Transformer coupled cathode follower power?

[Osvaldo de Banfield]

I would try a sweep tube as a pentode connected output follower - should work better than a 6L6 due to its lower internal losses, higher transconductance, etc..

The majority of sweep tubes had been designed to carry large amounts of current in short periods of time, between cutouts times. Also, anode/cathode distance is lower than in most pentodes. From this, the class A operation of them is usually a fiasco because the low plate voltage required to not exceed the anode discipation rating.

 

[piano3]

The published anode dissipation rating for sweep tubes is often rather conservative, though, considering the size of their anodes.

 

[Osvaldo de Banfield]

I'm not so sure. Several years ago I tried to make a Class A amplifier using a 6DQ6 to modulate another one used as Class C power output at 3500KHz @ 500V plate voltage, and when I used screen voltage more than 75V, it became red hot, so no more than few watts of output I could take from it. At this time, I haven't oscilloscope yet, so I don't know which the level of distortion I had in it, but for vocal modulation it wan insufficient.

 

[piano3]

Oscillation? I think some forum members have used 6DQ6 successfully.

 

[Osvaldo de Banfield]

Perhaps, but I doubt, if it would be oscillating, I would see it in the cathode's current indicator (Class C emitters almost always has any way to see the current trough it, and using it in the cathode not only it is more suitable from the security point of view, it also permits to adjust grid excitation and plate current). When you modulate the class C amplifier, usually a small increase in plate current can be observed, moreover in a pentode whose screen is also modulated.

Please also note than 3500KHz is the novice 80mts amateur band, now unusable because of the mhz range switching power supplies and multitude of unshielded low cost electronic apparatus. But about 1987 when I started ham radio, it was still listenable.

 

[Sodacose]

I think I'm going to re-work this with a couple of 9 pin tubes at the input rather than the two 6SN7s. That will give me a little more leeway to play with the choke-loaded drivers (6CG7, 12BH7, 5687 all look promising) and input tubes.

 

[Sodacose]

Here's the schematic rejiggered for a 12BH7A:



Lots of options for the input valve and I haven't decided on anything yet. I have everything on that list so I'll probably try more than one and see what I like the sound of the most.

edit: oh yeah, I also have a pair of EF86s that might be interesting...

 

[Lingwendil]

Lots of options for the input valve and I haven't decided on anything yet. I have everything on that list so I'll probably try more than one and see what I like the sound of the most.

edit: oh yeah, I also have a pair of EF86s that might be interesting...

This thing is just begging for fixed bias- and it would be easy to add a negative rail with a voltage multiplier of some sort off of one of your power rails usually... what windings do you have on your transformer?

If it was me, and I needed the voltage swing, I would go with a 12AT7, and either a 6CG7 or the 12BH7 you already have, and replace your output tube with a sweep tube like a 6AV5GA. run the screen at 175 volts or so, and go fixed bias. should be a nice stiff output stage and give more power, with better output impedance.

 

[Sodacose]

My power supply has a regulated B+ (variable 275-375V with decent current capacity) and a 6.3VAC heater. I built it on a separate chassis as a learning exercise, and that resulted in fewer possible power configurations for the audio portion of the circuit. The umbillical will feed the B+, ground, and 6.3VAC heaters (only four connections).

Pic:


I like the idea of fixed bias, but I'm not sure if it will be practical as I have the power portion configured. I could run the 6L6GC (or a sweep tube) as a pentode, but I worry that I'm already pushing the Vhk. Fixed bias would change that, of course.

I'm learning why we design the power supply to meet the amp needs and not the other way around I can always fall back on a regular single-ended design (would make for a pretty extreme 6V6 SET).