.Well you're only off by a factor of 10. (Ya just gotta love these armchair critics.)
It's like the astronomers always say, "what's a few orders of magnitude between friends?"
edit -- Please disregard these posts. I responded to Upupa Epops' posts under the mistaken assumption that they were referring to the schematic diagram I had posted. However Upupa Epops was actually referring to a *different* amplifier that had been brought up on the previous page.
It's like the astronomers always say, "what's a few orders of magnitude between friends?"
edit -- Please disregard these posts. I responded to Upupa Epops' posts under the mistaken assumption that they were referring to the schematic diagram I had posted. However Upupa Epops was actually referring to a *different* amplifier that had been brought up on the previous page.
Charles Hansen said:
I get the feeling from this that you have concluded that NFB is to be avoided all costs for ultimate performance high end amplifiers. Mmm, You may well be right.
I have less experience that you both in construction and in listening to different amps so for now my position has to be that I don't know.
but it's not that much trouble for me to continue on and test my ideas, it will not take that long now I already have most of the bits I need and more than that I really want to find out for myself how my ideas sound in reality.
I am encouraged by a few things.
I understand that SE ended designs with FB tend to sound less 'detailed' and more natural the fully complimentary designs with FB.
My current amps have an understated kind of natural and relaxed sound that I really like.
There are other amps around with FB, like the aleph, and the latest versions of the JLH simple class A, that very experienced listeners seem to regard very highly.
Anyway, if, when I get to do a comparison I prefer your amp then I probably save up for a bit and just buy one. God think of all the spare time I would have......
Many thanks for posting you schematic - very interesting and elegant. I will study it more closely.
Do you think I have hope in hell ?
regards
mike
Upupa Epops said:
First you think you can tell what the output impedance is by looking at the schematic. You were quite far off the mark there. Do you think that just maybe you might be wrong about the sound quality as well? It would be pretty tough to say, having never listened to it, no? Or do you suppose that if *you* can't make a zero-feedback amp with good bass, then nobody else can either? Just curious.
edit -- Please disregard these posts. I responded to Upupa Epops' posts under the mistaken assumption that they were referring to the schematic diagram I had posted. However Upupa Epops was actually referring to a *different* amplifier that had been brought up on the previous page.
analog_sa said:
Hello -
I don't know if there are any substitutes or if those outputs are even made any more. I don't use vertical MOSFETs any more, as explained in another thread:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=28853
What I did at the time was find the biggest available P-channel part (the RFG60P06, made by Harris which became Intersil which became Fairchild), and then measured a dozen or so likely candidates to find the best complement. That turned out to be the IRCP054. We used the version with the current-sensing leads in order to easily measure the bias current.
When using MOSFET output devices, they pretty much just keep sounding better as you raise the bias current. So you just turn it up until it's running as hot as you feel safe with. We used to use about 700 mA on this amp, but it required a pretty good heatsink to do so. The heatsink temperature was around 125 F, which you can lay your hand on for about 10 seconds before it becomes uncomfortably hot. More than that becomes a safety hazard if directly exposed.
Everything we make is fully balanced from input to output. All the power amps have a balanced bridge output stage. There are a lot of advantages to this approach.
Best regards,
Charles Hansen
Charles, thanks for the schematic. How did you stabilize the idle current without a source resistor? Is there a NPO-like point for those devices, or is the trick tight thermal coupling with the voltage source FET and picking the current through it to compensate for output device tempco? Sorry to ask such dumb stuff, but I don't have the data sheets for those devices.
You're welcome.
The first thing to remember is "don't believe everything you read". Kind of like the converse of "they wouldn't print it if it weren't true".
The zero TC point of a FET is the result of two opposite effects. As the temperature increases, the Vgs threshold voltage decreases, which makes the device turn on harder. (The tempco of the Vgs threshold voltage is also indirectly related to the drain current.) But at the same time, the channel resistance increases (like almost all conductors) as the temperature increases which limits the current in the device.
At low currentss the first mechanism dominates and you have a positive device tempco. At high currents the second mechanism dominates and you have a negative tempco. The point at which they cross over is the zero tempco point.
For vertical MOSFET power devices as used in the output stage of an audio power amplifier, the zero tempco point is typically between 5 and 15 amperes! So for normal usage in an audio amp you will need temperature compensation when using vertical MOSFETs. Source resistors really don't do much to help in this regard, although they do make a difference in other areas.
The first thing to remember is "don't believe everything you read". Kind of like the converse of "they wouldn't print it if it weren't true".
The zero TC point of a FET is the result of two opposite effects. As the temperature increases, the Vgs threshold voltage decreases, which makes the device turn on harder. (The tempco of the Vgs threshold voltage is also indirectly related to the drain current.) But at the same time, the channel resistance increases (like almost all conductors) as the temperature increases which limits the current in the device.
At low currentss the first mechanism dominates and you have a positive device tempco. At high currents the second mechanism dominates and you have a negative tempco. The point at which they cross over is the zero tempco point.
