well, now that i've actually read the review and test results ...
in the KR-X, we have a design that has:
1. great sound (according to the reviewer)
and
2. great test results, measurement results
achieved with no global feedback and it's a fully balanced design to boot!
good job charles; congratulations!
so, just how long before the "tech trickle down" to products i might have a chance of affording ...
where's that first born when you need to sell 'em
mlloyd1
in the KR-X, we have a design that has:
1. great sound (according to the reviewer)
and
2. great test results, measurement results
achieved with no global feedback and it's a fully balanced design to boot!
good job charles; congratulations!
so, just how long before the "tech trickle down" to products i might have a chance of affording ...
where's that first born when you need to sell 'em
mlloyd1
no global feedback does not mean no feedback.
We can for example make an RIAA amp with 3 stages with zero GNF - but, each of the stages has feedback around it.
Maybe there is a sonic benefit to this, maybe not.
Zero Global Feedback looks like it has become an important consumer (obviously audiophile in this case) trigger associated with high end, much like mentioning the word 'digital' in a product tag line also inferes 'quality' and 'leading edge' to a consumer in the street (of course, here we have a different opinion, but that's a different matter).
This is marketing.
We can for example make an RIAA amp with 3 stages with zero GNF - but, each of the stages has feedback around it.
Maybe there is a sonic benefit to this, maybe not.
Zero Global Feedback looks like it has become an important consumer (obviously audiophile in this case) trigger associated with high end, much like mentioning the word 'digital' in a product tag line also inferes 'quality' and 'leading edge' to a consumer in the street (of course, here we have a different opinion, but that's a different matter).
This is marketing.
Bonsai said:
Regardless of the semantics of feedback, Charles has done a great job here. While I do not agree that NFB is a bad thing, I also do not assert that a really fine design can only be done using NFB.
Hats off to Charles on this one. Whether or not the absence of NFB has anything to do with this accomplishment, Charles has probably done 1001 other things right and with attention to detail.
Cheers,
Bob
stinius said:
Yes, the Ayre MXR, for example.
However, let's be clear here. I believe that Charles has a pretty strict definition of no NFB. I believe that when Charles says no NFB, he means no NFB at all, not even fairly tight local NFB, BUT he DOES allow for local degeneration.
So, if one wants to argue even stricter semantics and include emitter degeneration and emitter followers as incorporating NFB, then even Charles' amplifier would not qualify as no NFB.
I think Charles' definition of no NFB, as I understand it, is fair and reasonable.
It is definitely a significant challenge to build a really high-performing amplifier with no NFB (except for emitter degeneration and emitter followers), but it definitely can be done.
Cheers,
Bob
Ok
I have not seen the shematics of the amp, but if he has bulit an amp with not even a "fairly tight" local NFB it's a good job.
In my opinion local degeneration is a NFB.
In general:
How many stages can be used in the feedback loop before it is called NFB?
Can you take the feedback loop from the driver stage and call it no NFB, or the more popular phrase no globall NFB?
Stinius
I have not seen the shematics of the amp, but if he has bulit an amp with not even a "fairly tight" local NFB it's a good job.
In my opinion local degeneration is a NFB.
In general:
How many stages can be used in the feedback loop before it is called NFB?
Can you take the feedback loop from the driver stage and call it no NFB, or the more popular phrase no globall NFB?
Stinius
stinius said:
You have certainly hit on an area for discussion.
I think the quick answer is that it depends on how evil you think negative feedback is. If you think negative feedback is REALLY evil, you do not even allow for negative feedback around one stage, such as shunt feedback, or around a single transistor pair, such as a CFP stage.
A more moderate and pragmatic "no NFB" position is that you allow for reasonably tight, wideband negative feedback loops that do not come from the output. For most people this would constitute no global negative feedback, but even here you might get some semantical differences of opinion.
I agree that, technically, even emitter degeneration is indeed a form of negative feedback, but I don't think its worth arguing a lot about.
