and in2 instead?
tried a volume pot at each channel to balance the sinewave amplitude?

I'll be interested to see what you learn and learn more from your experiment.

[dickjonesify]

Hi, I have some questions for your question for further fun and exploration. an oscilliscope would make fast work of this if you have one handy.

I've never made this circuit but if I was experimenting with it I'd be wondering some of the following things:

can you verify that the voltage peak to peak is the same for both sinewave signals?
would they cancel eachother out with just two mixing resistors (without any amplification)?
do you think the triodes in the tube are balanced? can you test this?
are Sin1 and Sin2 where your sine signals are connected?
if so do the sinewaves pass through coupling capacitors before the points on your schematic?
do the all resistors in one channel have the same reading in ohms as the other channel?
which signal is present at the output? Sine1 or Sine2
how loud is that sinewave signal if you remove the quieter or non existent signal?
does it increase in volume when second wave is removed?
have you tried in1 and in2 instead?
tried a volume pot at each channel to balance the sinewave amplitude?

I'll be interested to see what you learn and learn more from your experiment.

Excellent response and perfect screen name hahaha 

[ghostinthescope]

Thanks :)

[AlNewman]

I may be wrong here, but if you feed each side through a capacitor, that means you flip each signal 90 degrees.  So you ended up with a parallel gain stage.

[dickjonesify]

I may be wrong here, but if you feed each side through a capacitor, that means you flip each signal 90 degrees.  So you ended up with a parallel gain stage.

Yeah I was wondering something like that. Like in a tremolo oscillator, each resistor/capacitor pair shifts it 60° which is why it takes three to get to 180° but I’m not sure how that’s calculated.

A long tailed pair doesn’t have a cathode bypass cap. I wonder if that’s throwing it off or why do you have it there?

[astronomicum]

I'm having trouble understanding why the two sine waves aren't summing to "0"

No two triodes are perfectly matched. What are your scoped input and output voltages? Is this just a test or is there some reason you would need perfect cancellation?

[ghostinthescope]

I may be wrong here, but if you feed each side through a capacitor, that means you flip each signal 90 degrees.  So you ended up with a parallel gain stage.

Both signals would be shifted 90 degrees in the same direction and still be 180 degrees from one another.

[Merlin]

Two triodes are never perfectly matched, and they are also not perfectly linear, i.e the peaks and valleys are not amplified equally, so they cannot cancel perfectly. It sounds like you're getting a differential gain of roughly unity, whereas the common-mode gain is probably about 50 (for a 12AX7), so you are getting maybe 34dB of cancellation for out-of-phase input signals. Perfect cancellation is asking too much.

[Craftyjam]

Two triodes are never perfectly matched, and they are also not perfectly linear, i.e the peaks and valleys are not amplified equally, so they cannot cancel perfectly. It sounds like you're getting a differential gain of roughly unity, whereas the common-mode gain is probably about 50 (for a 12AX7), so you are getting maybe 34dB of cancellation for out-of-phase input signals. Perfect cancellation is asking too much.

So, the mixing resistors after the harmonic trem triodes in the 6g13a are used to do the actual mixing and cancel out the modulation signal?

 Based on demos I've heard, the 6g13a doesn't have a noticable ticking sound. If I wanted to get close to perfect cancelation, I would have to manually trim the modulation signals until they are being amplified equally. Maybe use the variable cathode bypass circuit from your guitar preamp book..?

Are the separate plate load resistors increasing the balance of the circuit to help alleviate some ticking? It seems like unless fender was using very closely matched tubes, they would run into similar cancelation issues. I'm using a 6AC10 that appears to be closely matched based on my mutual conductance tester.

Any one with a 6g13a that can comment on real world performance of ticking?

[ghostinthescope]

I’ve built that tremolo circuit into an amp. The common anode mixer isn’t there to cancel anything at all. One triode acts as a highpass filter and the other a lowpass filter. The tremolo oscillator runs through a cathodyne phase inverter. And the two 180 degree out of phase tremolo signal modulate the volume of the rc filters at opposite times. So the signal is heard alternating between the two rc filters. If you want to reduce ticking from the oscillator you have to play around with the capacitors after the intensity pot. Or reduce the maximum setting of said pot. But it’s all a balancing act, because you may be attenuating the tremolo signal and at the same time reducing the effect of the cool swirly tone modulation.

[Craftyjam]

I’ve built that tremolo circuit into an amp. The common anode mixer isn’t there to cancel anything at all. One triode acts as a highpass filter and the other a lowpass filter. The tremolo oscillator runs through a cathodyne phase inverter. And the two 180 degree out of phase tremolo signal modulate the volume of the rc filters at opposite times. So the signal is heard alternating between the two rc filters. If you want to reduce ticking from the oscillator you have to play around with the capacitors after the intensity pot. Or reduce the maximum setting of said pot. But it’s all a balancing act, because you may be attenuating the tremolo signal and at the same time reducing the effect of the cool swirly tone modulation.

The out of phase modulation signals should be canceling each other at the node where the two mixing resistors meet, so the triodes with the hp and lp filters will ideally amplify them equally to get most common mode gain like Merlin said. Obviously, fender must have added with separate plate load and mixing resistors for a reason, which I'm suspecting is to increase balance. I've seen the supposed circuit the 6g13 trem is based on, and it instead uses 3 equal resistors in the anode circuit, two connecting to each anode with the other end meeting with the third, which is then connected to HT. So they do double duty of mixing and splitting the load resistor between each triode.

[Merlin]

God gave us circuit simulators for a reason. Here is a 1Hz, 1Vpp LFO signal fed to both circuits (note the anodes are connected together in the second circuit). Both circuits give identical common-mode gain, but you do get superior cancellation with the separate anode resistors rather than the common-anode mixer.

[ghostinthescope]

But how many vpp is the modulating signal? Does it go low enough to cut off the triode? If so then the output wont be able to cancel out the other triode’s positive signal.

[Craftyjam]

God gave us circuit simulators for a reason. Here is a 1Hz, 1Vpp LFO signal fed to both circuits (note the anodes are connected together in the second circuit). Both circuits give identical common-mode gain, but you do get superior cancellation with the separate anode resistors rather than the common-anode mixer.

That's awesome! Does the simulation assume identical triodes?

[Merlin]

That's awesome! Does the simulation assume identical triodes?

Yes, I used identical triodes in sim.