davidsrsb said:
Resistors R13 and R3 will make it less of an Rail to rail.
I guess you want to use R13 and R3 to control bias.
I would suggest use a more normal feedback.
Join base of Q3 and Q6.
Remove R4, R5,C1 And also R13 and R3.
Problem with this type of amplifier is how to control the bias of output.
If this could be done, without adding resistors in output rail,
then we could have a rail-to-rail amplifier.
With use of MOSFETS this could be done, as they are controlled by a voltage by Gates.
I am almost sure an idea like yours, has been used before.
Maybe someone knows.
I am trying to produce a nested feedback design. I was trying to get an output stage which was linear because of local feedback. It has to have some gain to avoid clipping problems earlier, I chose x3.
The input stage could be a decent opamp, so I avoid the usual miller compensation arrangement with problems of signals being referenced to on of the (dirty) supply rails.
R3 and R13 define the quiescent current, they can be fairly low valued, so the voltage drop across them is very small compared with the usual ~5V or more of a complementary source follower.
I have been thinking about building a class A amp with as high efficiency as possible for a tweeter driver. The relatively low rails used by class A amps makes the normal follower configuration a poor choice.
I don't like the bias stability problems of running common source outputs without local feedback and depending on enormous loop gain and feedback to get a low output impedance.
The input stage could be a decent opamp, so I avoid the usual miller compensation arrangement with problems of signals being referenced to on of the (dirty) supply rails.
R3 and R13 define the quiescent current, they can be fairly low valued, so the voltage drop across them is very small compared with the usual ~5V or more of a complementary source follower.
I have been thinking about building a class A amp with as high efficiency as possible for a tweeter driver. The relatively low rails used by class A amps makes the normal follower configuration a poor choice.
I don't like the bias stability problems of running common source outputs without local feedback and depending on enormous loop gain and feedback to get a low output impedance.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.