Very Good Posts!!!
I just wanna say that the latest posts from Eric, Frank and Hugh are very much "spot on" and it seems as this knowledge is missing in some participants in this thread
Lumped THD+N figures means NOTHING on their own, at least at levels of 0.001% and higher.
High or low order, odd or even and also the PHASE of the harmonics plays a big role in how an amp sounds.
mikek wrote;
"I protest......why should you or anyone else consider '******' second best?
i suggest they, like amplifiers have their uses and provide an indispensable service... "
Somehow I´m not surprised a post like this comming from you.
/Peter
I just wanna say that the latest posts from Eric, Frank and Hugh are very much "spot on" and it seems as this knowledge is missing in some participants in this thread
Lumped THD+N figures means NOTHING on their own, at least at levels of 0.001% and higher.
High or low order, odd or even and also the PHASE of the harmonics plays a big role in how an amp sounds.
mikek wrote;
"I protest......why should you or anyone else consider '******' second best?
i suggest they, like amplifiers have their uses and provide an indispensable service... "
Somehow I´m not surprised a post like this comming from you.
/Peter
Less FB Is Better ?.
I have heard a couple of tube amplifiers with switchable NFB ratio - 0 FB, 10dB and 20dB.
In both cases, the sound changes quite markedly.
At 20dB, the frequency range was extended in highs and lows, and distortion was low.
In 0 Fb position, the response tailed off in the highs and the lows, the distortion was slightly higher, but the overall sound was rather smoother, and had a very nicely natural decay response (guassian).
This ties in with theory as regards higher harmonics production, and the ears dislike of such higher harmonics.
Eric.
I have heard a couple of tube amplifiers with switchable NFB ratio - 0 FB, 10dB and 20dB.
In both cases, the sound changes quite markedly.
At 20dB, the frequency range was extended in highs and lows, and distortion was low.
In 0 Fb position, the response tailed off in the highs and the lows, the distortion was slightly higher, but the overall sound was rather smoother, and had a very nicely natural decay response (guassian).
This ties in with theory as regards higher harmonics production, and the ears dislike of such higher harmonics.
Eric.
Visuals Should Not Matter.
There has been talk here of the percieved physical appearance of an amplifier affecting the percieved sonic characteristics.
I agree that this factor may influence the less experienced listener, but for me, I have seen the outside AND inside, AND repaired AND heard so many amplifiers, that appearances are totally irrelevant to my sonics apraisals.
Also, the fact that I do not own these amplifiers further divorces any attatchment.
To me an amplifier is an amplifier, and aesthetics do not influence me in the slightest.
My criterion is sonics, and I find it easy to quickly discern whether I like the sound of a particular amplifier or not, largely because I have a very large library of 'stored' sounds to compare against, and further because I am very familiar with live and studio sounds, vocals and instrumentals.
If the on-test amplifier satisfies my sonics criterion, then my second criterion is design and build quality, and long term reliability.
Visuals I do not care about, except to say that some 'high-end' examples are just silly.
In the pro-audio field, nobody cares a damm about the front panel visuals of an amplifier mounted in a rack.
What is cared about is reliability and sonics, and there is good correlation between what I hear in such amplifiers on my test bench, and what colleagues who use them say about their sonics.
This applies to mixing desks and outboard effects units too.
Amplifiers that I have commented on as being 'ear bleeders', are generally regarded in the same light.
Mackie amplifiers are a good example of harshness, and amplifiers like Amcrons are regarded as nicely neutral, and others fall in between.
Eric.
There has been talk here of the percieved physical appearance of an amplifier affecting the percieved sonic characteristics.
I agree that this factor may influence the less experienced listener, but for me, I have seen the outside AND inside, AND repaired AND heard so many amplifiers, that appearances are totally irrelevant to my sonics apraisals.
Also, the fact that I do not own these amplifiers further divorces any attatchment.
To me an amplifier is an amplifier, and aesthetics do not influence me in the slightest.
My criterion is sonics, and I find it easy to quickly discern whether I like the sound of a particular amplifier or not, largely because I have a very large library of 'stored' sounds to compare against, and further because I am very familiar with live and studio sounds, vocals and instrumentals.
If the on-test amplifier satisfies my sonics criterion, then my second criterion is design and build quality, and long term reliability.
Visuals I do not care about, except to say that some 'high-end' examples are just silly.
In the pro-audio field, nobody cares a damm about the front panel visuals of an amplifier mounted in a rack.
What is cared about is reliability and sonics, and there is good correlation between what I hear in such amplifiers on my test bench, and what colleagues who use them say about their sonics.
