The arrangements of Q25-Q26-D12-D7, Q27-Q28-D10-D11, Q12-Q13-D3-D4 and Q14-Q15-D5-D6 have no current setpoint defined. Despite the precise current setting through the diodes, due to their spreading, the involved transistors will be set to indefinite current values, or even fail.
I would make a similar point. At least add some current-defining resistors to D3-D4, Q12-Q13, etc.
Also, as this appears to be a circuit, any sim results? Single stage gain (folded cascode) not used widely because of lower OLG (meaning, distortion not particularly low). Some degree of distortion cancellation will be achieved with the balanced circuitry though.
Also, as this appears to be a circuit, any sim results? Single stage gain (folded cascode) not used widely because of lower OLG (meaning, distortion not particularly low). Some degree of distortion cancellation will be achieved with the balanced circuitry though.
Two 1N4148s can allow the driver stage (2SA1145/2SC2705) to generate about 1.6mA quiescent operating current, can this be used for use? I uploaded the simulation file(NI Multisim 14.0), hoping to help improve the circuit to make it more ideal.
my English is not good, the reply will be made through translation software.
my English is not good, the reply will be made through translation software.
Last edited:
How is this 1.6mA defined? By what? Do you expect the Q's and the D's have exactly similar Vbe - Vak characteristics? In simulation, all models behave ideal, but the real thing prefers to diverge no matter what.
LT Spice designer. I have them at work too...
Wonder about stability... By all means, cool that you are designing, and the above concerns as well as stability usually get fixed at the bench, when the real design starts, so go fer it.
Wonder about stability... By all means, cool that you are designing, and the above concerns as well as stability usually get fixed at the bench, when the real design starts, so go fer it.
Did you include temperature drift, variations, dependencies, and the lot too in the simulation?
What is that equation all about: Ic/Icob = e EXP(qVbe/kT -1). All in (simulated) control, not?
Yes, I have built this circuit for a long time and have obtained good data through simulation, but I still cannot determine whether the values at all levels are reasonable. For example, whether the driver level 1.6mA (2SA1145/2SC2705) can meet the requirements. Perhaps it is more appropriate to use the TO-126 packaged MJE340/MJE350?
For better stability, the slew rate of the circuit is limited to a lower level (C9/C6)
IRF510 is used for thermal compensation, and Q33/Q34/Q35/Q36 forms CMFB circuit to suppress output DC level
I understand what you write, but do I not read that you have build this circuit long ago with yet unknown results, but have obtained good data through simulation only? The build reality should answer your own question, because it is measurable.
And how about my post in #6? Is there some correlation with your above statement?
Once upon time I built something similar, with less stages (only 2) but had issue with common DC drifting away from gnd and going to one of PS rails slowly. You solved that by Q33-Q37 circuit I see.
At the end I settled with this:
https://www.diyaudio.com/community/...ttle-fb-160watt-amplifier.412261/post-7672583
It is much more simple and absolutely stable and proven over 20+ years. If you want (much) better THD on expense of complexity, one good dual opamp introduced in feedback loop will provide that, making it composite unit.
At the end I settled with this:
https://www.diyaudio.com/community/...ttle-fb-160watt-amplifier.412261/post-7672583
It is much more simple and absolutely stable and proven over 20+ years. If you want (much) better THD on expense of complexity, one good dual opamp introduced in feedback loop will provide that, making it composite unit.