For the last while I have been considering putting the volume pot. after the first stage of a 2-stage push-pull MOSFET Class-A amp. An advantage to having the attenuation after the first stage is that the first stage sees more of the signal and might lose less detail. The problem I've been worrying about lately is that running the first stage at full power (about 2W) and then attenuating its output might cause the output to be less linear since the first stage is running at full power and Class-A MOSFET amps are more linear (and thus less distortion) at lowish power. Any suggestions/opinions?
The circuit uses global feedback but would it be a good idea to use local feedback on the first stage as well or would that be redundant?
Are their any errors in circuit? I've caught a few errors in the circuit in that the bias volt divider output should be behind the gate resistor. I plan to breadboard the first stage but I wonder about the power stage since I don't know very much about source-followers and they seem to work in a complex way.
I plan to use resistors between the rails and MOSFETS for the first-stage to set the bias current like the power-stage. Would values of 1 or 2 ohms be suitable?
The circuit uses global feedback but would it be a good idea to use local feedback on the first stage as well or would that be redundant?
Are their any errors in circuit? I've caught a few errors in the circuit in that the bias volt divider output should be behind the gate resistor. I plan to breadboard the first stage but I wonder about the power stage since I don't know very much about source-followers and they seem to work in a complex way.
I plan to use resistors between the rails and MOSFETS for the first-stage to set the bias current like the power-stage. Would values of 1 or 2 ohms be suitable?
An externally hosted image should be here but it was not working when we last tested it.
The basic idea is not really new, it used by Conrad-Johnson Premier 350. They use lateral MOSFETs both for the VAS, and for the driver of the output BJTs.
Your desing based on same idea. The only problem is the vertical MOSFET temperature compensation, which is missing, so You will be faced with thermal runaway, on both stages.
Sajti
Your desing based on same idea. The only problem is the vertical MOSFET temperature compensation, which is missing, so You will be faced with thermal runaway, on both stages.
Sajti
Kees52 is working on a mildly related topology with two IRFP240's (in push-push floating ground mode) on the outputs and IRF610's for the VAS in the allFET Circlotron thread.
http://www.diyaudio.com/forums/solid-state/291767-allfet-circlotron.html
the issue with thermal runaway would require a Vgs (FET) or Vbe(BJT) multiplier to be mounted on heatsink to control temperature.
http://www.diyaudio.com/forums/solid-state/291767-allfet-circlotron.html
the issue with thermal runaway would require a Vgs (FET) or Vbe(BJT) multiplier to be mounted on heatsink to control temperature.
Yes, rather immature design.
Drain resistors don't make sense.
Biasing scheme is not good - each drain is referenced to ground/rail and impedance of each reference depends on particular trimmer's position.
No thermal feedback - with HexFETs this is subject to thermal runaway. I understand - you have in mind class A, but still.
This is an inverting setup with NFB connected to the input, but no resistor in series from the input. That means, the gain is strongly dependent on the signal source's output impedance, being way too high with normal low-impedance signal source.
0.1uF shunting the output - what for?
Drain resistors don't make sense.
Biasing scheme is not good - each drain is referenced to ground/rail and impedance of each reference depends on particular trimmer's position.
No thermal feedback - with HexFETs this is subject to thermal runaway. I understand - you have in mind class A, but still.
This is an inverting setup with NFB connected to the input, but no resistor in series from the input. That means, the gain is strongly dependent on the signal source's output impedance, being way too high with normal low-impedance signal source.
0.1uF shunting the output - what for?
The rail to MOSFET resistors are there just to take voltage measurements across to help set the bias current (with the aid of a chart). This is a fairly simple circuit which is why I am brave enough to try to build it (I'm a long way from being an Electrical Engineer) and I have heard should sound decent mainly because it is class-A. I bought a chassis with some serious heat sinks and a large power supply and caps so I'm pretty committed to building it. I didn't think MOSFETs suffered from thermal runaway problems and I'm not paralleling them.
In the above statement would using say a 25k resistor and 10k. pot be better? I don't understand what you are saying due to my lack of knowledge. There should still be around 4-5k on the pot. so is this not enough?
The 0.1uF shunting the output was on the circuit I copied but I wasn't planning on using it since it seemed like an extra part that wasn't crucial. Perhaps it would shunt too much of the output to ground as well?
In the above statement would using say a 25k resistor and 10k. pot be better? I don't understand what you are saying due to my lack of knowledge. There should still be around 4-5k on the pot. so is this not enough?
The 0.1uF shunting the output was on the circuit I copied but I wasn't planning on using it since it seemed like an extra part that wasn't crucial. Perhaps it would shunt too much of the output to ground as well?
- Status
- Not open for further replies.