Help? 6n23p cathode follower

[ericj]

I recently bought a pcb to build a hybrid tube/chip power amp. 6dj8 type cathode follower buffer feeding a pair of lm3886. Clearly the tube buffer is meant to add warmth and even-order harmonic distortion to the chip amp. My supposition is that the tube is meant to be nearly unity gain.

The relevant portion of the schematic is attached. (What? I can't attach photos?)



The designers, since they are chinese, designed around the chinese 6n11 tube. I can't find curves for this tube, but the vague internet rumors are that it's a chinese 6n1p, which would mean that it's pin compatible with a 6dj8 type but not really a 6dj8.

I've no intention of buying a 6n11 for $15. I have some 6n1p, 6dj8, and 6n23p here. I'm hoping that the 6n23p is a good choice for this application.

I've built a few valve amps based on other peoples designs but i am still essentially a tube dunce. I'm capable of not killing myself with high voltage but i don't know how to calculate the resistor values around a tube to set the operating point.

So, here is where i ask for help: Assuming that a 6n23p is a good match for this application, what are the ideal resistor values for R7, R8, and R9 at 90vdc b+? What should i consider adjusting if B+ turns out to be closer to 80v or closer to 100v?

Would i be better off with a more run-of-the-mill 6dj8 type? I've heard that 90v isn't really a great voltage for the 6n1p.

[ruffrecords]

Some simple math should get us in the right ball park. With a 90V supply you want the cathode to be around half that for maximum output swing - say about 45V. With 33K in the cathode lead the tube current is not likely to be much more than one milliamp (45/33) which is way too small for a 6DJ8 in a cathode follower. You should aim for something like 10mA so you could change the 33K to 4.7K. Trouble is, looking at the 6DJ8 anode curves, that would mean a rather low grid bias voltage mainly because the voltage across the tube is so small. So let's try 5 mA instead and use a 10K resistor. from the anode curves, 45V across the tube and 5mA current requires a bias of -1V - that's more like it. We now need to change the 470R resistor to drop 1V at 5mA - R = V/I =1/5m (K) = 200 ohms.

So try changing the 33K to 10K and the 470 to 200 ohms for a 6DJ8. The 6N23P is pretty close to this so the same values should work.

Cheers

Ian

[ericj]

Thanks for the suggestion.

I should have also asked for a pointer to a good website or book to learn more about this stuff, in terms that someone bad at math can understand. I've seen recommendations for Morgan Jones' book more than once - would you agree that it's a good place to start?

About the b+ -- the schematic doesn't specify it (I added that) but the board is configurable for either 65vac tube supply (= about 90 volts after rectification and smoothing) or to use the full wave of the dual supply for the chip amp.

I have selected a 25-0-25-0 toroid for the chip amp supply which would give me about 70v, which is almost certainly too low, so I am considering an off-board voltage doubler on the top half of the AC giving me closer to 100-105v.

It sounds as though i would not likely have to adjust your suggested resistor values for 15v greater b+. All the HV caps on the board are rated 160v so i have no worries there.

[SY]

Originally Posted by ericj

I've seen recommendations for Morgan Jones' book more than once - would you agree that it's a good place to start?

+1000.

Also see the cathode follower article ("The Heretical Preamp") on my website.
SYclotron Audio The Heretical Preamp

[ericj]

Originally Posted by ruffrecords

So let's try 5 mA instead and use a 10K resistor. from the anode curves, 45V across the tube and 5mA current requires a bias of -1V - that's more like it. We now need to change the 470R resistor to drop 1V at 5mA - R = V/I =1/5m (K) = 200 ohms.

So try changing the 33K to 10K and the 470 to 200 ohms for a 6DJ8. The 6N23P is pretty close to this so the same values should work.

Wait. R9 in the above schematic is 470K - corroborated by the BOM and the printing on the pcb.

Did you mean 200R or 200K?

[SY]

I think he was talking about R7 and R8.

[ericj]

Originally Posted by SY

I think he was talking about R7 and R8.

then where did he get 470 from?

[SY]

Easy to misplace numbers on a schematic. In any case, use his suggested resistor values in the cathode circuit (R7 and R8). The grid resistor is less critical, it can be anywhere from 10k-1M. It's bootstrapped in this circuit, so you can make the coupling cap a lot smaller if you keep the 470k grid resistor.

[ericj]

ok, makes sense i guess

[ruffrecords]

Originally Posted by ericj

Wait. R9 in the above schematic is 470K - corroborated by the BOM and the printing on the pcb.

Did you mean 200R or 200K?

I did not mention R9. I was talking about the 330 ohm resistor R7 which I mistakenly wrote down as 470. My apologies. There is no need to change R9.

And by the way, Morgan Jones book is ideal for beginners.

Cheers

Ian