1'st stage of PA

[baudouin0]

I am designing a first stage for a new PA I am building. I need a gain of about 10x and fairly low impedance output approx 10k. The output swing needs to be about +/-10v for full output. This is quite large as the GFB is taken to the second stage. The distortion needs to be quite low < .02%

I read the Mu verses CF with much interest but its quite lengthy.

The circuit I have come up with seems the best, but are open to suggestions which I can simulate. Its stereo so the other halves of the valves are used. I would have prefered no FB to the input but as the volume control is 10k that's OK. I seems that high gain with NF gives the best linearity.


I plan DC heater U1 and ac elevated (with caps) for U2. I think by using 12.6v the caps will be smaller.

 

[Ketje]

It looks like you like simulating
Try this one, performs perhaps better. Current needs ~5mA
Mona

 

[baudouin0]

It looks like you like simulating
Try this one, performs perhaps better. Current needs ~5mA
Mona

Yep that seems to work just as good without the bootstrap. I've gone back to the original input arrangement as that version seems to increase gain at 5Hz rather than rolling off. The key as you have pointed out is to run the first stage at much lower current. The 300V will be regulated, 10uF is the biggest mk2 I can get.

Many thanks I think that what I will prototype.

 

[Chris Hornbeck]

Since you're doing an anode follower, you might put the volume control in the feedback, a la Boegli.


All good fortune,
Chris

 

[baudouin0]

Many thanks, the volume control is in the pre-amp output connected via phono to the amp. Sorry I did not mention that!

 

[baudouin0]

It looks like you like simulating
Try this one, performs perhaps better. Current needs ~5mA
Mona

I've prototype the design and works perfectly. Heater is DC on first stage and elevated on the second hence the caps.








Next the VAS stage.

 

[Osvaldo de Banfield]

I prefer the neon lamp schema. Also, it seems to me that 5KΩ for a 12AX7/6AV6 is a bit high. Usually 2.2 to 2.7KΩ is there.

 

[baudouin0]

Sorry I used a diode. Should never conduct once warmed up. The plate has 330k so I do need about 4k7. The idea is to get as much voltage gain as possible to minimise distortion with the NF. The measured plate voltage is 165 and 162v on U1 which is very close to simulation.