Class AB2 P-P Design

[andrewRneumann]

Hello fellow nerds and geeks,

I've been trying to design a P-P power amp that drives a pair of 6V6S's (triode-strapped) as far into Class AB2 as possible. My research convinced me that the CF drivers need to be direct-coupled to the grids of 6V6 to provide the pulses of current required and also avoid blocking distortion. This causes quite a lot of havoc when it comes to biasing the whole setup.

There are different ways and places the bias could be applied, but it seemed best to apply a negative bias on the tail of the CF driver. This would have to be a low impedance bias source because it pulls DC while in operation, so that pretty much forces me to use the 50v (70VDC) bias tap directly (with diode and smoothing cap) with no voltage divider network.

Then to get the correct bias on the power tube, I had to go ahead and add some tail resistance. So this makes the power tube have a mixed fixed/cathode bias.

I'm totally in the weeds on this, and hoping what I've designed makes sense to somebody out there and my errors can be pointed out. Attached diagram is just one half of the P-P setup. Also if anyone can point me to a working schematic with a similar design, I'd much appreciate it.

 

[NTC]

From the not very helpful department: you could do this like they did in the early days of electronics and use an interstage coupling transformer. As an example look for a datasheet for a 19 tube. Of course, you would have to wind your own transformer.

 

[andrewRneumann]

From the not very helpful department: you could do this like they did in the early days of electronics and use an interstage coupling transformer. As an example look for a datasheet for a 19 tube. Of course, you would have to wind your own transformer.

Well... I was reading this article from a 1945 June issue of QST and it looks like it can be done without a transformer which looked really cool. I thought I'd try my own take on the circuit as Class AB2. I couldn't figure how they made the bias work on this one, so I tried to figure it out from scratch on my own. I just have no idea if my idea above would actually work.

 

[andrewRneumann]

I'm making modifications to this circuit. I'm trying to implement what Aiken speaks about here.

Add a DC-coupled cathode follower between the phase inverter and the grid of the output tubes, with the cathode follower cathode resistor returned to a high negative voltage, and the grid bias applied to the grid of the cathode follower. This effectively isolates the output tube grid circuit from the phase inverter and its associated AC coupling, and provides a very low impedance source for the output stage. This will prevent the output stage from going into grid clamp, and will eliminate the long time constant of the AC coupling. This method has the unfortunate side effect of requiring an extra tube and completely ruining the value of your vintage amp, so it is best used only on new designs, but is highly recommended. You will also get more power out of the output stage because it is now running in class AB2 or class A2 (the "2" suffix indicates grid current flows for a portion of the cycle).​

I think I'm getting close to a workable design that eliminates the cathode resistor/cap on the power tube. However, I need more negative bias on the cathode of the follower. So I'm looking at something new for me: voltage doubling circuits. If I double the 50vac bias tap I figure I can get about -138VDC available, which is more than enough. What I'm worried about is that this circuit, unlike a regular bias circuit, needs to carry a substantial amount of DC and AC current, and I'm reading that voltage doublers have poor regulation. But how poor? I estimate I will have 13mA of DC current and peak AC current of 25mA running into my voltage doubler. How can I tell whether it can handle this kind of load?

 

[Lowerleftcoast]

So I'm looking at something new for me: voltage doubling circuits. If I double the 50vac bias tap I figure I can get about -138VDC available, which is more than enough.

Cant you skip the 50v bias tap and just use one leg of the PT HT for higher voltage?

 

[2L man]

Voltage doubler circuit efficiency to deliver current is very poor. You need to test the circuit loading it to a current you need.

If you don't need to use a tube rectifier there are more power transformers which have multiple outputs. Or you could use a small toroid to do powerful enough negative?

Antek 50VA cost $28 and lowest HV is 120VA. 100VA are $38 and there are many. Naturally there is no need to have filament output on small transformer but if your test does not work you can use it somewhere else. HiFi builders sometimes seem to stack toroids on top of each other so they should fit quite fine to the space an EI core PT need?

I follow your design with high interest but just earlier today I did order (about $32) Hammond push pull Interstage to be used on AB2.

 

[andrewRneumann]

Cant you skip the 50v bias tap and just use one leg of the PT HT for higher voltage?

That's an interesting idea. I will put pencil to paper and see if it works out. Maybe I will ground the plate of the cathode follower and suck the current through from the cathode!

Voltage doubler circuit efficiency to deliver current is very poor. You need to test the circuit loading it to a current you need.

I think you may be right there. I will have to try it unloaded, and then wire in a dummy resistor that should give me around 25mA and see how far the voltage drops. As I type this, I just realized that I will probably have to fire up LTSpice--and I'm dreading it already.

 

[andrewRneumann]

Here is where the current design stands.
Changes:
1. Applied twice the negative voltage to tail of CF.
2. Removed cathode bias from power tube. (Bias to power tube controlled by bypassed resistor on cathode of drive).
3. Bias of drive controlled by voltage divider network.
4. Added build out resistor and grid stopper to coupling circuit. These are the maximum values recommended by Merlin. I want to minimize this resistance, but don't know if it will be possible and keep from oscillating. This will require experimentation. My calculations show that the 12BH7 can still push enough current to drive the 6V6S triode into Class 2 even with the extra 1K7 of impedance.
5. Pot in coupling circuit will be used to experiment with changing output power and overdrive character. Since a big chunk of the power this amp will produce comes from getting into Class 2, if we cut-off that Class 2 current, we reduce the power of the amp. I calculate 12W in Class AB2, but if we increase the grid resistance, that can be reduced to 4.7W or anywhere in between with the flick of a pot. I can't wait to see if this will work.

 

[2L man]

I'm afraid that circuit bias is not stable enough? Only potentiometer I see is R1 and there it won't do much adjusting?

If LTspice produce voltages which have "point something" accuracy, in actual circuit they can be what ever, because of component tolerances.