I was going to ask Douglas Self about the use of his blameless amps as a buffer . Sort of ties in with what you say . I suspect inverting buffer will work ? Might it tell us something useful ?
I notice that op amps that work well as buffers don't seem to me to be great at gains of lets say 2 to 10 . Above that they are great . At gain of 50 rather good . All by ear . Seems logical that they might be optimized for unity gain and that has consequences . At gain of 1 they seem not to have a sound .
I notice that op amps that work well as buffers don't seem to me to be great at gains of lets say 2 to 10 . Above that they are great . At gain of 50 rather good . All by ear . Seems logical that they might be optimized for unity gain and that has consequences . At gain of 1 they seem not to have a sound .
I am not aware of any technique to 'optimize them for gain of 1', whatever that means. I do know that some opamps are decompensated for gains above unity, to increase the available loop gain and hence the available feedback at higher gains.
jan
Thanks for two new names on the hunt for people moving forward. Found lots of new things to read.
Too it is correct. Didn't think with what it is connected?
To Jan, studying separate amplifiers, all always forget that a starting point of counting can be only temporary ratios in a signal yet not passing through the amplifier. Each link makes a certain contribution to change of temporary parameters and it doesn't pass completely.
Cordell and Self, too get to this trap, taking for a reference point the size of distortions. What price reaches decrease in distortions? Accuracy of performance of separate operations isn't connected with a signal transmission without distortions.
Baxandall too couldn't find direct link.
Last edited:
Hi Sergey,
Not sure I understand what your point is. Do you suggest that we should leave the non-linearities in the hope that they sound better? Thta's a valid reasoning - it has been shown again and again that people not necessarily prefer the cleanest 'wire with gain' amp.
Or do you suggest that making the amplifier more linear does somehow make it sound worse by definition? I have a problem imagining less accurate reproduction yet better sound? (Objectively speaking of course - subjective performance has nothing to do with technical performance unless supported by controlled testing results).
jan
Hello Jan, I think extent of influence of zones of synchronization of a signal and the amplifier is very poorly studied. For example - the moon doesn't come off the earth though force of an attraction of the sun is much more.
Therefore if zones of mutual synchronization of a signal and the amplifier are strongly carried in temporary space, there are distortions which we define in the known way. Their reduction at the expense of the speed of increase has restrictions.
Therefore if zones of mutual synchronization of a signal and the amplifier are strongly carried in temporary space, there are distortions which we define in the known way. Their reduction at the expense of the speed of increase has restrictions.
So far, I have never accepted the possibility that a goal for an audio reproduction system design should be anything other than the most-accurate reproduction, whether people preferred that sound or not. Maybe I will have to force myself to think about that again.
But even if it were true that people preferred something other than the most-accurate reproduction, and I wanted to provide what they prefer, my engineering approach might be to make the system as accurate as possible and then add mechanisms to introduce the "preferred" behaviors, whether it was upstream, downstream, or within the amplifier itself.
But even if it were true that people preferred something other than the most-accurate reproduction, and I wanted to provide what they prefer, my engineering approach might be to make the system as accurate as possible and then add mechanisms to introduce the "preferred" behaviors, whether it was upstream, downstream, or within the amplifier itself.
You don't get out much, eh?
I'm not talking audiophiles, I'm talking general public. The majority don't want accurate, the want good, cool, awesome, rockin!! They just want to enjoy it.Maybe I was restricting my thoughts to audiophile types. I am not in the music-reproduction business. So I have the luxury of usually caring only about what I, myself, think about a system. But I can say that several of my and my son's hi-fi-naïve friends who have listened to my system have enthusiastically volunteered the fact that they liked it more than anything else they had ever heard. Some of them never want to leave my listening seat. Of the very small sample, roughly one-third to one-half of them have had that reaction. Maybe some were just being overly polite. So, that agrees with your "majority don't want accuracy" but leaves a significant portion who probably do want it.
I still believe that the "good, cool, awesome, rockin", and all of the other possible types of baked-in goodness that please and thrill people, can and should come from the source, not the rest of the system. If nothing else, accurate reproduction by the rest of the system makes it more able to provide whatever it is that's preferred. i.e. "accurate" would be optimally versatile.
I still believe that the "good, cool, awesome, rockin", and all of the other possible types of baked-in goodness that please and thrill people, can and should come from the source, not the rest of the system. If nothing else, accurate reproduction by the rest of the system makes it more able to provide whatever it is that's preferred. i.e. "accurate" would be optimally versatile.
The market place is inundated with ammplifiers with strongly varying 'accuracy', and people buy all of those various amps. This proves without any doubt that people's preferences are not specifically for best accuracy.
jan
Last edited:
For gootee
I think many can tell too its, and will be right, at the same time ,it's systems different. There is the general factor which is present at them in different measure.
And if these systems are completely different? Including value harmonics and bandwidth.What is that factor?
I think many can tell too its, and will be right, at the same time ,it's systems different. There is the general factor which is present at them in different measure.
And if these systems are completely different? Including value harmonics and bandwidth.What is that factor?
Last edited:
A hunch , that's all . I will repeat the test using inverting vesrion . I feel it will be different .
Optimized for unity gain is typical a requirement . Some do it differently and offer external capacitor pins for unity . That sort of confirms my idea to me . The designer could have said just fit what it needs for unity , job done . The choice was not to .