Cheers,
Bob
Some people confuse 'local feedback' with 'global feedback'. We worry about 'global feedback' because it usually has a time constant included in it that is called the 'dominant pole' that is often in the audio range, and sometimes below 10Hz. This means that the amount of 'global feedback' is continually changing and that the open loop gain is 90 degrees out of alignment with the input signal, IF 'global feedback' is removed or disconnected. This is what separates 'global' from 'local' feedback. Charles and I have found 'global feedback' problematic, so our BEST preamps have no 'global feedback'. We win the awards so maybe the rest of you can learn something from our example.
john curl said:
Hi John,
I largely agree with your definition of global feedback and the importance of the distinction between it and local feedback.
As we both have pointed out, however, global feedback can be arbitrarily applied with wide open loop bandwidth, and yet even under these conditions some don't like negative feedback, like Charles. Indeed, I'm pretty sure, as I stated above, that Charles does not like local NFB either.
It is also true that we can have large amounts of global NFB while still having a wide open-loop bandwidth if we have a gain crossover out in the 1 MHz to 2 MHz range. This permits 34 to 40 dB of global NFB with an open-loop bandwidth of 20 kHz. Indeed, I believe that if one wants to do this, the best way is to limit the LF gain of the VAS by means of local negative feedback (rather than passive loading of the VAS, for example).
Cheers,
Bob
To Bob Cordell and John Curl
I’m sorry that I mentioned the word “feedback” this was a tread about the brilliant Ayre amp and it was not my intention to take the focus away from the start of the tread.
I’m not fanatic against either local or global feedback, but we can discuss that matter another place. It’s absolutely possible to make a good sounding amp either with or without global feedback; it’s also very possible to make a bad sounding amp without or with global feedback, but to make an amp with absolutely no NFB seems to me like as “Bambi” on ice. and very thin ice, Unstable.
To Bob
I have read a lot of your postings today, and have read your articles years ago. I agree in a lot of the things you are writing, also about active x-over for speakers, and that the best speaker would be a combination of closed box and reflex loading. Personally I don’t like the “one note” pumping from a reflex loaded speaker, I like the sound of a closed box with qtc = 0.5 with a controlled lift in the bottom a lot better.
To John
You have also built a lot of good sounding amps like the Halo, and JC-2 as examples.
I understand that you are against global feedback, but are using local feedback. I absolutely respect that.
By the way I met one that worked in the sound crew for Grateful Death when I was in New York last year. Since I have been touring as a sound engineer for many years myself I always talk to the sound engineers when I’m at a concert.
I have worked with the construction of an amp with the ThermalTrak transistors for a while and done some simulations, quite a lot of simulations actually. After I have read different treads on this forum I see that I have used wrong SPICE models both for the transistors and the diode, lucky me. I really can’t see the point in making a data sheet, app note and spice model that is lacking important information or are completely wrong and useless.
I’m also working on a loudspeaker (system) that looks very promising.
Stinius
I’m sorry that I mentioned the word “feedback” this was a tread about the brilliant Ayre amp and it was not my intention to take the focus away from the start of the tread.
I’m not fanatic against either local or global feedback, but we can discuss that matter another place. It’s absolutely possible to make a good sounding amp either with or without global feedback; it’s also very possible to make a bad sounding amp without or with global feedback, but to make an amp with absolutely no NFB seems to me like as “Bambi” on ice. and very thin ice, Unstable.
To Bob
I have read a lot of your postings today, and have read your articles years ago. I agree in a lot of the things you are writing, also about active x-over for speakers, and that the best speaker would be a combination of closed box and reflex loading. Personally I don’t like the “one note” pumping from a reflex loaded speaker, I like the sound of a closed box with qtc = 0.5 with a controlled lift in the bottom a lot better.
To John
You have also built a lot of good sounding amps like the Halo, and JC-2 as examples.
I understand that you are against global feedback, but are using local feedback. I absolutely respect that.
By the way I met one that worked in the sound crew for Grateful Death when I was in New York last year. Since I have been touring as a sound engineer for many years myself I always talk to the sound engineers when I’m at a concert.
I have worked with the construction of an amp with the ThermalTrak transistors for a while and done some simulations, quite a lot of simulations actually. After I have read different treads on this forum I see that I have used wrong SPICE models both for the transistors and the diode, lucky me. I really can’t see the point in making a data sheet, app note and spice model that is lacking important information or are completely wrong and useless.
I’m also working on a loudspeaker (system) that looks very promising.
Stinius
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.