This applies to mixing desks and outboard effects units too.
Amplifiers that I have commented on as being 'ear bleeders', are generally regarded in the same light.
Mackie amplifiers are a good example of harshness, and amplifiers like Amcrons are regarded as nicely neutral, and others fall in between.
Eric.
Re: Less FB is Better?
Eric:
>This ties in with theory as regards higher harmonics production, and the ears dislike of such higher harmonics.<
I have only read the Matti Otala paper in Japanese, so it is possible that there were discrepencies from the original document, but at least it appeared to me that in a couple of places the math didn't make sense. With the equations "corrected" (and hence subject to whatever limitations my own abilities and insights may entail), I did not feel that low NFB was necessarily the conclusion that I would have drawn, were I in Matti's shoes. John Iverson independently worked out a very similar theory as Otala, but chose to keep the results for himself. It would have been nice to have spent time with John to note his insights and perspective.
OTOH, I will not deny that there is a definite relation between greater global NFB and increased emphasis on higher-order distortion products. And I agree that there are perceptible sonic effects as well as calculated and measured ones. But is it appropriate to make observations about subjectively appropriate amounts of global NFB from one amplifier design and try to apply those observations to a completely different design?
I have conducted some controlled listening tests regarding the amount of global feedback and subjective sonic preferences. And when I am preparing a new design, I will always conduct such tests. Note that I prefer to control the amount of global feedback through resistors soldered to the circuit board rather than through wired switches.
On my Connoisseur 4.0 amplifier, the outcome was that we were able to apply well in excess of 50dB of global feedback before I and my listening panel began to feel that there were audible detriments. But in other situations with different circuits, the general subjective perception was that much lower amounts of global NFB were preferable. Hence, my present thoughts about global NFB is that the appropriate amount should be considered in relation to the specific design, rather than being applied as a blanket generalization in every situation.
hth, jonathan carr
Eric:
>This ties in with theory as regards higher harmonics production, and the ears dislike of such higher harmonics.<
I have only read the Matti Otala paper in Japanese, so it is possible that there were discrepencies from the original document, but at least it appeared to me that in a couple of places the math didn't make sense. With the equations "corrected" (and hence subject to whatever limitations my own abilities and insights may entail
), I did not feel that low NFB was necessarily the conclusion that I would have drawn, were I in Matti's shoes. John Iverson independently worked out a very similar theory as Otala, but chose to keep the results for himself. It would have been nice to have spent time with John to note his insights and perspective.OTOH, I will not deny that there is a definite relation between greater global NFB and increased emphasis on higher-order distortion products. And I agree that there are perceptible sonic effects as well as calculated and measured ones. But is it appropriate to make observations about subjectively appropriate amounts of global NFB from one amplifier design and try to apply those observations to a completely different design?
I have conducted some controlled listening tests regarding the amount of global feedback and subjective sonic preferences. And when I am preparing a new design, I will always conduct such tests. Note that I prefer to control the amount of global feedback through resistors soldered to the circuit board rather than through wired switches.
On my Connoisseur 4.0 amplifier, the outcome was that we were able to apply well in excess of 50dB of global feedback before I and my listening panel began to feel that there were audible detriments. But in other situations with different circuits, the general subjective perception was that much lower amounts of global NFB were preferable. Hence, my present thoughts about global NFB is that the appropriate amount should be considered in relation to the specific design, rather than being applied as a blanket generalization in every situation.
hth, jonathan carr
NFB Is Not Just NFB
Hi Jonathon,
I did not mean that NFB is a bad thing per se, nor did I really mean that low NFB is always better.
Of course the overall circuit is of paramount importance, and NFB implementation is similarly important.
I did recently hear an amplifier that uses two feedback paths - one from the output point to ensure DC stability, and another from the output of the drivers stage for above 4kHz I think.
I have not seen the schematic so that is all the info that I have to go on at present.
I need to have a better listen on my own gear to properly learn this amplifier, but first impressions were that it is not reactive in the sense of a typical NFB amplifier, and maybe erred on the side of politeness, maybe.
Implementation of NFB is parameter in amplifier design amongst many.
Eric.
Hi Jonathon,
I did not mean that NFB is a bad thing per se, nor did I really mean that low NFB is always better.
Of course the overall circuit is of paramount importance, and NFB implementation is similarly important.
I did recently hear an amplifier that uses two feedback paths - one from the output point to ensure DC stability, and another from the output of the drivers stage for above 4kHz I think.
I have not seen the schematic so that is all the info that I have to go on at present.
