Cathode follower - lowering cathode resistor for 12dw7 or ecc823

[Guitarist]

Have you experimented with the cathode follower resistor? I tend to think that encouraging the low Z aspects of this circuit might be good. Amps often insert an effects send on this too. I also suspect that the MV pot might get pushed towards transparency. So if there's an area to make better then I try and do just that!

Also it a rare opportunity for the u7 side to shine. It would be easy to switch in a lower R for going back to a 12ax7 if wanted.

[Tony Bones]

The low output resistance of a CF comes from the tube itself, not the resistor. Roughly, it's 1/gm. And don't believe it when someone tells you that a CF sources current through the tube but sinks current through the resistor. If biased properly, then it sinks and sources through the tube. You could replace the resistor with a CCS and the CF will still work.

On that note, a 12AU7 really doesn't make a very good cathode follower. It's better than a 12AX7, but not as a good as a 12AT7. Better yet, 5687 or 7119!

You don't want to know what I really think of 12AU7s (or 12DW7s) ...

[loudthud]

... And don't believe it when someone tells you that a CF sources current through the tube but sinks current through the resistor. If biased properly, then it sinks and sources through the tube.

Perhaps a picture would persuade you. Have you ever looked at how poorly a 12AX7 drives the tone stack in a 5F6A or Marshall 2203 ? It does OK until the signal gets really big. Then you start to see the low speed on negative transitions and premature clipping on positive transitions.

On that note, a 12AU7 really doesn't make a very good cathode follower. It's better than a 12AX7, but not as a good as a 12AT7. Better yet, 5687 or 7119!

Can't find any reference to 7119. Did you mean 7199 ? The triode in a 7199 is pretty close to a 12AU7. The 5687 has more Gm than a 12AU7, but burns three times the heater current.

You could replace the resistor with a CCS and the CF will still work.

Not much difference really. Too little current available on negative transitions.

[Guitarist]

I have some of those variations and don't mind using more heater current and play little heed to melting a PT- as never have yet. I tend to play with the preamp on 10.
Having to ability to switch in that cathode follower resistor from 22KΩ to 100kΩ might be cool.

[R.G.]

If you want low output impedance and are willing to switch out devices to do it, consider a high voltage MOSFET source follower. A source follower has dramatically higher gain - and hence lower output impedance - than a cathode follower.

No heater current at all.

[Tony Bones]

Perhaps a picture would persuade you. Have you ever looked at how poorly a 12AX7 drives the tone stack in a 5F6A or Marshall 2203 ? It does OK until the signal gets really big. Then you start to see the low speed on negative transitions and premature clipping on positive transitions.

It's slow because the 12AX7 is a low current, low gm tube. The asymmetrical clipping is because the tube is not biased properly. The 5F6A direct coupled design is flawed in that way.

Can't find any reference to 7119. Did you mean 7199 ? The triode in a 7199 is pretty close to a 12AU7. The 5687 has more Gm than a 12AU7, but burns three times the heater current.

The 7119 is similar to 5687, but more of the same thing. Datasheet

Not much difference really [with CCS]. Too little current available on negative transitions.

Because there was too little current at idle. The low output resistance of a CF is all looking up into the cathode. The resistance looking down to ground is large by comparison. If the CF runs out of current then blame the tube, not the load resistor or CCS. If clipping is asymmetric, then correct the bias. Same as with a plate loaded triode.

[Helmholtz]

Have you ever looked at how poorly a 12AX7 drives the tone stack in a 5F6A or Marshall 2203 ? It does OK until the signal gets really big.

I am convinced that the non-linear transfer characteristic of the DCCF is an essential contributor to the sound of many great guitar amps.

Lowering the cathode resistor below 50k with a 12AX7 may force the CF triode into grid conduction even at idle, depending on individual tube characteristics, making the stage even more non-linear.

[nickb]

Have you experimented with the cathode follower resistor? I tend to think that encouraging the low Z aspects of this circuit might be good. Amps often insert an effects send on this too. I also suspect that the MV pot might get pushed towards transparency. So if there's an area to make better then I try and do just that!

Also it a rare opportunity for the u7 side to shine. It would be easy to switch in a lower R for going back to a 12ax7 if wanted.

It's true that some of the other suggested tubes or FETS will work better than a 12AU7, but it not without it's virtues. So, if you are trying to use one up go ahead and save it's feelings. It was a stalwart oscillator tube in TV receivers for years. It can run at substantial currents meaning you can use a much lower cathode resistor and yet still draw no grid current. The price paid is power dissipation, lots of it.

Here's 12AX7 compared with a 12AU7 driving a 10nF cap. I used a 400V supply, grid one at 125V and cathode resistor just 13K. Input is a healthy 200Vpp.



Of course, as has been said, the non-linearity of the original design is a big part of the sound of those amps.