Don't even contemplate, even in your deepest, darkest moments -
, such a course of action, Tom! I have never been stung by following the route of hearing more and more what's in the recording - and, over the years when I've heard expensive systems doctored to sound "good" by adding syrup, discarding the "unpleasant stuff", rounding off of sharp edges, I find them quite tiresome and irritating to listen to ...I just read on another forum, a member with a strongly objectivist streak who has what should be a nicely decent, revealing, musical system, proclaiming that "All but the last of Charlie Parker's recordings are pretty horrible" -- and I think, huuuh?? Okay, they're not 'brilliant' recordings, but I find a double album I have of the material very satisfying to listen to -- some tracks will brilliantly highlight every last defect in the playback system quality, makes them excellent 'measuring devices' - the recording nicely serves a double function ...
That's exactly right . Thanks .
http://www.analog.com/static/imported-files/tutorials/MT-096.pdf
The typical " blameless amplifier " is an op amp . It is subject to the same rules .
As I said in another thread we can manipulate the LTP input to have harmonics of distortion in the fashion of Jean Hiraga having suspension bridge harmonics if we like . That is easier if not using a current mirror . That will require a sensible slew rate rather than ones to impress . As far as I can see slew rate is about low impedance drive and has nothing to do with the supposed 100 watts the tweeter might require ( irony ) . High slew rate amps can sound better . It isn't the speed . They sound better playing 78's . On reflection cassettes , 78's are possibly the most demanding source . Even with a little colour it can be a blameless amp . That is one 20 dB lower in distortion than the accepted -60 dB limit . That has to extend to the sub 1 watt as best we can .
Crossover distortion . Not required . I feel class AB ( B ) can sound better than class A . Simply because the PSU isn't saturated for want of a better word . If I notice a difference it is that A is less opaque at low volume .Without a reference almost the same . The speakers I use have almost too much depth , width and space so the amp can be slightly less good there . When I say too much I play them too loud . I guess they have less than 1% distortion at 100 dB and 0.1 % at 1 watt . They will have a stab at a square wave .
They need an amp with grip . I love valves and hate distortion . The OTL of Croft might be OK . 4 ohms on a good day as the load . 100 watts 2 R with high damping factor required . It even comes close to my blameless - 80 dB . Don't much care for the look of valves , just the sound . The looks suggest it will be forever going wrong as TV's did . The Croft sounds very fast and open when I heard it . That might be 1/3 harmonic ? Sounds good .
Welcome to Croft Acoustics +44(0)1723 355111 (The official site for Croft amplification)
Don't even contemplate, even in your deepest, darkest moments -
I just read on another forum, a member with a strongly objectivist streak who has what should be a nicely decent, revealing, musical system, proclaiming that "All but the last of Charlie Parker's recordings are pretty horrible" -- and I think, huuuh?? Okay, they're not 'brilliant' recordings, but I find a double album I have of the material very satisfying to listen to -- some tracks will brilliantly highlight every last defect in the playback system quality, makes them excellent 'measuring devices' - the recording nicely serves a double function ...
I made an amplifier recently that was OK and about 5 times less distortion than typical of it's type . It's greatest quality which I have no explanation for is mediocre stuff sounds OK . The layers can be heard . Not nice distortion if wondering as a design concept . MP3 is still MP3 yet good . I had hoped for this . So many of the prototypes were very poor . One thing I did try was sharing the virtues of the amp around . The more I did that the simpler the design had to be . In the end everything was like the tale of the Three bears , baby bear was just right . The guy I made it for instantly upped the power . That was difficult for me as the test results showed it to be less than optimum . I am building him a class A2 amp now . I don't do valves usually . Seem to understand them OK . I hinted on a forum how I did the amp . I was told I would never build two the same . I have news for them that was what baby bear was all about . I have boxes of used valves . Within reason all measured the same in this amp . This is because it is baby bear and not daddy bear . The reason it works is an oversize device running 25% it's rating does most of the work . That is against 80% that is not unusual . Devices discarded years ago in the box are happy at 25% . I use resistor cathode bias also as I feel it helps the baby bear quality ( just right and durable ) . The amp is in it's own world is blameless . When thinking typical speaker distortion it is almost blameless . It will meet DIN 45500 and has no loop feedback . It does have UL . I had a prejudice against that . I was very wrong .
This to me is a very, very good marker. I would say you've done a very nice job here, and further to that you've managed to significantly reduce the distortion that typically severely mars "bad" recordings
And this could be a key aspect of what allowed the amp to perform so well - understressed electronics is a really, really good thing when you want to maximise SQ ...
The choice was not to because if you would use this opamp at larger gains, it would be overcompensated meaning that you don't get all the loop gain (and curative feedback) that you could.
Therefor, they make the comp node(s) are externally accessible, so you can set the comp for whatever gain you want, so you have max open loop gain at each required closed loop gain.
There's no optimisation for gain=1 as such, but you CAN set the compensation for stability at gain=1. But then again, you can set compensation for stability at any gain you fancy.
But maybe my understanding of the word 'optimised' is different from yours.
jan
It is a feature of humans (espcially some engineers) that when they don't understand the theory of something they will invent rules-of-thumb to enable them to successfully use the thing they don't understand. All fine. The snag comes if later they come to regard their ROT as a theory and then share it to others as a theory. Theory should always beat ROT, because a good theory will always include some idea about where the boundaries of the domain of applicability are. ROTs rarely recognise their own boundaries, so can lead people astray.
The issue of unity gain opamps, as explained by jan, is a manifestation of the general issue that feedback should not just be slapped on an amplifier (or removed) without thinking in advance about how much feedback will be needed.
The issue of unity gain opamps, as explained by jan, is a manifestation of the general issue that feedback should not just be slapped on an amplifier (or removed) without thinking in advance about how much feedback will be needed.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.