I need to have a better listen on my own gear to properly learn this amplifier, but first impressions were that it is not reactive in the sense of a typical NFB amplifier, and maybe erred on the side of politeness, maybe.
Implementation of NFB is parameter in amplifier design amongst many.
Eric.
Eric:
>I have not seen the schematic so that is all the info that I have to go on at present.<
The behavior of an amplifier is strongly affected by the board layout and physical construction, not only the schematic. I have taken amplifiers that showed definite signs of overshoot and high-frequency ringing, and made them measurably stable by simply rebuilding the feedback network so that it was physically as short and direct as possible. In the process, the operating temperature dropped somewhat, too.
For the same reason, the phase compensation network needs to be considered on the basis of the specific environment, including the board layout and physical construction as well as the schematics.
regards, jonathan carr
>I have not seen the schematic so that is all the info that I have to go on at present.<
The behavior of an amplifier is strongly affected by the board layout and physical construction, not only the schematic. I have taken amplifiers that showed definite signs of overshoot and high-frequency ringing, and made them measurably stable by simply rebuilding the feedback network so that it was physically as short and direct as possible. In the process, the operating temperature dropped somewhat, too.
For the same reason, the phase compensation network needs to be considered on the basis of the specific environment, including the board layout and physical construction as well as the schematics.
regards, jonathan carr
NFB.
Hi,
Absolutely, blanket generalisations are always dangerous.
If at all possible I'd rather use moderate amounts of local FB than huge amounts of global NFB.
Wherever feasable I think it's wise to choose the building blocks on a basis of required gain and linearity.
On some occasions adding an extra building block may be preferable to allow for less global feedback.
What I do regret however is that due to the bad design of CDPs amplifier designers need to resort to "euphonic" distortion mechanisms to make it sound "right" again.
While it may seem pleasant at first it still is a deviation from the truth and while I can't wait for better recordings to appear I also think the problem is also present in the digital replay chain.
On better replay systems some CDs that where hitherto "unlistenable" on more modest gear can really shine on the better ones.
Pitch stability often seems to be the cause and no doubt OS techniques and jitter may be other culprits .
I doubt that all these timing deviations etc. are really easily measure and countered for, at least that would explain why so many manufacturers are unable to produce CDPs that are listenable long term.
As far as amplifiers go the same often applies, when one might find oneself at awe with looks, specs and a gigantic pricetag...if however I find myself not tapping my foot to the rythm of the music on the appropriate material the whole amp is not worth a dime to me.
For want of a better word, if it isn't "musical" than I can't care less how well it measures.
That's of course just my opinion but to me at least that's where the basics of a good design should be.
In conclusion, a distortion figure on it's own is rather meaningless.
What's important IMHO is to know what it's made of, i.e. its' harmonic content.
\stop rant.
Cheers,
Hi,
Absolutely, blanket generalisations are always dangerous.
If at all possible I'd rather use moderate amounts of local FB than huge amounts of global NFB.
Wherever feasable I think it's wise to choose the building blocks on a basis of required gain and linearity.
On some occasions adding an extra building block may be preferable to allow for less global feedback.
What I do regret however is that due to the bad design of CDPs amplifier designers need to resort to "euphonic" distortion mechanisms to make it sound "right" again.
While it may seem pleasant at first it still is a deviation from the truth and while I can't wait for better recordings to appear I also think the problem is also present in the digital replay chain.
On better replay systems some CDs that where hitherto "unlistenable" on more modest gear can really shine on the better ones.
Pitch stability often seems to be the cause and no doubt OS techniques and jitter may be other culprits .
I doubt that all these timing deviations etc. are really easily measure and countered for, at least that would explain why so many manufacturers are unable to produce CDPs that are listenable long term.
As far as amplifiers go the same often applies, when one might find oneself at awe with looks, specs and a gigantic pricetag...if however I find myself not tapping my foot to the rythm of the music on the appropriate material the whole amp is not worth a dime to me.
For want of a better word, if it isn't "musical" than I can't care less how well it measures.
That's of course just my opinion but to me at least that's where the basics of a good design should be.
In conclusion, a distortion figure on it's own is rather meaningless.
What's important IMHO is to know what it's made of, i.e. its' harmonic content.
\stop rant.
Cheers,
Problem Circuit Boards ?.
Hi Jonathon,
Thankyou, that is very good information.
Regarding layouts, when doing output null testing, I find that mirror imaged layouts are usually bad, and dual mono is usually better.
I agree that board and component placement parasitics have large influence.
I once built a P-P wired cdp Dac and output/filter stage.
This gave the biggest cleanest and most dynamic replay that I have ever heard, but alas was physically impractical, and some bits got salvaged for the next audio experiment.
