High praise indeed, Fred.
Coming from you that fills me with a nice warm glow 😎
Hope you didn't mind me stealing the circuit description under 'interesting circuits'.
Andy.
Coming from you that fills me with a nice warm glow 😎
Hope you didn't mind me stealing the circuit description under 'interesting circuits'.
Andy.
Hi,
Allow me to go off topic for a minute...it's hot enough as it is, right?
Regarding xover notch distortion, I have a paper by one designer that no such thing exists!!!
Assuming he meant switching from class A to B operation...
While I hope it doesn't, I have a hard time believing the claim...
Any quick thoughts?
BTW, the designer of the NAIM amps as I know them is/was Julian Vereker.
Cheers,😉
Allow me to go off topic for a minute...it's hot enough as it is, right?
Regarding xover notch distortion, I have a paper by one designer that no such thing exists!!!
Assuming he meant switching from class A to B operation...
While I hope it doesn't, I have a hard time believing the claim...
Any quick thoughts?
BTW, the designer of the NAIM amps as I know them is/was Julian Vereker.
Cheers,😉
THIS JUST CAUGHT MY EAR...
Hi,
I don't agree with that statement.
Doing away with any content above and beyond 20K is audible IMHO.
But that's another topic, I reckon...
😉
Hi,
I don't agree with that statement.
Doing away with any content above and beyond 20K is audible IMHO.
But that's another topic, I reckon...
😉
Koinichiwa,
Nope.
Actually, most came through the workshop for testing. Some because they where blown up by some idiots too.
Actually, I not anywhere I believe claimed that I am "a well known designer with a long pedigree of commercial successes behind you". As for Naim gear, all but the most recent generation has been passed through my hands. Some of it was measured and listened to and other only listened to.
As to your point of: "The input signal is then limited by a very high quality low-pass filter within the preamp, that limits bandwidth to that which is audible. It does this without introducing any musically important anomalies (e.g. phase etc)."
Actually, the lowpass filter does introduce phaseshift. It is really easy to measure too. As are the various distortions (nor am i the first by far to point them out).
As for the audibility of signals above 20KHz, maybe you should read:
T. Oohashi et al, "High-Frequency Sound Above the Audible Range Affects Brain Electric Activity," AES Preprint No.3207 91st Convention, New York
I have little interest to argue further on the subject of Naim gear. It has it's followers and their usual religous zeal about it is sufficient to make me avoid long arguments.
If anyone wants to know what is really going on I recommend to take measurements by themselves. Naim gear is not that hard to find.
Sayonara
ALW said:
Nope.
ALW said:
Actually, most came through the workshop for testing. Some because they where blown up by some idiots too.
ALW said:
Actually, I not anywhere I believe claimed that I am "a well known designer with a long pedigree of commercial successes behind you". As for Naim gear, all but the most recent generation has been passed through my hands. Some of it was measured and listened to and other only listened to.
As to your point of: "The input signal is then limited by a very high quality low-pass filter within the preamp, that limits bandwidth to that which is audible. It does this without introducing any musically important anomalies (e.g. phase etc)."
Actually, the lowpass filter does introduce phaseshift. It is really easy to measure too. As are the various distortions (nor am i the first by far to point them out).
As for the audibility of signals above 20KHz, maybe you should read:
T. Oohashi et al, "High-Frequency Sound Above the Audible Range Affects Brain Electric Activity," AES Preprint No.3207 91st Convention, New York
I have little interest to argue further on the subject of Naim gear. It has it's followers and their usual religous zeal about it is sufficient to make me avoid long arguments.
If anyone wants to know what is really going on I recommend to take measurements by themselves. Naim gear is not that hard to find.
Sayonara
Well having done the tests myself, and also having good corroboration from some well-respected audio types, pre-biased to agree with your statement, I'd modify it accordingly: -
"Doing away with any content above and beyond 20K badly is audible IME"
The Listening Test
As with many things in audio, details matter. If the cure is worse than the disease, there may be a better cure, rather than live with the disease (and not dealing with the bandwidth of the signals you amplify, results in some serious extension of the performance criteria for subsequent stages, all of which is unnecessary).
Andy.
OKIDOKI...
Hi,
Fair enough...
I do prefer a wide bandwidth design but I figure that's a matter of taste.
Oh, I am not biased in any way, I just know what I like and what not...I regulate my tubes in an unconventional way already so, I'll be the last to critise anyone elses little sins.😉
Hi,
Fair enough...
I do prefer a wide bandwidth design but I figure that's a matter of taste.
Oh, I am not biased in any way, I just know what I like and what not...I regulate my tubes in an unconventional way already so, I'll be the last to critise anyone elses little sins.😉
Kuei Yang Wang said:
Yes, that claim caught my attention too. I'd be interested in hearing how they manage to implement a low-pass filter without any phase shift.
However, ANY delay will result in phase shift. Phase shift is meaningless in and of itself. What counts is whether all frequencies are arriving at the same time. Or rather at their proper relative times. And phase shift does not, in and of itself, change arrival times. A simple, frequency-independent delay, i.e. one which delays all frequencies equally, will result in phase shift, yet it won't effect arrival times.
