Odd you should say that, John, after I has been midly criticised for my insistence on wide open loop response and low levels of GNFB not too long ago.
Not that I am phased, mind you, such amps still tend to sound better to me than those using lots of GNFB and I will stick with them.
In fact, all the talk was that one didn't have to have a wide open loop response to avoid TIM phenomena and all the rest of it, and in fact NFB was good because it linearizes the midband and the low end, and so forth, about everything except the resulting sound.
Perhaps we may find that the op amp manufacturers may, if they already haven't, make an op amp with a lower THD factor but also with a lower gain factor, with some nice degeneration resistors inside ...
Not that I am phased, mind you, such amps still tend to sound better to me than those using lots of GNFB and I will stick with them.
In fact, all the talk was that one didn't have to have a wide open loop response to avoid TIM phenomena and all the rest of it, and in fact NFB was good because it linearizes the midband and the low end, and so forth, about everything except the resulting sound.
Perhaps we may find that the op amp manufacturers may, if they already haven't, make an op amp with a lower THD factor but also with a lower gain factor, with some nice degeneration resistors inside ...
CFA DSL drivers? - TPA6120 is one with lots of audio performance plots - given the "low" open loop gain of ~70 dB it has to be very linear open loop to get those single digit ppm numbers
op amps have ben engineered for much better performance - Pease didn't plot anything from a newer isolated semi process designed for high performance in the past decade or so
SG-Acoustics · Samuel Groner · IC OpAmps has done a few though
I really don't see that you have a path forward - or any idea what direction forward is when you keep returning to 30-40 year old suppositions, speculations - and then don't appear to look at anything done since to advance understanding in the theory, modeling, measurement - and psychoacoustic controlled listening that puts those oldies into a modern perspective
op amps have ben engineered for much better performance - Pease didn't plot anything from a newer isolated semi process designed for high performance in the past decade or so
SG-Acoustics · Samuel Groner · IC OpAmps has done a few though
I really don't see that you have a path forward - or any idea what direction forward is when you keep returning to 30-40 year old suppositions, speculations - and then don't appear to look at anything done since to advance understanding in the theory, modeling, measurement - and psychoacoustic controlled listening that puts those oldies into a modern perspective
Yeah, wide open loop bandwidth is known to be BAD for audio. I mean hey, it produces all of 0.3% THD 20-20,000 Hz at full blast into 4 Ohms, and that can't be good, right? And then when you add 20 dB of GNFB, that sinks down to 0.05%, and that's even worse, right, because I haven't used my obligatory 60 dB of GNFB. So my mids can't be right?
I should listen more to the new kids on the block, the wanna-bes, who are quite happy to use more than 60 dB of GNFB and produce a slew of amps which sound mediocre at best, and I'm being kind. Also, they design five amps a day that way. which is cost efficient, right?
BTW, I never said anything to the contrary of what I know to be a fact, and that is that op amps have come a loooooong way since way back then. Quite to the contrary, I did say just that several times right here - look it up in the log. I also stated quite clearly that I always keep a stock of op amps I happen to like, mostly by AD and some from National, which is a contradiction to what you are saying.
Not that it matters - you will likely continue to be aggressive anyway, as the only possible way is your way.
I should listen more to the new kids on the block, the wanna-bes, who are quite happy to use more than 60 dB of GNFB and produce a slew of amps which sound mediocre at best, and I'm being kind. Also, they design five amps a day that way. which is cost efficient, right?
BTW, I never said anything to the contrary of what I know to be a fact, and that is that op amps have come a loooooong way since way back then. Quite to the contrary, I did say just that several times right here - look it up in the log. I also stated quite clearly that I always keep a stock of op amps I happen to like, mostly by AD and some from National, which is a contradiction to what you are saying.
Not that it matters - you will likely continue to be aggressive anyway, as the only possible way is your way.
Some of Pease's testing is in National's AN-1485, with photos at the end.
