Partial feedback (anode grid) with cathode follower

[swordsman]

Reading Gary Pimm Tabor and John Broskie article about anode grid feedback, aka Partial Feedback scheme, I'm thinking of trying something like this.

Using kt88 as output tube, a feedback resistor from it's plate to the driver tube (cascode) plate. This resistor will be the plate load for the driver, RC coupled to the next stage, very much like what JB article spelled out. Here come the question, can I add a cathode follower between the driver and the output stage?

The cathode follower be be direct coupled to the kt88 like the ancient cathode drive class AB amp. My purpose is to use the cathode drive to eliminate the need of grid resistor and provide more drive current. Will this work?

 

[Alastair E]

CF isnt inverting, so should be ok....

 

[swordsman]

The thing is, with this feedback scheme the output tube has a very low input impedance therefore will take in lots of current. Whereas there is little voltage swing at the driver plate. According to John Broskie, it is a V-I converter at the driver and I-V converter at the output tube. Im wondering if sticking a cathode follower in between will work?

 

[Wavebourn]

Here come the question, can I add a cathode follower between the driver and the output stage?

The cathode follower be be direct coupled to the kt88 like the ancient cathode drive class AB amp. My purpose is to use the cathode drive to eliminate the need of grid resistor and provide more drive current. Will this work?

Yes. Such a way you eliminate "Capacitor Sound" that actually is the result of non-linear load of the coupling capacitor. Of course, feedback should stay as is: from anode (plate) of output tube to anode of the upper triode in cascode. It is fine to load cascode on low resistance.

However, you will need a source of significant negative voltage to drive grid of Kt-88 negatively, about 3 times more than was needed for negative bias.

 

[swordsman]

Wavebourn, thanks for replying.

Can you explain the rational why the need for 3 times the negative bias?

 

[Wavebourn]

Wavebourn, thanks for replying.

Can you explain the rational why the need for 3 times the negative bias?

Don't you see that if through capacitor AC applied to the grid develops twice of negative bias on negative peaks, while positive peaks go to zero volt?
In case of cathode follower this negative voltage is limited by negative supply voltage. Even if CF is loaded on CCS you can't get it more than negative supply minus voltage drop on CCS, before nasty clipping. For some margin you need more.

Here is an example: http://wavebourn.com/forum/viewtopic.php?p=25729#p25729

 

[swordsman]

I see. I was thinking there will not be much voltage swing at the driver plate since the principle is to swing current there. Say the output tube need 6ma input for full output. Then the cathode follower will need to swing the bias voltage plus 6ma.