I have been playing with this, off and on, for a year. I posted the schematic in another thread but it didn't get much attention there. SO, I post it again here, to see if I can wrangle some comments out of the learned members.
I made it more accessible by changing the parts but it still operates the same way. Class A, currently at ~10 watts/8 ohms but I'd like to work this to be around 25-30 watts.
Thoughts? Comments?
I made it more accessible by changing the parts but it still operates the same way. Class A, currently at ~10 watts/8 ohms but I'd like to work this to be around 25-30 watts.
Thoughts? Comments?
Nelson Pass said:How does it sound?
I have a prototype up and running but haven't actually listened yet. Next step, I guess.
Steve Dunlap said:I thought the sound didn't matter as long as it simms well.
Ah Steve, sim first, build second and listen last, right? Adjust if needed.
OR...
just throw a bunch of parts down and wire them up!
Member
Joined 2009
Paid Member
I like this MJL, nice design indeed.
The schematic for my friends Chinese hybrid shows global feedback, since removed. What's interesting is that in it's modified form it's close (but not the same) to what you have - SS EF output stage and non-diff front end without feedback. And it sounds much nicer that it started so that should be a good omen.
It does use +/- rails but of course there's a cap between the tubes and the VBE.
A big difference is that the tube hybrid is ClassAB
I would love to see the spice FFT.
The schematic for my friends Chinese hybrid shows global feedback, since removed. What's interesting is that in it's modified form it's close (but not the same) to what you have - SS EF output stage and non-diff front end without feedback. And it sounds much nicer that it started so that should be a good omen.
It does use +/- rails but of course there's a cap between the tubes and the VBE.
A big difference is that the tube hybrid is ClassAB
I would love to see the spice FFT.
Nelson Pass said:Before you rush off and do that, consider that you can probably do
that front end with +/- supply rails and get rid of the output cap.
I did accidentally run the prototype on a split supply (hastily hooked up - not paying attention) and it looked fine on the scope. Bigger output but that's to be expected
Thanks for the suggestion. I'll do some experimenting and see how it sims.
Member
Joined 2009
Paid Member
MJL21193,
As soon as I posted my previous message, I realized I was wrong. I assumed that the output section had both CE and CC stages. Obviously, I didn't look very closely. I hoped no one would notice my error, but I suspected you would.
The input stage is especially innovative. In effect, Q1, Q2, Q3 and the associated components form a differential input transconductance amplifier. The closed loop AC gain from the base of Q1 to the collectors of Q2 and Q3 is approximately (R13+R21)/R21. In addition, the bias pot will allow you to set the output DC level to about 25 volts. A pot may not even be needed, but it doesn't hurt.
The output stage appears to be fairly typical, but setting the bias level might be trickier than for a circuit that uses global feedback. I say this because you will need to bias the output transistors with a high enough current to minimize crossover distortion, but at a low enough current to prevent excessive power dissipation. Circuits with global negative feedback can probably get by with a lower bias current in the output section.
Anyway, it's interesting to see something that is not just a variant of other stuff.
Rick
As soon as I posted my previous message, I realized I was wrong. I assumed that the output section had both CE and CC stages. Obviously, I didn't look very closely. I hoped no one would notice my error, but I suspected you would.
The input stage is especially innovative. In effect, Q1, Q2, Q3 and the associated components form a differential input transconductance amplifier. The closed loop AC gain from the base of Q1 to the collectors of Q2 and Q3 is approximately (R13+R21)/R21. In addition, the bias pot will allow you to set the output DC level to about 25 volts. A pot may not even be needed, but it doesn't hurt.
The output stage appears to be fairly typical, but setting the bias level might be trickier than for a circuit that uses global feedback. I say this because you will need to bias the output transistors with a high enough current to minimize crossover distortion, but at a low enough current to prevent excessive power dissipation. Circuits with global negative feedback can probably get by with a lower bias current in the output section.
Anyway, it's interesting to see something that is not just a variant of other stuff.
Rick
sawreyrw said:
Thanks for the clarification Rick.
I ran the prototype on a 40VDC supply (no separate regulated higher voltage for the front end yet), idle current set to ~1 amp and had a listen. Connected to my older 2-way TL, it sounds pretty damn good - clean and undistorted.
I need to do some more tests and tinkering but I'm pleased so far. I'm not turned off by the output cap - I have another class A amp with the same and I can't hear any short comings.
Nothing nasty when I remove power and it stabilizes quickly (no notable thump) on power up.
There is a slight hum but that is due to the makeshift power supply (just one 10mF cap). I expect that will subside when I do a proper supply with the higher voltage regulated front end supply.
It's running very stable, laying on the bench with wires dangling everywhere and no output inductor or Zobel.
There is a slight hum but that is due to the makeshift power supply (just one 10mF cap). I expect that will subside when I do a proper supply with the higher voltage regulated front end supply.
It's running very stable, laying on the bench with wires dangling everywhere and no output inductor or Zobel.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.