Dejan. If you ever have time could you try to find the point where feedback is so low your amplifier suffers. I suspect you could reduce yours to 12 dB and have some surprising outcomes. The bass will be slightly less tight. VAS degeneration could be changed to win a little back. I suspect what you might hear is a speed and delicacy. There is a risk of crossover distortion showing. As you use higher bias levels I suspect not. Gain of the amp will be 8 dB higher which might be nice. Hiss might intrude. What I supect is if it works you will be bemused and like it a lot. If 5 and 7 th harmonics are below - 80dB up to 40 kHz it should be OK. That's a big if I guess? Easiest way to do it is reduce lower gain resistor by shunting it ( 2 minute job ).
the intellectual and physical tools have been about for a while - I really don't see the "sounds good" crowd even trying
but I will point out when they abuse circuit, signal, feedback theory - claiming they have explanations for "amp sound" when those are wrong or easy to demonstrate highly unlikely in light of EE and/or Psychoacoustics
but I will point out when they abuse circuit, signal, feedback theory - claiming they have explanations for "amp sound" when those are wrong or easy to demonstrate highly unlikely in light of EE and/or Psychoacoustics
Like speakers, opamps should be invisible in use - if one can hear a difference when they're in circuit then they're not being used properly: wrong opamp, or poorly implemented in the device. The first power amplifier that really impressed me 30 years ago used a decent opamp, and nothing else, as the voltage gain mechanism - made all the conventional amplifiers around at the time sound pretty feeble ...
|
Please note the catches within this statement.
I'm just writing this for completeness.
You see, a speaker needs to be invisible first, like a crystal window, before we can hear any supposed invisibility which is further down in the links, in this case the amplifier. Prior to the amplifier is the DAC. Prior to the DAC is the ADC and the studio.
Just to note quickly. We are listening to the studio right? Yes, naturally, however only the vivid parts of it.
It's not very likely we can hear the faulty capacitors with high distortion or the perfect measuring capacitors in their studio equipment. If we can, in theory, that seems like a very, very fine and difficult difference.
Returning to the point......
If the speaker is tainted or shaded, i.e. not transparent, then the taintedness is very likely in some cases to up the differences it is supposed to be revealing.
Right?
Point in case......
Without invisible speakers, we can not hear the...... in theory...... invisible audio equipment.
______
Related......
What many people call "burn-in effect" in transducers, I believe is mostly just sonic, perceptual calibration.
For instance, the first time you try a pair of glasses, everything looks a little distorted and confusing. Apparently, after a week or so your vision becomes calibrated, --with and without-- the glasses.
You have developed a calibration software so to speak, so as soon as you put on the glasses, everything looks normal. However, reality still looks normal as well.
There are most likely snooker players which can play well with and without glasses.
Is audio perception any different than the above example? Likely not.
I noticed this when I was using the Etymotic ER-4 for the first time. At first, it didn't seem very transparent, I was a little disappointed in it, I couldn't hear the differences I usually hear very well.
However, over time, it became a little more transparent to me. Most likely I calibrated to it's THD, speed, time, tone, soundscape or whatever else.
I can not go back in time to check, without a time-machine I can not A/B this, you see, it's yet another example where ABX is not applicable to "the doors of perception".
Nonetheless, this phenomenon should exist, as per my visual analogy to glasses and playing snooker.
______
Lastly, there is the constant highly active filter in our minds, which is discarding information all the time.
If that filter in us can very slightly change over time, then it can let in and discard more accurately.
|
Last edited:
This is where things can get confusing, at least for some folks, 
! IME, speakers don't become invisible, systems do! But what everyone tunes into, perceives to be the most important, are the speakers - because they're the big things in front of you, they're obvious, they must be the bits that are making themselves invisible, ; because when they are not invisible you can clearly "see" the drivers working! But to me this is like blaming the tyres when a car goes off the road unexpectedly - they're the bits actually touching the road, therefore that's the part that's not right; forget about the suspension, the speed of the vehicle, the behaviour of the driver ... .