For vertical MOSFET power devices as used in the output stage of an audio power amplifier, the zero tempco point is typically between 5 and 15 amperes! So for normal usage in an audio amp you will need temperature compensation when using vertical MOSFETs. Source resistors really don't do much to help in this regard, although they do make a difference in other areas.
To Charles Hansen
Was here talked, that any boy from Moscow was build an amp, which have damping factor 2 - 3. This schematic I don't know, 'cos similar amps aren't interesting for me, because they have " muddy " bass and with this I'm stand alone. Nothing another I do not talk and I'm surprised, that you thus attack me. I don't know your amp. You or somebody may to build an amp without overall feedback and this amp may good play bass - in case, if will have good DF. If not, his bass will be muddy. It's physics, no blables.
Was here talked, that any boy from Moscow was build an amp, which have damping factor 2 - 3. This schematic I don't know, 'cos similar amps aren't interesting for me, because they have " muddy " bass and with this I'm stand alone. Nothing another I do not talk and I'm surprised, that you thus attack me. I don't know your amp. You or somebody may to build an amp without overall feedback and this amp may good play bass - in case, if will have good DF. If not, his bass will be muddy. It's physics, no blables.
For the record, both Charles and I prefer to build open loop circuits, if we can.
We find them to sound better, all else being equal.
On my testbench is a preamp that is amazingly similar to what Charles Hansen builds (for good reason, since first I saw his design ). It runs open loop. When I look at the harmonic series on a scope or FFT analyzer, it is amazingly 'pure'. This means that it contains very few higher order distortion products. Feedback circuits that I make with similar design concepts have more higher order products, not much perhaps, but much more easily measurable, than open loop designs.
Of course, feedback amps and preamps have their place, and are often necessary to meet a spec. This spec could be damping factor, THD, IM or some other spec. This does not necessarily improve the sound, and often tends to compromise it, if anything.
Now Charles might build an amp with a damping factor that is 'let's say' 30 or so. I might, with feedback, build a similar amp with an effective damping of 300 or so. Is this better? Probably no, but it looks good on a mid fi spec sheet.
Regarding the complementary feedback pair vs darlington connection.
It is difficult to separate 'feedback' from 'degeneration' as both terms are used.
The best reference of the practical difference between a complementary feedback pair and a darlington follower comes from 'Analysis and Design of Analog Intergrated Circuits' by Paul Grey/ Robert Meyer' pp 398-399.
"The major problem with the configuration 'complementary feedback pair' is the POTENTIAL INSTABILITY of the local feedback loop formed ... particularly with capacitive loads." I said it before and I say it again: The complementary feedback pair is LESS potentially free from oscillation, all else being equal.
We find them to sound better, all else being equal.
On my testbench is a preamp that is amazingly similar to what Charles Hansen builds (for good reason, since first I saw his design ). It runs open loop. When I look at the harmonic series on a scope or FFT analyzer, it is amazingly 'pure'. This means that it contains very few higher order distortion products. Feedback circuits that I make with similar design concepts have more higher order products, not much perhaps, but much more easily measurable, than open loop designs.
Of course, feedback amps and preamps have their place, and are often necessary to meet a spec. This spec could be damping factor, THD, IM or some other spec. This does not necessarily improve the sound, and often tends to compromise it, if anything.
Now Charles might build an amp with a damping factor that is 'let's say' 30 or so. I might, with feedback, build a similar amp with an effective damping of 300 or so. Is this better? Probably no, but it looks good on a mid fi spec sheet.
Regarding the complementary feedback pair vs darlington connection.
It is difficult to separate 'feedback' from 'degeneration' as both terms are used.
The best reference of the practical difference between a complementary feedback pair and a darlington follower comes from 'Analysis and Design of Analog Intergrated Circuits' by Paul Grey/ Robert Meyer' pp 398-399.
"The major problem with the configuration 'complementary feedback pair' is the POTENTIAL INSTABILITY of the local feedback loop formed ... particularly with capacitive loads." I said it before and I say it again: The complementary feedback pair is LESS potentially free from oscillation, all else being equal.
john curl said:
Great. Open loop circuits are great when you're needing voltage gains greater than 1. But when you need to reduce output impedance without reducing signal voltage what do you do? What would you recommend other than a follower of some sort for this task?
I don't know that the two words are so far apart.
I mean, negative feedback is degenerative is it not?
Yes. But then that's never been in dispute.
se
feedback x non-feedback
Well, I tried the same topology, same circuit with and without global feedback. Push-pull class A amplifier with very low values of THD and IMD. Even for non-feedback circuit it is in order of 0.001% for low power, 0,01% at 60% of output power and 0,1% close to limitation. With global NFB the values are 10x better. IMD is also very very low.
