Hey thanks! You could always copy the approach if you want the same results!
I started a new thread here on the completed amp design where I will present detailed measurements of the final configuration.
Yeah, I could. I still have a 35T breadboard (that started out as Michael Koster's Meteor), sitting on one corner of my bench, it would be fun to try your approach but . . . . . . I did mean it when I said I'd like to hear what you think after you've had the amp in your listening room for a few months .
In my experience, some nits don't become apparent for a while. If you say you really like the sound I'd be game to see if I can make it work with the 35T.
Right now, I'm working on bringing a different circuit around to what I hope will some day be worth talking about. It sounds very good on the bench speaker but doesn't look so good on a scope and I'd like to have both up to snuff. . . . . . . . . so the timing of your posts is motivating me to keep at it.
In my experience, some nits don't become apparent for a while. If you say you really like the sound I'd be game to see if I can make it work with the 35T.
Right now, I'm working on bringing a different circuit around to what I hope will some day be worth talking about. It sounds very good on the bench speaker but doesn't look so good on a scope and I'd like to have both up to snuff. . . . . . . . . so the timing of your posts is motivating me to keep at it.
Attachments
I'll report on how I like it some time later. So far it just sounds very authoritative and clean. It doesn't seem to color the sound. I like that as I tend to listen to electronic music and hard rock.
The initial versions of my experiments that had 1% distortion at 1W fell apart with hard rock and electronic music cranked up, but they sounded interesting with something simple like Norah Jones. There was definitely a little added color there, but I don't prefer that overall. I want an amp that holds it together when I crank it up during a weightlifting session.
If you like more color, the version with the input tube cathode running direct into the feedback network is still pretty clean but has some higher distortion. If you like closer to 1% at 1W, keep the feedback local to the output stage rather than including the two stages together like this.
The 35T looks like a really cool tube. I'm seriously tempted to buy some 40W output transformers and get my hands on some high impedance transmitting triode or pentode with a 100W plate and make a 40W version of this.
Edit: I really miss Michael Koster's contributions here.
The 35T is pretty neat and when I first started with Michael's circuit I was really impressed with the balls that little thing had. But listening to it I'd always find myself thinking "Pretty great. . . . . .well, that's enough for now." and turn it off. I had a conversation with Michael about it and he pointed me to a notch in the grid current at the transition from - to + Vg. I found it the same on my build. So I started looking at other drivers but didn't find anything to work better.
A conversation with Reinout got me to try the 35T in zero bias but whether it was the components I happened to have on hand to try it with or what, I don't know but every topology I've tried with it so far just sounds better with a different tube. I leave the small mono breadboard on the bench though, in hopes an idea will come along that will get the best out of it.
Yeah, I miss Michael too. He could dance with the best but was also very generous with time to help others of us get off the launch pad - and defended beginner level exploration when peers were pushing the usually prescribed path.
A conversation with Reinout got me to try the 35T in zero bias but whether it was the components I happened to have on hand to try it with or what, I don't know but every topology I've tried with it so far just sounds better with a different tube. I leave the small mono breadboard on the bench though, in hopes an idea will come along that will get the best out of it.
Yeah, I miss Michael too. He could dance with the best but was also very generous with time to help others of us get off the launch pad - and defended beginner level exploration when peers were pushing the usually prescribed path.
This approach may help control the grid current transition better for the 35T, since the driver and output stage are both in the loop, and the feedback is strong.
I believe the 826 grid stays completely positive until pretty high power levels, so I don't know if I would hear the transition area until the sound level is hurting my ears anyway. My garage test speakers are some early '70s Sansui speakers that are pretty efficient so the amp doesn't have to work too hard to sound big.
Keep in mind, this approach does create a problem that needs to be dealt with, and that's the bias of the input stage. The plate drifts quite a bit due to the high gain and I'll probably have to build an op-amp bias servo into a final design.
I have a board design that I've used in other amps, so that's what I'll be employing when the time comes.
As it stands I monitor the plate voltage of the input stage on a scope and I usually have to tweak it once per test session when it gets more than 50V from where I want it to be (midway between B+ and GND).
I believe the 826 grid stays completely positive until pretty high power levels, so I don't know if I would hear the transition area until the sound level is hurting my ears anyway. My garage test speakers are some early '70s Sansui speakers that are pretty efficient so the amp doesn't have to work too hard to sound big.
Keep in mind, this approach does create a problem that needs to be dealt with, and that's the bias of the input stage. The plate drifts quite a bit due to the high gain and I'll probably have to build an op-amp bias servo into a final design.
I have a board design that I've used in other amps, so that's what I'll be employing when the time comes.
As it stands I monitor the plate voltage of the input stage on a scope and I usually have to tweak it once per test session when it gets more than 50V from where I want it to be (midway between B+ and GND).