Hi,
tell us why your decision to insert 2 input caps and keep the bases of the LTPs separate?
The feedback to the left side inverting LTP seems wrong. There is a missing DC block on the lower leg and an extra pair of caps in line with the feed.
Why do you call it ClassB? Depending on output pair Iq it could be A, AB or B
No input RF filter, no output Thiel network to reject RF and load the output at high frequency
How about a cap bypass across the Vbe multiplier
The DC block on the input could be non electrolytic with some advantage.
Output offset adjustment? LTP emitter resistors? Stability control components?
Decoupling? Separation of clean and dirty grounds?
Overall not much thought so far, or is this a leg pull?
tell us why your decision to insert 2 input caps and keep the bases of the LTPs separate?
The feedback to the left side inverting LTP seems wrong. There is a missing DC block on the lower leg and an extra pair of caps in line with the feed.
Why do you call it ClassB? Depending on output pair Iq it could be A, AB or B
No input RF filter, no output Thiel network to reject RF and load the output at high frequency
How about a cap bypass across the Vbe multiplier
The DC block on the input could be non electrolytic with some advantage.
Output offset adjustment? LTP emitter resistors? Stability control components?
Decoupling? Separation of clean and dirty grounds?
Overall not much thought so far, or is this a leg pull?
HI Mastertech,
If you combine the input stage(s) at their bases you can have up to 90% cancellation of bias current for those bases - and with unity DC gain (your C to GND in the feedback leg), you could leave out input capacitors altogether. It my belief that no C is the best C.
This is what I have done in the S.K.A where bias current cancellation results in Ib nett of only 0.1uA. The input C (provided for on the board) is linked out for normal operation.
Cheers,
Greg
If you combine the input stage(s) at their bases you can have up to 90% cancellation of bias current for those bases - and with unity DC gain (your C to GND in the feedback leg), you could leave out input capacitors altogether. It my belief that no C is the best C.
This is what I have done in the S.K.A where bias current cancellation results in Ib nett of only 0.1uA. The input C (provided for on the board) is linked out for normal operation.
Cheers,
Greg
Hi my friend Kanwar, youve been absent for a while it's great
yo see you back on the forum, what is going on these days
with classes i read on the forum that class-d is going to
dominate the whole amplifier domain even the high-end
I am primarily focused on class-a at present but this doesnt
mean that ive given up on all other classes, it is intresting to
see other people making progres or rather so they claim
Hi Greg all classes have pros and cons depending on the application and class-a to me is the most challenging from
the engineering point of view i think it has not been fully
developed for obvious reason i see great potential in class-a
in terms of evolving for high-end
regards to both of you
cheers
yo see you back on the forum, what is going on these days
with classes i read on the forum that class-d is going to
dominate the whole amplifier domain even the high-end
I am primarily focused on class-a at present but this doesnt
mean that ive given up on all other classes, it is intresting to
see other people making progres or rather so they claim
Hi Greg all classes have pros and cons depending on the application and class-a to me is the most challenging from
the engineering point of view i think it has not been fully
developed for obvious reason i see great potential in class-a
in terms of evolving for high-end
regards to both of you
cheers
Hi mastertech
I think that your application of double differential stage is a bit risky. There is a threat of a DC-inbalance because of different DC values on every base. The changes might be slight but with high transconductance of a bjt it can result in diff'l stage bjt working in not very linear region of operation.
You could minimize that effect with relatively high value emitter resistors ( >150R), good matching etc.
What do you mean by 'class B' in terms of quescient current?
<20mA or <2mA or <0.2mA ?
I have an inpression that divison between class B and AB is a bit unclear and more historical now.
regards
I think that your application of double differential stage is a bit risky. There is a threat of a DC-inbalance because of different DC values on every base. The changes might be slight but with high transconductance of a bjt it can result in diff'l stage bjt working in not very linear region of operation.
You could minimize that effect with relatively high value emitter resistors ( >150R), good matching etc.
What do you mean by 'class B' in terms of quescient current?
<20mA or <2mA or <0.2mA ?
I have an inpression that divison between class B and AB is a bit unclear and more historical now.
regards
Hi Greg,
I wouldn't choose to separate Class-B and Class-AB that way.
You can have a very small quiescent current that eliminates discontinuity. In such a situation, each device (of a standard push/pull config.) would conduct for very close to 50% of an output sine cylce. If you are to define this as Class-AB, there isn't a destinction for output stages that have a significant quiescent current, thereby operating in class-A for low output powers.
