Hi
It is a standard symmetrical differential input circuit but with only single stage as VAS/output like F5 amplifier from Pass for example. However it is a VFA type compared to the so called “CFA” type as the F5.
The input stage current is very high to be able to drive the mosfet output. Therefore, for input transistors power consideration the voltage should not be higher than about +/-12vdc max (180mw) for the selected transistors. Also, the possibly small open loop gain of the amp could impose a class A bias current to keep THD to a reasonable level....however the transconductance of the mosfet seems very high....
As for the selected output mosfet they seem to be for high switching current with very high capacitance. They would not be stable in temperature without source resistors installed ...
Fab
It is a standard symmetrical differential input circuit but with only single stage as VAS/output like F5 amplifier from Pass for example. However it is a VFA type compared to the so called “CFA” type as the F5.
The input stage current is very high to be able to drive the mosfet output. Therefore, for input transistors power consideration the voltage should not be higher than about +/-12vdc max (180mw) for the selected transistors. Also, the possibly small open loop gain of the amp could impose a class A bias current to keep THD to a reasonable level....however the transconductance of the mosfet seems very high....
As for the selected output mosfet they seem to be for high switching current with very high capacitance. They would not be stable in temperature without source resistors installed ...
Fab
Last edited:
I'm not convinced about A/B operation of amplifiers, as I have not realized any advantage of a zero voltage bias current to limit THD. It is just a waste of energy and it destabilizes the output stages. 0.1..1% THD is ok for me. The open loop voltage amplification is 10k. This is ok.
The output MOSFETs work thermally stable, as it is a pure B amplifier which is operated at just 1% of its current limit.
In fact I have never considered class B amplifier because I had learned a long time ago that they were prone to high crossover distorsions. It seems you know how to prevent that....
Regarding 1% THD level is it at 1W or close to max power? THD profile ?
Regarding open loop gain can you explain how you determine the transconductance of this mosfet at very low bias current from the datasheet.
Thanks
Fab
Regarding 1% THD level is it at 1W or close to max power? THD profile ?
Regarding open loop gain can you explain how you determine the transconductance of this mosfet at very low bias current from the datasheet.
Thanks
Fab
Member
Joined 2009
Paid Member
This is a better design: TGM7 - an amplifier based on Greg Ball SKA
It provides stronger drive for the output FETs, otherwise very similar to what you posted. Works fine.
It provides stronger drive for the output FETs, otherwise very similar to what you posted. Works fine.
There is some amount of "crowbar current" (also called "shoot thru current") that flows through the output MOSFETs when the input is zero. The magnitude of this current depends upon many things
Several of those are temperature dependent. Which means the crowbar current can vary with temperature. It might even have a propensity to thermal runaway, which would be truly unfortunate.
You might be forced to make R1 and R2 into trimmer potentiometers, to handle the very real possibility that M1 and M2 need different amounts of gate drive to achieve the DC bias condition you have chosen. This would really stink because it guarantees the two halves of the circuit have different gain (!). You could try making R6 and R7 into trimmers instead, but this affects the gm of the differential pairs, and thus you get unequal gain.
- The current in Q6, times (R1 / 2)
- The current in Q5, times (R2 / 2)
- The turn-on threshold voltage of P channel MOSFET M1
- The transconductance of M1
- The turn-on threshold voltage of N channel MOSFET M2
- The transconductance of M2
Several of those are temperature dependent. Which means the crowbar current can vary with temperature. It might even have a propensity to thermal runaway, which would be truly unfortunate.
You might be forced to make R1 and R2 into trimmer potentiometers, to handle the very real possibility that M1 and M2 need different amounts of gate drive to achieve the DC bias condition you have chosen. This would really stink because it guarantees the two halves of the circuit have different gain (!). You could try making R6 and R7 into trimmers instead, but this affects the gm of the differential pairs, and thus you get unequal gain.
Member
Joined 2009
Paid Member
In fact R1 and R2 are trimmers, as Vgsth differ by about 0.1 V. The transconductance differ by a factor of 1.5. But it is ok.
When you say that transconductance N/P ratio difference is 1.5 is it measured at specific current or from the datasheet. From the datasheet it seems to vary a lot in the low current region.
For thermal stability even if you operate at very low current compared to rated current it could have a tendancy to rise with music playing at medium/high power because of mosfet temperature coefficient.... source resistors is of good help and maybe compatible with vgs/ID curve matching exercise...
I use thermistors in input stage to regulate output current in that type of configuration for class A and A/B but maybe not a good idea for class B amp because of thermistors difference...
I forgot to say that it seems a good simple project.
Fab
For thermal stability even if you operate at very low current compared to rated current it could have a tendancy to rise with music playing at medium/high power because of mosfet temperature coefficient.... source resistors is of good help and maybe compatible with vgs/ID curve matching exercise...
I use thermistors in input stage to regulate output current in that type of configuration for class A and A/B but maybe not a good idea for class B amp because of thermistors difference...
I forgot to say that it seems a good simple project.
Fab
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.