I never think to myself that the speaker has a certain, wrong "sound"; to me it's always the system that's at fault - the difficulty then is tracing the underlying cause, the more complex the combination of components the harder it becomes to properly diagnose - but IME unless one does that then a vicious cycle of never-ending fiddling and changing ensues, with relatively low chances of striking a lucky combination that one is satisfied with ...
! IME, speakers don't become invisible, systems do! But what everyone tunes into, perceives to be the most important, are the speakers - because they're the big things in front of you, they're obvious, they must be the bits that are making themselves invisible, ; because when they are not invisible you can clearly "see" the drivers working! But to me this is like blaming the tyres when a car goes off the road unexpectedly - they're the bits actually touching the road, therefore that's the part that's not right; forget about the suspension, the speed of the vehicle, the behaviour of the driver ... .I never think to myself that the speaker has a certain, wrong "sound"; to me it's always the system that's at fault - the difficulty then is tracing the underlying cause, the more complex the combination of components the harder it becomes to properly diagnose - but IME unless one does that then a vicious cycle of never-ending fiddling and changing ensues, with relatively low chances of striking a lucky combination that one is satisfied with ...
I understand that if you don't power a tube's heater, the gain suffers.
Hey, that's actually a good idea. There's a preamp (don't remember which one) where each input goes to a buffer with a cathode follower, with one cathode resistor and coupling cap shared by all. The input is selected by connecting that tube's heater. All this to avoid a selector switch contact.
I think you follow my idea, Apple earbuds are not invisible, the STAX SR-003, Sony EX1000, Olasonic TH-F4N, Vsonic GR07 mk2, Ortofon E-Q8 or Etymotic ER-4B is somewhere closer to invisible.
I'm not very familiar with high-end speakers or the closest to invisible speakers. I suspect that speakers have not reached invisibility.
If someone asserts that perfect frequency response and vanishing low THD is invisibility, well, they can assert that if they wish, it's just a theory right?
I'd like to see something close to perfect time by the way, like, a speaker with perfect impulse response / spectral decay.
Last edited:
O.k., thanks for the tip, I'll look around for measurements of full-range ribbons one day.
Last edited:
You are probably right, Nige, and I did some experimenting a while ago, but the outcome is curiously the same. If my GNFB dips below approximately 17 dB, I get that feeling that the amp is not quite in focus, that it's a bit blurred.
For all I know, it could be my imagination, but I have to live with that as well as with the sound.
If anything, it's a repetitorum. I have excatly the same feelings with my H/K 6550 and 680 integrated amps. The smaller 6550 has 17 dB of GNFB and sound better than it has any right to, in prefect focus. The 680 is more powerful (by about 2 dB nominally, but it can do MUCH more impulse current), but is in my view just a bit "out of focus", and it uses just 12 dB of GNFB. By way of exmple, the Dennsen zero GNFB also sounds a little off to me. Good sound overall, but a bit out of focus.
The odd man out is my H/K Citation 24 power amp. That one also uses 12 dB of GNFB, but sounds perfectly in focus and is fully composed.
This is where things can get confusing, at least for some folks,
I never think to myself that the speaker has a certain, wrong "sound"; to me it's always the system that's at fault - the difficulty then is tracing the underlying cause, the more complex the combination of components the harder it becomes to properly diagnose - but IME unless one does that then a vicious cycle of never-ending fiddling and changing ensues, with relatively low chances of striking a lucky combination that one is satisfied with ...
Not entirely reasonable, is it, Frank?
What if a speaker has a say pronounced hump between say 2 to 10 kHz? Which easily shows up on measurement? That's a bright speaker no matter which amp is driving it, and in fact, the better the amp, the more obviouis it becomes that the speakers are bright.