Some day, I will make a better and more practical example.
This implementation did teach me about the importance of a proper ground common, and power supply reticulation.
I had to use only two lengths of wire between stages (12mm long), and I found the type of wire and the direction of the wires to have very strong influence, surprisingly so.
Eric.
Hi Jonathon,
Thankyou, that is very good information.
Regarding layouts, when doing output null testing, I find that mirror imaged layouts are usually bad, and dual mono is usually better.
I agree that board and component placement parasitics have large influence.
I once built a P-P wired cdp Dac and output/filter stage.
This gave the biggest cleanest and most dynamic replay that I have ever heard, but alas was physically impractical, and some bits got salvaged for the next audio experiment.
Some day, I will make a better and more practical example.
This implementation did teach me about the importance of a proper ground common, and power supply reticulation.
I had to use only two lengths of wire between stages (12mm long), and I found the type of wire and the direction of the wires to have very strong influence, surprisingly so.
Eric.
Re: Leach FB
No, I haven't heard it, but I'm consedering building one to
find out how it sounds. Anyway, the feedback principle should
be possible to migrate to almost any other amp with global
NFB. Might be interesting to compare two feedack principles
within the same topology.
P.S. this will be my last post for a couple of days, so don't expect
me to continue discussion right now.
mrfeedback said:
No, I haven't heard it, but I'm consedering building one to
find out how it sounds. Anyway, the feedback principle should
be possible to migrate to almost any other amp with global
NFB. Might be interesting to compare two feedack principles
within the same topology.
P.S. this will be my last post for a couple of days, so don't expect
me to continue discussion right now.
Re: Problem Circuit Boards ?.
Although it may seem hard to believe to some this is very true. I'm using 3 pcs of 1" wire in my GC for ground connection. Just out of curiosity I tried solid core copper and solid core silver in two amps for comparison. Everybody preffered copper, since silver didn't give enough "edge" and was simply too smooth.
Also, the appearance of the amp doesn't mean nothing to me.
If I find myself not tapping my foot to the rythm of the music on the appropriate material the whole amp is not worth a dime to me. Right, Frank?
mrfeedback said:
Although it may seem hard to believe to some this is very true. I'm using 3 pcs of 1" wire in my GC for ground connection. Just out of curiosity I tried solid core copper and solid core silver in two amps for comparison. Everybody preffered copper, since silver didn't give enough "edge" and was simply too smooth.
Also, the appearance of the amp doesn't mean nothing to me.
If I find myself not tapping my foot to the rythm of the music on the appropriate material the whole amp is not worth a dime to me.
Right, Frank?
Wires Can Be Problems Too.
Peter,
In my experience, you should make those connections non-directional.
See one of my recent posts for how to make non-directional interconnects.
Imo, this directional conductor issue is a PITA actually, but the effects can be much lessened by intelligent usage of non-directional connections - upstream problems can be reduced markedly.
In the case of the P-P Dac, the sonic effects of differing wire types and their direction were sonically extremely apparent.
I tried a half dozen wire types and got six quite strongly different sounds.
In this sample set were telephone wire, multi-strand wire, resistor lead wire, litz wire and super pure copper with silver plating.
The silver plated super pure copper wire had a thin teflon sleeve that was found to influence the sound very strongly in the Dac application, and also when covering interconnect cable strands.
Just some of my experiences.....
Eric.
Peter Daniel said:
Although it may seem hard to believe to some this is very true. I'm using 3 pcs of 1" wire in my GC for ground connection. Just out of curiosity I tried solid core copper and solid core silver in two amps for comparison. Everybody preffered copper, since silver didn't give enough "edge" and was simply too smooth.
Also, the appearance of the amp doesn't mean nothing to me.
If I find myself not tapping my foot to the rythm of the music on the appropriate material the whole amp is not worth a dime to me.
Peter,
In my experience, you should make those connections non-directional.
See one of my recent posts for how to make non-directional interconnects.
Imo, this directional conductor issue is a PITA actually, but the effects can be much lessened by intelligent usage of non-directional connections - upstream problems can be reduced markedly.
In the case of the P-P Dac, the sonic effects of differing wire types and their direction were sonically extremely apparent.
I tried a half dozen wire types and got six quite strongly different sounds.
In this sample set were telephone wire, multi-strand wire, resistor lead wire, litz wire and super pure copper with silver plating.
The silver plated super pure copper wire had a thin teflon sleeve that was found to influence the sound very strongly in the Dac application, and also when covering interconnect cable strands.
Just some of my experiences.....
Eric.
- Status
- Not open for further replies.