Konnichiwa,
Actually, much what you mention helps INHERENT LINEARITY, meaning it stops or reduces alterations to the wholey, holey and only signal. It does so, if well applied without resort to either local or global positive or negative feedback.
Note anything with a positive and negative input plus an applied feedback loop is a "Op-Amp", even the single FET in the PassZen or a single triode with Anode to grid loop feedback.
You mean the category which uses "bad" op-amps with feedback?
I belong to the zeroeth category which regards feedback as a tool to get certain results, which like any tool needs application with care. Feedback is no pankea to solve all problems and in my own system at home I prefer to do without.
From MC Pickup to Speaker Output I have 4 active devices, no feedback, not even "local" feedback, all supplies are unregulated. Of course, most of the time I add a digital equaliser via it's analogue inputs into the mix, which has at least 100's of active stages (real or virtual), loads of Op-Amp's, negative feedback etc.
But of course, I did not design the EQ (I did modify though) I just find it works.
Of course. That is 1950's stuff, read Shorter, Olson and Crowhurst from those days when the JAES was not "just toing the party line". There are plenty more publicatiosn. Including some which suggest higher degrees of "ear congruent" distortion can enhance audibility by a process best translated as "residiium hearing"
Actually, feedback is NOT a subtractive process. It is ALLWAYS additive.
Secondly, feedback does "subtract" distortion, it redistributes it into different domains (PIM/TIM) and spectrae, hencemy insistence on INHERENT linearity. If nothing goes wrong to start with there is no fallout from fixing it.
Of course, on the additive nature of feedback, the sign of some of the added components may be negative, but the feedback node is ALLWAYS and INHERENTLY additive.
So, at best feedback only adds the signals with signs and magnitudes.
In reality it adds many others... Simply killing the common mode element in this additive process (using inverting mode) helps tremendously, interestingly BOTH objectively and subjectively.
Actually, it sounds like a completely irrational religion AND it is very far from the truth. At least from where I stand.
Agreed. Valves sound much better and in the dark they glow much longer....
Assuming your devices are "good enough", which sadly excludes solid state almost wholesale, Admiral Nelsons protestations ("It's only a fleshwound") nothwithstanding.
Solid state requires the added complexity that makes chips near mandatory to actually work well enough for music. Oh, it requires almost allways some for of (looped) feedback too.
Sayoanra
nobody special said:
Actually, much what you mention helps INHERENT LINEARITY, meaning it stops or reduces alterations to the wholey, holey and only signal. It does so, if well applied without resort to either local or global positive or negative feedback.
nobody special said:
Note anything with a positive and negative input plus an applied feedback loop is a "Op-Amp", even the single FET in the PassZen or a single triode with Anode to grid loop feedback.
nobody special said:
You mean the category which uses "bad" op-amps with feedback?
I belong to the zeroeth category which regards feedback as a tool to get certain results, which like any tool needs application with care. Feedback is no pankea to solve all problems and in my own system at home I prefer to do without.
From MC Pickup to Speaker Output I have 4 active devices, no feedback, not even "local" feedback, all supplies are unregulated. Of course, most of the time I add a digital equaliser via it's analogue inputs into the mix, which has at least 100's of active stages (real or virtual), loads of Op-Amp's, negative feedback etc.
But of course, I did not design the EQ (I did modify though) I just find it works.
nobody special said:
Of course. That is 1950's stuff, read Shorter, Olson and Crowhurst from those days when the JAES was not "just toing the party line". There are plenty more publicatiosn. Including some which suggest higher degrees of "ear congruent" distortion can enhance audibility by a process best translated as "residiium hearing"
nobody special said:
Actually, feedback is NOT a subtractive process. It is ALLWAYS additive.
Secondly, feedback does "subtract" distortion, it redistributes it into different domains (PIM/TIM) and spectrae, hencemy insistence on INHERENT linearity. If nothing goes wrong to start with there is no fallout from fixing it.
Of course, on the additive nature of feedback, the sign of some of the added components may be negative, but the feedback node is ALLWAYS and INHERENTLY additive.
So, at best feedback only adds the signals with signs and magnitudes.
In reality it adds many others... Simply killing the common mode element in this additive process (using inverting mode) helps tremendously, interestingly BOTH objectively and subjectively.
nobody special said:
Actually, it sounds like a completely irrational religion AND it is very far from the truth. At least from where I stand.
nobody special said:
Agreed. Valves sound much better and in the dark they glow much longer....
nobody special said:
Assuming your devices are "good enough", which sadly excludes solid state almost wholesale, Admiral Nelsons protestations ("It's only a fleshwound") nothwithstanding.
Solid state requires the added complexity that makes chips near mandatory to actually work well enough for music. Oh, it requires almost allways some for of (looped) feedback too.
Sayoanra
Re: Re: Re: Re: Re: Re: Re: summarizing...
I respectfully declare this total and utterly nonsense. First comes the topology, the type of components (FET? BP? Tubes?) the the detailed diagram, then the implementation. And then IF you get all this optimised, you can indulge in 100$ resistors, and convince yourself it really makes a difference. But make a non-optimal choice in one of the basics, and no amount of money thrown at it makes a difference (except on the price).
Jan Didden
(Yes I know you were only joking, but a lot of people here may actually BELIEVE you, for pete's sake)
millwood said:
I respectibly disagree. As you long as you pay >$1K for your cap and $2K for your resistors, your amp will sound fabulous, regardless of its topology, or build quality.
Oh, don't forget about damping,
I respectfully declare this total and utterly nonsense. First comes the topology, the type of components (FET? BP? Tubes?) the the detailed diagram, then the implementation. And then IF you get all this optimised, you can indulge in 100$ resistors, and convince yourself it really makes a difference. But make a non-optimal choice in one of the basics, and no amount of money thrown at it makes a difference (except on the price).