While I will not argue that the electronics do have a very important part in it all, one cannot say that they determine the entire sound of the system and simply disregard the speakers. They do have a sound of their own, but they also present a certain kind of load to the amp, which could be anything from easy and cosy to attrocious. Mine happen to be a very easy load for any imp, including the Technics sold in the supermarket, but other speakers, otherwise with a good sound, could be almost killers - the Infinity reference comes to mind, the old AR 3A Improved, the old Yamaha NS 1000, etc. Even modern speakers, like those from say B&W, or Focal, can be nasty, after all, the manufacturers themselves declare their impedance as 8 Ohms nominal, but quote minimum dips down to 3 Ohms. Such speakers definitely do have a sound of their own.
Last edited:
I guess it depends upon people are sensitive to, Dejan - some may react to subtle FR changes, but personally it's low level distortion that matters, that's what makes the big difference for me. Think of it this way: I'm listening to a live performance, and various drapes with differing acoustic properties are placed between me and the performers, attenuating the sound reaching me in different parts of the spectrum, to match the responses of different speakers. To me what would count is that I can still tell in each case that I'm listening to live music, the fact that the FR differs for each is irrelevant.
An obviously bright speaker to me says that it's emphasising treble distortion, so my solution is to fix the distortion, not change the FR.
An obviously bright speaker to me says that it's emphasising treble distortion, so my solution is to fix the distortion, not change the FR.
And how would you externally fix a frequency bump, Frank?
Look for an amp which has an equivalent depression there?
In my view, if something has abuilt in fault, that's it, you need an octave equalizer to fix that, plus a sure way of measuring the state of affairs.
I own some old Magneplanar SMGa. They have excellent bass and are durable, They sound like a Quad ESL57/63 and Spendor. They are visible in so much that they have a narrow field of stereo. Other than that they are very good. The later ones sound metallic to me.
My Eminence 12 Lta on polystyrene 2 x4 foot baffle speakers are less invisible. However they sound like real music. That is the colouration is easily mistaken as not being from them. If you like nice additions. I found by adding a 10 R series resistor I could kill the correctness of them. This was nothing to do with frequency response as I can do much to correct that actively. Simply put a PA speakers needs iron grip from the amplifier to make the miracle of a wizzer cone a low colouration device. In the 12 Lta it is very low. I would argue less is lost in this wizzer than many passive crossovers.
From the little I have done passive crossovers are very bad news. Take the SB acoustics 5 inch driver. It has a resonance at about 8 kHz. Run it on a polystyrene baffle as a full range it is highly acceptable. Add active EQ and super-tweeter, job done. Put it in a box and that 8 kHz resonance now has an enhancement device ( the box ). So you add a choke. It gets worse as now you loose the iron grip of the amp to damp it. Suddenly this lovely driver is relegated to the 1960's and chopped starting at 1.5 kHz. Bloody shame as it is linear up to 5 kHz. The solution looks to be a Seas unit made from good old paper. I will build a Voight pipe version if I can get the domestic ones to work. 7 foot should be about the same bass as the SMGa.
The great shame of this is in 20 minutes I had designed a good quality 6 pole active filter to attempt 6 kHz - 3 dB. There would also be a bass boost filter and sub bass cut filter. Lets be clear I would be using the very best op amps possible as that is a tonne of circuit to use. I won't be used as there is no market for it. I stopped short of doing it as it would depress me more if it did work.
To Joachim all of this is a walk in the park. To me it is the first time I took this exact walk and I feel very honoured to know my own stuff.
I noticed as crossover elements were added the difference between amps becomes greater (much, much). I would suggest the di/dt etc does matter when reactive.
I have to say people who say current drive a great concept I have great doubts.
Invisible speakers can also be extremely boring speakers. I have to conclude that much of the amplifier debate is linked to domesticated loudspeakers. They hold us back in the 1960's. Great fun for amp designers as something they really do not know can be inferred by listening. Those that listen most and try harder are rewarded. The customer knows nothing, if they did they would build something. Thus for all of our science listening is still King. All the massive debating that goes on here proves mostly we don't know what we a talking about. However most of us by experience know a bad one and a good one.
