Hi 🙂
I have discussed this subject for some time with some of the other members here, and we found that the information in that white paper contains to little information 🙁
However we all agree on, that it could be interesting to try to "copy" this bias circuit 😉
So if anyone has a ML amp with adaptive bias, please lift the lit on it and help us 😀
I have discussed this subject for some time with some of the other members here, and we found that the information in that white paper contains to little information 🙁
However we all agree on, that it could be interesting to try to "copy" this bias circuit 😉
So if anyone has a ML amp with adaptive bias, please lift the lit on it and help us 😀
This circuit was developed by firm Sansui, not by ML. Look at J. Audio Eng Soc.,Vol. 29, No. 3, 1981 March. Author is Susumu Tanaka...
![tanaka-lapaz[1].gif](/community/data/attachments/58/58003-36f0a1d4203fcb140b6f11d960055ff8.jpg?hash=NvCh1CA_yx)
These schemes are most similar to the "optical bias" circuit,
my patent #4,752,745 from 1988, where bias feedback is
taken from the Bases (or Gates) of the output devices.
😎
my patent #4,752,745 from 1988, where bias feedback is
taken from the Bases (or Gates) of the output devices.
😎
Gentlemen;
here is my contribution to the topic.
I have mine self-adaptive bias. Plus, it protects output transistors against overload, and when overdriven it sounds like a vacuum tube.
The idea is, playing with feedbacks. In "normal" mode the feedback is 100% global over both transistors of the shoulder and feedback diodes that also bias and thermo-stabilize the amp, when overdriven it slowly becomes a feedback by current in the 1'st (driver) stage, when the 2'nd stage amplifiers the current limiting it on the level of I*B, where I is the current through the 1'st transistor, B is HFE of the second one.
Errors on the schematic: pots to the ground (max output current) are connected to the ground, top right pot (bias) is R3, bottom right is R4.
The values for the resistors are for A+B class. For C class pots R3,R4 may be omitted together with series resistors.
For A class pots R1, R2 and associated Shottky diodes may be eliminated, for more bias current R3, R4 and resistors in series must be decreased.
Feedback diodes (silicone fast diodes) need to be thermally coupled with driver transistors. Theoretically, there may be the same transistors as used for drivers, with B-C junction used as a diode, B-E shorted, but I did not try such a version.
here is my contribution to the topic.
I have mine self-adaptive bias. Plus, it protects output transistors against overload, and when overdriven it sounds like a vacuum tube.
The idea is, playing with feedbacks. In "normal" mode the feedback is 100% global over both transistors of the shoulder and feedback diodes that also bias and thermo-stabilize the amp, when overdriven it slowly becomes a feedback by current in the 1'st (driver) stage, when the 2'nd stage amplifiers the current limiting it on the level of I*B, where I is the current through the 1'st transistor, B is HFE of the second one.
Errors on the schematic: pots to the ground (max output current) are connected to the ground, top right pot (bias) is R3, bottom right is R4.
The values for the resistors are for A+B class. For C class pots R3,R4 may be omitted together with series resistors.
For A class pots R1, R2 and associated Shottky diodes may be eliminated, for more bias current R3, R4 and resistors in series must be decreased.
Feedback diodes (silicone fast diodes) need to be thermally coupled with driver transistors. Theoretically, there may be the same transistors as used for drivers, with B-C junction used as a diode, B-E shorted, but I did not try such a version.
I did a Dynamicaly Biased Headphone Amp a few years ago attached is the Schematic of the Output Stage.
they create additionial voltage drop so as the VCE of the driver transistors are about 1.6 volts this keeps them conducting otherwise there would be only a few mV VCE and would not turn on
this is a Dimond Buffer topology so looking at the upper half D1 moves the emiter of Q23 about 700 mV away from the b of Q24 Giving an added 700mV of VCE for Q23 D3 moves Q24's emiter up the other driection away from the Base of Q23 giving another 700 mV since the emiter of Q24 would normaly have only 20 mV from the output and the out put and input should both be the same this gives about 1.4 volts VCE for Q23. CRD1 is a constant current source keeping Q24 from compleatly turning off at the Zero crossover point.
Attached is the Class Ab output stage as now used in the PPA headphone Amp as seen hear http://www.tangentsoft.net/audio/ppa/amp2/
A discussion on the Dyn-O-Bias buffer can be found Hear http://www6.head-fi.org/forums/showthread.php?t=185143
A discussion on the Dyn-O-Bias buffer can be found Hear http://www6.head-fi.org/forums/showthread.php?t=185143
Hi,ppl.How to apply your bias system(yes,bias system only) to a normal push pull output stage???
- Status
- Not open for further replies.