Perfecting Fixed-Bias Cathodyne

[Bruce / Mission Amps]

Yup.
Exactly my point... the real world has that circuit driving real power tubes... you can't ignore the reason you are doing it in the first place.
The photo shot of my scope is simply showing the X10 probe across a high wattage, 8 ohm, nonreactive resistor, right off the output transformer secondary.
Except not a speaker... it's what really happens.

[Enzo]

The cathode swings up to 100V as the plate swings down to 200V so there is a 100V 'gap' between the two which is wasted headroom.

I am still back on this. 100v of wasted headroom? are we expecting the tube to drag its cathode and plate so close together as to make the 100v go away?

[jazbo8]

Its my understanding that higher Rp/Rk will yield higher current capabilities. I can see how 56k would be a suitable choice in my application, but Im still unsure of what the grid voltage should be. I will be experimenting with the software. In the meantime, additional input is much appreciated

I think you meant lower Rp/Rk will yield higher current capabilities, since I=V/R. Does your school not use SPICE for teaching basic circuit analysis? It is a pretty common teaching tool these days at engineering schools... Besides LTSpice, there are many other free simulators available: Win-SPICE, TINA-TI, Simetrix, etc. There is even an online version at CircuitLab.

Jaz

[Dave H]

I am still back on this. 100v of wasted headroom? are we expecting the tube to drag its cathode and plate so close together as to make the 100v go away?

No, not the full 100V, the tube will probably saturate at 30 to 40V plate to cathode meaning there is an extra 60V you could use if you were to bias the cathode closer to 1/4 of B+ (75V) rather than 1/6 B+ (50V)

[loudthud]

No, not the full 100V, the tube will probably saturate at 30 to 40V plate to cathode meaning there is an extra 60V you could use if you were to bias the cathode closer to 1/4 of B+ (75V) rather than 1/6 B+ (50V)

With a 100K (or 112K) load line on a 12AX7, it's closer to 100V. The OP want's to use 2X 100K so 50V is possible.

[Dave H]

With a 100K (or 112K) load line on a 12AX7, it's closer to 100V. The OP want's to use 2X 100K so 50V is possible.

It was between 30 and 40V in the sim. Is it 100V where the loadline crosses the Vg=0 curve? I think in a real circuit the other half of the 12AX7 driving the cathodyne can actually provide a little grid current to drive the grid positive so the plate to cathode voltage can go below 100V.

[loudthud]

The mu of the tube in the sim is about 20. I specified 12AX7 which is what the OP said he used in post #1.

[Dave H]

The mu of the tube in the sim is about 20. I specified 12AX7 which is what the OP said he used in post #1.

I don't understand. I used a 12AX7 in the sim. Perhaps I should get the scope on the real amplifier?

I've simulated a 12AX7 and a 12AT7 now and they were both about 30V

[walkman]

What would be the ideal conditions for a 12at7 with 280v HT ? 56k anode and cathode resistors ? Actually the tube I'm interested in is the 6dx8 triode similar to 12at7 but I think it needs less current? A12at7 might have a bias resistor of 3k , but looking at the data sheet for the ecl84/6dx8 it wouldn't need to draw as much current as the 12at7.

[Wombaticus]

You've evidently set up the sim incorrectly. Here's what the differential output impedance looks like when both outputs are considered simultaneously, which is how the circuit behaves in the real world, so long as the loads remain equal. The guru arguments basically boil down to a silly semantic argument about the technical definition of output impedance -- but in real circuits the two outputs are never operating independently from one another, so it makes no sense whatsoever to assess them individually. The kind of artifacts that Bruce shows reflect well-known real world issues with cathodynes, but these can be largely mitigated by Merlin's recommendations of oversized grid stoppers on the cathodyne itself, small output coupling caps, and larger than standard grid stoppers on the output tubes.



EDIT: Note that the sim is set up to inject 1A of current out-of-phase at each output, so the number of volts on the y axis of the graph corresponds exactly to the number of ohms of output impedance at the nodes shown (output_A and output_k)

[jazbo8]

You've evidently set up the sim incorrectly. Here's what the differential output impedance looks like when both outputs are considered simultaneously, which is how the circuit behaves in the real world, so long as the loads remain equal. The guru arguments basically boil down to a silly semantic argument about the technical definition of output impedance -- but in real circuits the two outputs are never operating independently from one another, so it makes no sense whatsoever to assess them individually. The kind of artifacts that Bruce shows reflect well-known real world issues with cathodynes, but these can be largely mitigated by Merlin's recommendations of oversized grid stoppers on the cathodyne itself, small output coupling caps, and larger than standard grid stoppers on the output tubes.

[ATTACH=CONFIG]22806[/ATTACH]

EDIT: Note that the sim is set up to inject 1A of current out-of-phase at each output, so the number of volts on the y axis of the graph corresponds exactly to the number of ohms of output impedance at the nodes shown (output_A and output_k)

No need to beat this thing to death - we all agree that under real-world operating condition, it's the difference between the two outputs that matters. Like you say it boils down to silly semantics, but I guess some of us are not comfortable to say that "Zp is equal to Zk, period". BTW, the circuit works whether Zp is equal to Zk or not - take the most extreme example, disconnect one of the outputs, so Zx is open, or Zx >> Zy.

Jaz

[Dave H]

I am still back on this. 100v of wasted headroom? are we expecting the tube to drag its cathode and plate so close together as to make the 100v go away?

You are correct. I was wrong. It won't pull the cathode and plate to 40V as I said. The simulation I was using is nothing like the real 12AX7s I have now measured on the scope. I apologize for posting incorrect data. I’ll have to try and correct the 12AX7 model

[jazbo8]

But I thought you had the sim matched up perfectly with the amp before? Which model are you using, Steve Bench's? It should pretty accurate.

Jaz

[Dave H]

But I thought you had the sim matched up perfectly with the amp before? Which model are you using, Steve Bench's? It should pretty accurate.

Jaz

The model is NH12AX7 from Duncan Munro’s DMTRIODEP.INC file.

I thought it was matched but I obviously hadn’t checked it thoroughly enough (I did it a while ago). I should have done it again before I posted. It’s strange. If you run the sim with the tube connected grid to cathode it sits as you’d expect from the data sheet for Vg=0 (100V plate to cathode) but when you drive it hard with a sine wave or connect the grid to B+ with a high value resistor it drops to less than 40V (the grid is driven positive). The three real 12AX7s I tried don’t do that. I’m done with simulation. I don’t trust it. The real plots are below.

Cathodyne 3M3 1.pdf
Cathodyne 3M3 2.pdf

[Bruce / Mission Amps]

The model is NH12AX7 from Duncan Munro’s DMTRIODEP.INC file.
... The three real 12AX7s I tried don’t do that. I’m done with simulation. I don’t trust it. The real plots are below.

[ATTACH]22817[/ATTACH]
[ATTACH]22818[/ATTACH]

Like I mentioned.. sims are cool but isn't real life amazingly educational!! ha ha