What you want to look at insead is not phase shift, but group delay. If group delay is flat, then arrival times are not altered, regardless of any phase shift.
se
What I'm certain you hear is the effect of filtering, not the effect of bandwidth limitation.
Not easy to solve though, but it is possible.
Andy.
Not easy to solve though, but it is possible.
Andy.
HOLD ON...
Hi,
Not sure whom you're responding to...
Well, I don't filter at all and both pre and amp run well into the + 100kHz range...
If and when I resort to filtering I can't think of any better way than using a xformer...again, just my preference.
Cheers,😉
Hi,
Not sure whom you're responding to...
Well, I don't filter at all and both pre and amp run well into the + 100kHz range...
If and when I resort to filtering I can't think of any better way than using a xformer...again, just my preference.
Cheers,😉
this is becoming a full time job........
"Interestingly, non of these circuits are commercial, the SE Amplifiers that are actually being manufactured with minimal exceptions all operate with outputs that amounts to common cathode/emitter/source (shunt) mode. Moreover, so do most DIY Amp's."
And once one observes the distortion spectrum and the load interaction of a common anode/collector/drain (aka follower) mode circuit and actually listens to the resulting sound it becomes clear why."
Commercilal offering have the constrain of efficiency and power output ratings that are not concerns of many DIY builders. The Aleph amps are not strictly singal ended but have a dynamic current source for the increase in efficiency. Thre are some single ended follower amps on the commercial market. the Pathos amps come to mind. I and many others prefer the sound of follower output stages. Heavily biased Mosfet or BJT followers are some of the best sounding amps I have heard. No loop feedback makes them very easy to make stable. I think it is the topology of choice for amateur constructors for that reason. Compensating feedback amps with negative feedback can be a real challenge for the amateur designer and a pain for even experienced designers. I don't think the distortion argument is really that legitimate either as the non linearity of the transconductance is the dominant mechanism in either topologies. A common source output stage has even more interaction with the load since the load now controls the open loop voltage gain with serious variations in gain with frequency with most speaker loads. The open loop voltage gain of a follower circuit is much more independent of load. You won't get low output impedance with common source without a lot of negative loop feedback. There is no such thing as a free lunch.
http://www.pathosacoustics.com/indexeng.htm
I have nothing against shunt regulators since they have the potential for excellent PSSR. The factors influencing transient response noise, stability, and output impedance are the same as those for a series regulator. You can design a poor shunt regulator about as easy as a poor series regulator and the same basic topology can be applied to either with change in output device. You can even design a shunt regulator with a three terminal series regulator IC with not much trouble. I like simple topologies and single ended follower output has the advantage of less complexity and the lack of need for bias circuits and emitter degeneration. Simplicity makes it much easier to obtain stability for high bandwidth and gain feedback circuits like the Jung regulator. if more linearity is required it is quite easy to increase the preloaded DC current for shunt or series regulators. It even works on three terminal regulators. A regulator is an amplifier and the advantages of Class A operation and simple topologies apply to regulators just like to audio amplifiers.
"Interestingly, non of these circuits are commercial, the SE Amplifiers that are actually being manufactured with minimal exceptions all operate with outputs that amounts to common cathode/emitter/source (shunt) mode. Moreover, so do most DIY Amp's."
And once one observes the distortion spectrum and the load interaction of a common anode/collector/drain (aka follower) mode circuit and actually listens to the resulting sound it becomes clear why."
Commercilal offering have the constrain of efficiency and power output ratings that are not concerns of many DIY builders. The Aleph amps are not strictly singal ended but have a dynamic current source for the increase in efficiency. Thre are some single ended follower amps on the commercial market. the Pathos amps come to mind. I and many others prefer the sound of follower output stages. Heavily biased Mosfet or BJT followers are some of the best sounding amps I have heard. No loop feedback makes them very easy to make stable. I think it is the topology of choice for amateur constructors for that reason. Compensating feedback amps with negative feedback can be a real challenge for the amateur designer and a pain for even experienced designers. I don't think the distortion argument is really that legitimate either as the non linearity of the transconductance is the dominant mechanism in either topologies. A common source output stage has even more interaction with the load since the load now controls the open loop voltage gain with serious variations in gain with frequency with most speaker loads. The open loop voltage gain of a follower circuit is much more independent of load. You won't get low output impedance with common source without a lot of negative loop feedback. There is no such thing as a free lunch.
http://www.pathosacoustics.com/indexeng.htm
I have nothing against shunt regulators since they have the potential for excellent PSSR. The factors influencing transient response noise, stability, and output impedance are the same as those for a series regulator. You can design a poor shunt regulator about as easy as a poor series regulator and the same basic topology can be applied to either with change in output device. You can even design a shunt regulator with a three terminal series regulator IC with not much trouble. I like simple topologies and single ended follower output has the advantage of less complexity and the lack of need for bias circuits and emitter degeneration. Simplicity makes it much easier to obtain stability for high bandwidth and gain feedback circuits like the Jung regulator. if more linearity is required it is quite easy to increase the preloaded DC current for shunt or series regulators. It even works on three terminal regulators. A regulator is an amplifier and the advantages of Class A operation and simple topologies apply to regulators just like to audio amplifiers.