I learned my craft by repairing ( 23 year + ). Amplifiers from very different schools of thought can sound good. Instead of looking to how they differed I looked to see how they were the same. Often layout and power supplies were similar. Also the integrity of a design. It can be seen in cars also. I asked why the parts of Renault that are individually good yet do not make a good design. The mechanic gave a fantastic answer. Renault build to save money. The order in which they build saves them the critical amount. They know it will be very difficult to repair and take the risk.
If you are looking for a bargain amp Sugden A48 has build integrity second to none. By that I mean not a penny too much or a penny too little. Exquisite details of build using low cost materials. His brothers turntable is the forgotten Connoisseur Craftsman which is often talked about by turntable designers (that probably means never on forums). The two companies were sold when both brothers were offered teaching job I heard.
My Eminence 12 Lta on polystyrene 2 x4 foot baffle speakers are less invisible. However they sound like real music. That is the colouration is easily mistaken as not being from them. If you like nice additions. I found by adding a 10 R series resistor I could kill the correctness of them. This was nothing to do with frequency response as I can do much to correct that actively. Simply put a PA speakers needs iron grip from the amplifier to make the miracle of a wizzer cone a low colouration device. In the 12 Lta it is very low. I would argue less is lost in this wizzer than many passive crossovers.
From the little I have done passive crossovers are very bad news. Take the SB acoustics 5 inch driver. It has a resonance at about 8 kHz. Run it on a polystyrene baffle as a full range it is highly acceptable. Add active EQ and super-tweeter, job done. Put it in a box and that 8 kHz resonance now has an enhancement device ( the box ). So you add a choke. It gets worse as now you loose the iron grip of the amp to damp it. Suddenly this lovely driver is relegated to the 1960's and chopped starting at 1.5 kHz. Bloody shame as it is linear up to 5 kHz. The solution looks to be a Seas unit made from good old paper. I will build a Voight pipe version if I can get the domestic ones to work. 7 foot should be about the same bass as the SMGa.
The great shame of this is in 20 minutes I had designed a good quality 6 pole active filter to attempt 6 kHz - 3 dB. There would also be a bass boost filter and sub bass cut filter. Lets be clear I would be using the very best op amps possible as that is a tonne of circuit to use. I won't be used as there is no market for it. I stopped short of doing it as it would depress me more if it did work.
To Joachim all of this is a walk in the park. To me it is the first time I took this exact walk and I feel very honoured to know my own stuff.
I noticed as crossover elements were added the difference between amps becomes greater (much, much). I would suggest the di/dt etc does matter when reactive.
I have to say people who say current drive a great concept I have great doubts.
Invisible speakers can also be extremely boring speakers. I have to conclude that much of the amplifier debate is linked to domesticated loudspeakers. They hold us back in the 1960's. Great fun for amp designers as something they really do not know can be inferred by listening. Those that listen most and try harder are rewarded. The customer knows nothing, if they did they would build something. Thus for all of our science listening is still King. All the massive debating that goes on here proves mostly we don't know what we a talking about. However most of us by experience know a bad one and a good one.
I learned my craft by repairing ( 23 year + ). Amplifiers from very different schools of thought can sound good. Instead of looking to how they differed I looked to see how they were the same. Often layout and power supplies were similar. Also the integrity of a design. It can be seen in cars also. I asked why the parts of Renault that are individually good yet do not make a good design. The mechanic gave a fantastic answer. Renault build to save money. The order in which they build saves them the critical amount. They know it will be very difficult to repair and take the risk.
If you are looking for a bargain amp Sugden A48 has build integrity second to none. By that I mean not a penny too much or a penny too little. Exquisite details of build using low cost materials. His brothers turntable is the forgotten Connoisseur Craftsman which is often talked about by turntable designers (that probably means never on forums). The two companies were sold when both brothers were offered teaching job I heard.
- Status
- Not open for further replies.