Oh dear! I was afraid that someone with such profound misunderstanding would turn out to be commercially involved with audio. This seems to be surprisingly common.atmasphere said:
Please stop bleating on about paradigms. There are not two different definitions of output impedance. The definition is not based on a paradigm; it truly does not care whether you want current drive, voltage drive or something in between. Output impedance is output impedance. There are not multiple definitions, depending on which universe you happen to live in or which is your favourite colour. There is output impedance, and lots of other things which are not output impedance but which may be confused by some with output impedance.
If the amp is non-linear (in gain or impedance) then this will give you a false result. In this case you would not be measuring output impedance but some measure of driving ability. If the amp is linear then this will give exactly the same result as a correct measurement using the 'voltage paradigm' or the 'current paradigm' or the 'balance a crystal on my nose' paradigm.
You are arguing with Jan, SY, me and others. You are arguing with every electronics textbook which has ever been written. It is foolhardy.
It is a characteristic of some who don't really understand a subject (in some cases despite making a living doing it) that they brush aside valid criticism as being from a different paradigm. This means they want their own personal version of physics, although they rarely understand enough physics to realise that this is what they are doing.
Looks interesting and good of you to share. Is this tried and tested or just a simulator exercise? Does it sound good?
Hi MiiB,
One would assume this is what those amps I've been reading glowing reviews of in the Hifi mags are doing if they really are zero GNFB. I suppose a CFP output stage would fit this "gain of one heavy feed back OPS" category? But more likely to be something fancier like dadod's schematic?
Looks interesting and good of you to share. Is this tried and tested or just a simulator exercise? Does it sound good?
There ia athread http://www.diyaudio.com/forums/solid-state/260308-gainwire-ngnfb-classb-poweramp.html. It is simulated only, but uses parts as IPS and OPS tested in different amps, Error Correction is not tested, at least by me.
It would be heroic effort to build such circuit as diy experiment, just to see how it works. But for some lab where the staff is generously funded by private grants, it's feasible.🙂
Last edited:
Dadod, this ops can run with a simpler voltage stage, no buffer needed because of the JFet input.
Spechs are much better than any error correction as it uses heaps of feedback, it does ppm at 20 kHz close to rail svinging 120V P/P.
The schematic show is concept, the actual circuit built has different current sources and also a servo for offset.
Spechs are much better than any error correction as it uses heaps of feedback, it does ppm at 20 kHz close to rail svinging 120V P/P.
The schematic show is concept, the actual circuit built has different current sources and also a servo for offset.
Attachments
Sorry the schematic can be seen here
http://www.diyaudio.com/forums/soli...eedback-huge-benefit-audio-2.html#post4408051
Here's the actual build
http://www.diyaudio.com/forums/soli...eedback-huge-benefit-audio-2.html#post4408051
Here's the actual build
Attachments
Please tolerate one more silly question. Can feedback be given to specific frequency range ?
Regards.
Regards.
Jan it's actually an OPS with just the VBE onboard, then I can add different IPS, so I can test builds of different topologies. Both in measurements and how they sound in my systems.
Thank you sir. By that I was thinking if one can reduce distortion to specific range which are prominently discernible leaving the rest as it is. Best of both the worlds ?
Regards.
Regards.
That would then introduce a non linear amplifier,with frequency dependant gain. Not very good for fedelity but maybe good for tone control
Last edited:
And Henry has been dead for a while.
Sounds like that is the keeper comment. Nelson, IMO is one of the leading designers in the field, and the sort of thing on which high end audio thrives.
Field coils have been a rising star in high end audio FWIW. They did not go away in the old days because of inferior performance. They went away because they were more expensive. But they have a property similar to ESLs, which is that the magnetic field in the gap is unperturbed by the application of current in the voice coil (relative to a permanent magnet) because the motive field is powered by an external power supply. This makes them lower distortion.
I did note that your comment about the paper has changed- now its about a 50s problem (still with us) rather than just open baffles.