Re: HOLD ON...
Works for me too. At least for RF filtering (above the 100kHz bandwidth of the transformer).
se
fdegrove said:
Works for me too. At least for RF filtering (above the 100kHz bandwidth of the transformer).
se
Sorry, I assumed you've tried it.
If you don't band-limit the input signals, then wider bandwidth elsewhere may well sound better, that wasn't my point.
If you bandwidth limit the input signals badly, then the degradation is audible.
Do it with the filter response above, and you may well find things sound even better.
Before you ask though, I'm not publishing that design - sorry (it's not mine).
Andy.
Re: more b********* from the psychic
You have a high-pass element in there somewhere? Just wondering why your group delay's increasing down there in the low end.
Oh, and seeing as you've got time for plots, how 'bout those FFT plots of battery noise you promised?
se
ALW said:
You have a high-pass element in there somewhere? Just wondering why your group delay's increasing down there in the low end.
Oh, and seeing as you've got time for plots, how 'bout those FFT plots of battery noise you promised?
se
Yep.
I can't find them at present, but haven't looked too hard to be honest - they're archived on a CD somewhere.
I was intending to re-do them in more detail though, and post them on my website for reference.
Not had much time of late though, sorry, but have every intention.
Andy.
ALW said:
Ok, thanks.
By the way, see the addendum to my post (which was done while you were replying, sorry).
se
Excellent discussion, gents!
ALW, your web site is fantastic. I hope to see more stuff from you on the site soon, so get busy 🙂
I love this place.
mlloyd1
who has obviously, once again, stayed at work far too many hours
ALW, your web site is fantastic. I hope to see more stuff from you on the site soon, so get busy 🙂
I love this place.
mlloyd1
who has obviously, once again, stayed at work far too many hours
...so get busy
I AM busy 😉
Thanks for the comments, the good stuff takes a lot of time, especially if fundamental research is required.
There will be more, but it will take time.
It'll be worth it though, I promise 🙂
A.
I AM busy 😉
Thanks for the comments, the good stuff takes a lot of time, especially if fundamental research is required.
There will be more, but it will take time.
It'll be worth it though, I promise 🙂
A.
Re: more b********* from the psychic
Koinichiwa,
Well, according to your test gear you have 50 degree phaselag at 20KHz. To me that is much more than your claim quoted below would suggest:
"The input signal is then limited by a very high quality low-pass filter within the preamp, that limits bandwidth to that which is audible. It does this without introducing any musically important anomalies (e.g. phase etc)."
BTW, the HF filter curves I noted on older Naim gear where more like -1.5..-2db @ 20KHz and the phaseshift was huge (out of neccesity). Which Naim pre is your measurements from?
The current series preamplfiers have a somewhat wider bandwidth than the older. The Amplifiers also have (subjectively) improved. I should have perhaps made clear that my comments applied to the pre "5 Series" equipment. The changes in the 5 series are not large but seem to have been suficient to remove the "classic Naim" sound.
Sayonara
PS, I was perfectly aware that JV designed the older Naim gear, primarily using circuits BTW out of the "recommended applications" section of Transistor manufacturers books. What he did then with these circuit was somewhat different indeed.
Koinichiwa,
ALW said:
Well, according to your test gear you have 50 degree phaselag at 20KHz. To me that is much more than your claim quoted below would suggest:
"The input signal is then limited by a very high quality low-pass filter within the preamp, that limits bandwidth to that which is audible. It does this without introducing any musically important anomalies (e.g. phase etc)."
BTW, the HF filter curves I noted on older Naim gear where more like -1.5..-2db @ 20KHz and the phaseshift was huge (out of neccesity). Which Naim pre is your measurements from?
The current series preamplfiers have a somewhat wider bandwidth than the older. The Amplifiers also have (subjectively) improved. I should have perhaps made clear that my comments applied to the pre "5 Series" equipment. The changes in the 5 series are not large but seem to have been suficient to remove the "classic Naim" sound.
Sayonara
PS, I was perfectly aware that JV designed the older Naim gear, primarily using circuits BTW out of the "recommended applications" section of Transistor manufacturers books. What he did then with these circuit was somewhat different indeed.
Re: Re: more b********* from the psychic
Yes, he did overstate on the issue of phase shift. But again, phase shift in itself is irrelevant. If you'll look at the group delay plot, you'll see that it's flat out to 1 MHz (excepting the low frequency group delay due to the high pass element). Meaning that all frequencies are arriving when they're supposed to and there's no alteration of the waveform.
se
Kuei Yang Wang said:
Yes, he did overstate on the issue of phase shift. But again, phase shift in itself is irrelevant. If you'll look at the group delay plot, you'll see that it's flat out to 1 MHz (excepting the low frequency group delay due to the high pass element). Meaning that all frequencies are arriving when they're supposed to and there's no alteration of the waveform.
se
- Status
- Not open for further replies.