Surely the name Class-AB implies a hybrid of Class-A and Class-B? And you wouldn't suggest that an amp. with a low quiescent current could be considered as having a significant region of output power where it is operating in Class-A?
I would suggest that an amp. with discontinuity in conduction of output devices is a Class-C output stage and as you rightly say, should not be used in audio amplifiers.
If you were to define class-B as each output device conducts for exactly 50% of the time, then good luck making such a thing. IMHO, it's best to use Class-B to refer to amps. with low quiescent current (and therefore no significant class-A region) and to use Class-AB for amps. with much more significant quiescent current.
I wouldn't choose to separate Class-B and Class-AB that way.
You can have a very small quiescent current that eliminates discontinuity. In such a situation, each device (of a standard push/pull config.) would conduct for very close to 50% of an output sine cylce. If you are to define this as Class-AB, there isn't a destinction for output stages that have a significant quiescent current, thereby operating in class-A for low output powers.
Surely the name Class-AB implies a hybrid of Class-A and Class-B? And you wouldn't suggest that an amp. with a low quiescent current could be considered as having a significant region of output power where it is operating in Class-A?
I would suggest that an amp. with discontinuity in conduction of output devices is a Class-C output stage and as you rightly say, should not be used in audio amplifiers.
If you were to define class-B as each output device conducts for exactly 50% of the time, then good luck making such a thing. IMHO, it's best to use Class-B to refer to amps. with low quiescent current (and therefore no significant class-A region) and to use Class-AB for amps. with much more significant quiescent current.
This reminds me somewhat of the input stage of the Musical Fidelity A1 amplifier. I don't really see what advantages having seperate input filter stages gives over eg the Leach design where one input filter is used to serve both the upper and lower LTP's. If anything youre going to get inaccuracies brought in by the different component tolerances in both filters.
Seperate feedback paths, again im unsure, but perhaps this could be useful to counter the differences between NPN and PNP output devices?
Seperate feedback paths, again im unsure, but perhaps this could be useful to counter the differences between NPN and PNP output devices?
Hi harryDymond,
...."I wouldn't choose to separate Class-B and Class-AB that way.
You can have a very small quiescent current that eliminates discontinuity. In such a situation, each device (of a standard push/pull config.) would conduct for very close to 50% of an output sine cylce. If you are to define this as Class-AB, there isn't a destinction for output stages that have a significant quiescent current, thereby operating in class-A for low output powers.
Surely the name Class-AB implies a hybrid of Class-A and Class-B? And you wouldn't suggest that an amp. with a low quiescent current could be considered as having a significant region of output power where it is operating in Class-A?"
Class B is already commonly understood to be NO bias, while Class AB is understood to be more finely defined as low or high bias, the latter being a transition towards Class A. I think my discontinuity statment is fair, given that I make low bias (25-50mA per device) Class AB amps and they certainly don't have a discontinuity so I wouldn't like them bundled into Class B - just because they're energy efficient. My Simple Killer Amp for example runs from +/- 55V supplies idling at 90mA ( 2 pair) for 10W heatsink dissipation (only) above a pure Class B with the bias turned to 0mA.
Cheers,
Greg
...."I wouldn't choose to separate Class-B and Class-AB that way.
You can have a very small quiescent current that eliminates discontinuity. In such a situation, each device (of a standard push/pull config.) would conduct for very close to 50% of an output sine cylce. If you are to define this as Class-AB, there isn't a destinction for output stages that have a significant quiescent current, thereby operating in class-A for low output powers.
Surely the name Class-AB implies a hybrid of Class-A and Class-B? And you wouldn't suggest that an amp. with a low quiescent current could be considered as having a significant region of output power where it is operating in Class-A?"
Class B is already commonly understood to be NO bias, while Class AB is understood to be more finely defined as low or high bias, the latter being a transition towards Class A. I think my discontinuity statment is fair, given that I make low bias (25-50mA per device) Class AB amps and they certainly don't have a discontinuity so I wouldn't like them bundled into Class B - just because they're energy efficient. My Simple Killer Amp for example runs from +/- 55V supplies idling at 90mA ( 2 pair) for 10W heatsink dissipation (only) above a pure Class B with the bias turned to 0mA.
Cheers,
Greg
Actually, the classic definition deals with % duty cycle the current flows in each polarity or cycle. I'd rather keep it that way since it's accepted and what is taught. Anything else muddies the waters. And yes, I know how difficult it is to make a class B amplifier and have it stay there. Let's classify those by the designer's intent (English common law).
-Chris
-Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.