Hello Jan, both you and DF96 below don't seem to grasp that I was outlining how the paradigm works. Its incorrect to say that I don't understand this stuff. Even though many consider me to be very woo because I make an OTL, often they are quite surprised that I rely very heavily on bench tests and math to sort out performance (listening is last to confirm).
I totally get what you are saying because I'm in contact with life and was taught this stuff in school. The reason I brought up the power paradigm is because its a method and a means of achieving a lack of tonal coloration without feedback. I don't remember what first put me on to it, but being old enough to remember some of this stuff when it was new and also noticing the rise of SETs about 25 years ago led to me winnowing out how the stuff worked. I was already aware at the time that feedback if poorly applied to an amplifier was worse than none at all, provided the amp could make flat frequency response on the speaker used. So I made it a bit of a hobby to sort out how things were being done. Mind you, there was no-one back in the old days (as I pointed out earlier, apparently you missed that post) that thought they were using power paradigm rules as opposed to voltage rules. So I documented what the rules were and are, and can now point out equipment that use those rules as in the world of high end audio they are still being used in some sectors. Sure its a niche, but what I find interesting is that after all this time this is still going on. But after all this time the LP is also increasing in sales (has since 1993) and vacuum tubes (which are much more sophisticated devices than they appear at first blush) are still very much around too.
Also as I mentioned before usually the succeeding art eclipses the prior art which fades into history. I used side valves in an engine as an example. But the nature of high end audio is such that a lot of stuff goes on- tubes being a great example- where what's old is new again. You don't have to know anything about electronics to know why tubes are still around, all you need to know is market forces- people like them even though considered 'obsolete'.
not to put too fine a point but you seem to have not comprehended my words in quoting them above; in fact are contradicting them as if you want to draw me into an argument. This is a classic example of trolling.
Reason why mr Pass found some good in current drive for ribbons are that they basically are the only drivers with linear impedance, they don't suffer from resonances and as such they can be driven with current as well as voltage. It would not matter, as long as you don't try to make a standard passive crossover on it. Try to drive any other electromechanical driver with current and your in for a lot of trouble.
Last edited:
However Nelson designed a proper transconductance amplifier and bothered to document it thoroughly how to adjust both ends of the spectrum as well as crossover design. Very different from just munging the feedback. He also is at great pains to point out that this is a specific application and not something most would worry about.
No, I am pretty sure I said on first reading it was a 50s problem. And although I didn't mention it that doesn't mean I have changed my mind. For the record I like OB designs.
post 227:
However the text following this statement confirms your position. So we both were right on that one. I like OB designs too. I saw one that Nelson built up at RMAF a few years ago. It did an excellent job of playing bass, especially when you consider how small the whole thing was!
It seems to me unlikely that you can just 'munge' (if that is a word) the feedback on most amps without extremely deleterious results. You have to built up an amplifier that is linear without feedback. They do exist- although often cited for poor specs, SETs (if used correctly, which is to say with a speaker of sufficient efficiency) have a property of distortion linearly decreasing to unmeasurable at low power unlike most push-pull and class D amps. This is why they are known for such great 'inner detail' as most listening occurs below 1 watt.
If you push an SET past about 20% of full power, the higher ordered harmonics come into play (and a very interesting interplay insues between the ear/brain system and the way the amp distorts, leading many audiophiles to say that SETs are very dynamic belied by their low power. IMO you can safely replace the word 'dynamics' with 'distortion' in 90% of all audiophile conversation without changing the meaning of the conversation at all. I'm serious about that so no smilie.) .
So IMO an SET is improperly set up if the speaker used lacks efficiency such that the higher orders come into play. This is also why I am not a fan of SETs, as its really not the easy to find a speaker that works with them. Its even harder than it is with OTLs!
Well I think you can safely assume that both DF96 and I know how (possibly frequency dependent) output impedance of an amp interacts with the frequency dependent impedance of a speaker.
I for one have still no idea what the 'power paradigma' is so that part of your explanation unfortunately failed.
Jan