Jan Didden
(Yes I know you were only joking, but a lot of people here may actually BELIEVE you, for pete's sake)
Re: Re: Re: Re: Re: summarizing...
Well, just for reference, my Nad is a Monitor Series Receiver 7000.
My money couldn't buy better about 12 years ago.
It's extensively tweaked.
Hey millwood, even with all those things you wouldn't go far if you didn't replace those nasty JRC op-amps on the preamp.
My Nad has OPA2132.
And it plays very well indeed, but still not enough current/PSU (?) to drive my Epos 11 speakers properly.
But let's say this amp, as it is now, is good hi-fi.
My GC is high-end.
millwood said:
Well, with a few tweaks, y ou can get excellent sound out of that amp, including that woeful NAD. start with gold-plated resistors, plentium coated caps, tungsgen reinforced heatsinks, 14K white gold jacks, and silver-plated cover, you don't have to worry about those sucky transistors anymore.
BTW, don't forget about using aluminum oxide insulation pads.
As they say in banking, there is no bad amp, just with the right resistors,
Well, just for reference, my Nad is a Monitor Series Receiver 7000.
My money couldn't buy better about 12 years ago.
It's extensively tweaked.
Hey millwood, even with all those things you wouldn't go far if you didn't replace those nasty JRC op-amps on the preamp.
My Nad has OPA2132.
And it plays very well indeed, but still not enough current/PSU (?) to drive my Epos 11 speakers properly.
But let's say this amp, as it is now, is good hi-fi.
My GC is high-end.
One disadvantage of chip amplifiers I haven't seen mentioned explicitly yet:
I don't know for sure if this applies to the National Semiconductors chips (I would be surprised if it didn't, though), but with some expensive exceptions, most bipolar IC processes have only very bad lateral PNP's with a very low current handling available. This is in fact why quasi-complementary output stages are so popular in monolithically integrated amplifiers. Making a fully complementary output stage with lateral PNP's would be difficult to do and would require an enormous chip area.
I don't know for sure if this applies to the National Semiconductors chips (I would be surprised if it didn't, though), but with some expensive exceptions, most bipolar IC processes have only very bad lateral PNP's with a very low current handling available. This is in fact why quasi-complementary output stages are so popular in monolithically integrated amplifiers. Making a fully complementary output stage with lateral PNP's would be difficult to do and would require an enormous chip area.
Konnichiwa,
Actually, not only IC's have usually "bad" PNP Transistors. In virtually ANY "complementary pair" the NPN (or N-Channel) transistor has a higher Beta, better Beta linearity and a higher Ft.
That's solid state for you. There where good reasons why the old masters never made a "P-Channel" Tube....
Sayonara
MarcelvdG said:
Actually, not only IC's have usually "bad" PNP Transistors. In virtually ANY "complementary pair" the NPN (or N-Channel) transistor has a higher Beta, better Beta linearity and a higher Ft.
That's solid state for you. There where good reasons why the old masters never made a "P-Channel" Tube....
Sayonara
Konnichwa Eddy San,
Sure.
So, you are saying that all the people who listen to fully complementary Solid State Amp's REALLY only like to listen to distortion? Works for me, Let's run with it....
Sayonara
Steve Eddy said:
Aye, but then that could be seen as beneficial as it would produce an asymmetry between the pairs increasing second/even order distortion giving it a somewhat more "single-ended" characteristic.
se
Sure.
So, you are saying that all the people who listen to fully complementary Solid State Amp's REALLY only like to listen to distortion? Works for me, Let's run with it....
Sayonara
Kuei Yang Wang,
Discrete PNP's indeed tend to be somewhat slower than discrete NPN's. I'm no device physicist, but I guess this has something to do with the lower mobility of holes compared to electrons.
However, the differences between discrete PNP's and NPN's are totally negligible compared to the differences between vertical NPN's and lateral PNP's in a typical bipolar IC process. A factor of three difference in current gain, a factor of 100 in peak fT and a factor of 100 in current handling capability for devices with similar sizes is not unusual.
Discrete PNP's indeed tend to be somewhat slower than discrete NPN's. I'm no device physicist, but I guess this has something to do with the lower mobility of holes compared to electrons.
However, the differences between discrete PNP's and NPN's are totally negligible compared to the differences between vertical NPN's and lateral PNP's in a typical bipolar IC process. A factor of three difference in current gain, a factor of 100 in peak fT and a factor of 100 in current handling capability for devices with similar sizes is not unusual.
Some of the more modern Sanken, Sanyo, Toshiba, NEC, Moterola devices are amazingly complementary. Sadly, many of these have already gone out of production (the evil influence of surround sound??).
How symmetrical can you get in a quasi-complementary configuration? My impression was always that quasi-complementary was bound to be less symmetrical than a good symmetrical pair.
Regards,
Eric
How symmetrical can you get in a quasi-complementary configuration? My impression was always that quasi-complementary was bound to be less symmetrical than a good symmetrical pair.
Regards,
Eric
capslock said:
In a quasi-complimentary scheme, a PNP device is still the controlling device so I don't see how you really get away from the NPN/PNP issue.
Personally I've preferred the true complimentary hybrid power opamps over the quasi-complimentary monolithic power opamps that I've tried. Though there are a lot more differences beside the output stage topology so I've really no idea how much is due to the true complimentary nature of the hybrids I've used.
For example, the devices I use (the Apex PA-16 and PA-02) are also a bit unusual in that the output is driven from the collectors of the output devices rather than the emitters.
se
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.