Great! Thanks Ian!
I think that about does it.Post 1 "Well, there's the challenge. A wideband non switching Auto Bias amplifier circuit that works, doesn't oscillate and blow up, and is not too complex."
Great project Ian, it's good to see a worthwhile project defined, undertaken and brought to a successful completion. Chapeau!Hi Minek, Thanks for asking.
Hi All,
I think that about does it.
BTW none of my circuits are patented, they are all in the public domain and free for all to use.
I have updated Post 1 summary of key posts.
Thanks to all who have already contributed to this thread!
Anyone is welcome to add more to this thread. Please do.
Cheers,
Ian Hegglun
The tricky bit is Q1 needs to be mounted on one of the Schottky diodes. I think it can be done by omitting the COM pin. Q1 base then uses 'COM' pin and Q1 emitter then uses 'Vtop' pin. Bias resistor R14 could be a pick and place resistor with machined plugs as part of the discrete 8 pin DIP (or maybe an EEPROM pot so the DIP can be sealed).BTW, I'm still fiddling with auto bias circuits, one the holy grails in audio. If one gets it truly right (without adding any distortion), that would mean a major step forward. Post 3308
An interesting variation of the LT1166 autobias variant by ttlhacker
https://www.diyaudio.com/community/...deband-auto-bias-amp-design-and-build.392549/
and I have modified it for higher voltages and reduced the number of stages while still giving good sim performance (not bench tested).
H Ian,... The interesting thing with this type of autobias compared to my last design is the Schottky diode temperature change does not much affect the idle current so is free from thermal time delays in both the power transistors and Schottky diodes. ...
I intend to use the previous scheme on lower power 10-25W amps, different need than yours. The plan is to have as uniform power dissipation as practical from idle to max power output to reach a somewhat steady state thermal equilibrium. However a compensation is absolutely required to stabilize IQ against ambient ΔT, an NTC as you suggested would be simple enough.... In my last design the Schottky diode temperature change, first at start and after full power, is tricky to accurately thermally compensate and there is a time delay....
In the ttlhacker autobias the Schottky diodes are only slightly conducting at idle - plot below. Most of the idle current passes through emitter resistors R5,R8 and not the Schottky's.Could you elaborate a bit more on the reasons of how ΔVF of the Schottky diode due to ΔT (ambient and self dissipation) does not effect idle current as much as the autobias action that depends on ΔVF due to ΔIF.
Yes. I prefer autobias arrangements that can interchange bipolar power transistors with MOSFET's without any change to the circuit; usually only a small re-tweak of the idle current. I confirmed that I could change from BJT's to MOSFETs in my previous autobias design in bench tests, without any resistor value changes apart from a re-trim of the idle pot.Do you see any disadvantage in directly detecting VBE of output transistor instead of using a Schottky for this scheme?
What do you think about the one I showed in Post#106?... D4, D5, Q3, and Q4 adds four ...
That circuit avoids switching, but I would call it a form of Vbe multiplier rather than a circuit that senses and regulates current.What do you think about the one I showed in Post#106?
Pretty sure autobias can actually do that! Here's the output stage of my original autobias design, as posted over in the thread about it: Original thread (has the LTSpice ASC)The auto-bias circuit is very clever. I was not aware that auto-biasing was possible at audio frequencies.
However, it has a downside: the switching of D4, D5, Q3, and Q4 adds four kinks to the transfer curve. I'll call this "auto-bias distortion".
I like to design amplifiers with good open-loop linearity. I don't think auto-bias can do that.
Ed
Moving the sensing resistors/diodes to the collectors greatly reduces the non-linearity, but does not eliminate it because the collector voltage still changes (slightly). While Q6 and Q8 may not cut off, they come close.Since this version of the amp has the current sense resistors in the output transistors' collectors (instead of at the emitters), no semiconductors in the signal path ever switch. Only the clamping diodes do switch (very softly), but they're in the collector paths, so they don't influence the output voltage at all. No transistor ever turns off.
ThanksSuper nice, thumbs up! ...