About listenning tests - global NFB version is definitely better. Better resolution, cleaner sound. More tough bass. I love to listen to the classical music - L.V. Beethoven, H. Berlioz, Shostakowich, Borodin. The non-feedback amp is far worse in resolution, instrument definition and overall sound. Most of the famous amps bring their "tail", "odeur" into this kind of music, especially non-feedback. This "odeur" might be interesting for someone (not for me) when listenning to the smoked jazz sax, but not at complex music. The feedback has to be closed around the stages as linear as possible, not curing switching spikes.
Well, I tried the same topology, same circuit with and without global feedback. Push-pull class A amplifier with very low values of THD and IMD. Even for non-feedback circuit it is in order of 0.001% for low power, 0,01% at 60% of output power and 0,1% close to limitation. With global NFB the values are 10x better. IMD is also very very low.
About listenning tests - global NFB version is definitely better. Better resolution, cleaner sound. More tough bass. I love to listen to the classical music - L.V. Beethoven, H. Berlioz, Shostakowich, Borodin. The non-feedback amp is far worse in resolution, instrument definition and overall sound. Most of the famous amps bring their "tail", "odeur" into this kind of music, especially non-feedback. This "odeur" might be interesting for someone (not for me) when listenning to the smoked jazz sax, but not at complex music. The feedback has to be closed around the stages as linear as possible, not curing switching spikes.
traderbam,
the amps can be seen on my website (click button below). PM-A1 was the non global feedback version, PM-A2 is a global NFB version (schematics shown). They have 2 stages with OA, power stage is a 6 transistor class A push-pull CFP follower. This follower was outside or inside NFB.
Frequency response is about 2Hz - 200kHz/-3dB for both non-NFB and NFB version. The damping factor is about 80 for non-NFB and about 500 for NFB.
the amps can be seen on my website (click button below). PM-A1 was the non global feedback version, PM-A2 is a global NFB version (schematics shown). They have 2 stages with OA, power stage is a 6 transistor class A push-pull CFP follower. This follower was outside or inside NFB.
Frequency response is about 2Hz - 200kHz/-3dB for both non-NFB and NFB version. The damping factor is about 80 for non-NFB and about 500 for NFB.
Re: To Charles Hansen
To Pavel Dudek,
I am *very* sorry. I thought your original post was referring to the schematic I had posted. I didn't realize that you were referring to the Russian amplifier from the previous page. Please accept my apologies for any disrepectful comments I made. I will edit the original posts to clarify my mistake, in case anyone is reading the thread from the beginning.
Best regards,
Charles Hansen
edit -- I tried to edit the original posts but cannot. There must be some sort of time limit as to how long the edit function works after posting.
Again Pavel, please accept my sincere apologies.
Upupa Epops said:
To Pavel Dudek,
I am *very* sorry. I thought your original post was referring to the schematic I had posted. I didn't realize that you were referring to the Russian amplifier from the previous page. Please accept my apologies for any disrepectful comments I made. I will edit the original posts to clarify my mistake, in case anyone is reading the thread from the beginning.
Best regards,
Charles Hansen
edit -- I tried to edit the original posts but cannot. There must be some sort of time limit as to how long the edit function works after posting.
Again Pavel, please accept my sincere apologies.
Re: feedback x non-feedback
Pavel, I really appreciate that you do not only report your listening
experiences but also tell what kind of music you evaluate with.
Many or most people never add that piece of information, and those
who do often seem to use non-classical music. Being an almost
exclusively classical listener myself I always wonder about the
relevance, for me, of the listening experiiences reported in most
cases. Some will claim that it doesn't matter what kind of music
is used for evaluation, but I stronlgy question the validity of that
argument. For instance, I get the impression that B&W speakers
are much more often favoured by classical listeners than others.
That said, also classical listener may differ in preference, but at
least we have some kind of reference, assuming we also attend
live performances.
Also, don't get me wrong that I am trying to generalize from
single data points like Pavels impressions to draw any conclusions
about NFB/non-NFB in generel. I just want to point out that I
would like many more people to mention what music they use
for evaluation, so we can make some estimate of the relevance
for our taste of music.
PMA said:
Pavel, I really appreciate that you do not only report your listening
experiences but also tell what kind of music you evaluate with.
Many or most people never add that piece of information, and those
who do often seem to use non-classical music. Being an almost
exclusively classical listener myself I always wonder about the
relevance, for me, of the listening experiiences reported in most
cases. Some will claim that it doesn't matter what kind of music
is used for evaluation, but I stronlgy question the validity of that
argument. For instance, I get the impression that B&W speakers
are much more often favoured by classical listeners than others.
That said, also classical listener may differ in preference, but at
least we have some kind of reference, assuming we also attend
live performances.
Also, don't get me wrong that I am trying to generalize from
single data points like Pavels impressions to draw any conclusions
about NFB/non-NFB in generel. I just want to point out that I
would like many more people to mention what music they use
for evaluation, so we can make some estimate of the relevance
for our taste of music.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.