Guys, thanks for the suggestions. Bonsai, I can't see there's anything here to challenge Mr Self with anyway - he's done the measurements that show there is no problem technically.
I know as much about feedback from control theory at uni as any other EE that hasn't used it since (ie not a huge amount, but I remember finding the transform maths tedious), but I agree that technically there should be no problem with the way it's used in audio amplifiers per se. The trouble is, I'm fairly sure what I'm hearing is not something that will be picked up using conventional measurements so it seems pointless to insist that feedback can't be the culprit. Certainly no different than speculating that it's due to opamp class B operation when we already have the measurements that suggest it isn't. I'm describing something that I hear. I don't expect to be able to offer some new previously unknown technical phenomenom by way of explaning what I'm hearing - the best I can hope for is offer my opinion that as is, the design sounds a certian way, but by changing this or that it sounds better/different to me but I don't have any rational explanation why. I don't think this will be at all useful to anyone that isn't also building the design and agrees with my opinions on SQ
I think the first thing I'll try is increasing the value of the 680R at the inverting input of the final gain stage in the active gain control - the buffer fed from the pot wiper is driving this and since the other end is virtual gnd this is a pretty taxing load on the buffer, even though the signal voltage level is lower here. I'll try 1k - this should be an easier load that is less likely to cause any class B operation while not increasing noise significantly. Then I might try the class A bias trick but I'm not keen on this idea to be honest. Might be better to then try just bypassing the whole AG stage as previously discussed. I'll try and let you know how it goes by the end of week if I get a chance.
I know as much about feedback from control theory at uni as any other EE that hasn't used it since (ie not a huge amount, but I remember finding the transform maths tedious), but I agree that technically there should be no problem with the way it's used in audio amplifiers per se. The trouble is, I'm fairly sure what I'm hearing is not something that will be picked up using conventional measurements so it seems pointless to insist that feedback can't be the culprit. Certainly no different than speculating that it's due to opamp class B operation when we already have the measurements that suggest it isn't. I'm describing something that I hear. I don't expect to be able to offer some new previously unknown technical phenomenom by way of explaning what I'm hearing - the best I can hope for is offer my opinion that as is, the design sounds a certian way, but by changing this or that it sounds better/different to me but I don't have any rational explanation why. I don't think this will be at all useful to anyone that isn't also building the design and agrees with my opinions on SQ
I think the first thing I'll try is increasing the value of the 680R at the inverting input of the final gain stage in the active gain control - the buffer fed from the pot wiper is driving this and since the other end is virtual gnd this is a pretty taxing load on the buffer, even though the signal voltage level is lower here. I'll try 1k - this should be an easier load that is less likely to cause any class B operation while not increasing noise significantly. Then I might try the class A bias trick but I'm not keen on this idea to be honest. Might be better to then try just bypassing the whole AG stage as previously discussed. I'll try and let you know how it goes by the end of week if I get a chance.
Unfortunately I don't have access to any test gear with the required resolution at the moment. I checked waveforms with a CRO for stability and level, then ran a frequency response test using a DSP speaker measurement system that I've tricked and calibrated for this purpose. All looks ok so I would assume it would also measure well wrt THD and noise.
That may be a good article, however my hobby budget is in big trouble with SWMBO now, so it seems a bit pricey - I agree, is it available elsewhere?
I'm familiar with that name - he's the guy behind the Hypex class D modules isn't he? A company I used to work for insisted on using them in active speakers - I am not at all encouraged by the way they sounded...
Despite this, and despite the fact that you have never demonstrated that you can distinguish this preamp from any other in a blind test, you are unconcerned about any possible financial damage you may be doing to another member?
Could you please clarify the last part of that statement? Are you suggesting that expressing my opinion on the sound quality of audio equipment will affect someone financially?
If you had been reading this thread earlier on you'd know that I've done some blind testing and had no trouble picking the preamp - I don't expect this to impress any hard line objectivist though.
If you had been reading this thread earlier on you'd know that I've done some blind testing and had no trouble picking the preamp - I don't expect this to impress any hard line objectivist though.
I'll take that as a yes, in which case thank you, I'm flattered. Unfortunately I think you're mistaking me for someone claiming to know all the amazing secrets of the art of audio or some incredible technical whiz and therfore having much influence on this forum. Not the case - I'm just some guy y'know. Can't see anyone being too bothered by my opinion either way
counter culture I have previously asked you to desist from this line of badgering. What you are in fact advocating is censorship, one more comment of this nature and you will spend the next week in the bin. I suggest strongly that your participation in this thread is at an end if you cannot adhere to moderation directives. FINAL WARNING 
It's online here:
http://www.linearaudio.net/images/onlinearticlesPDF/volume1bp.pdf
If you find the math intimidating, at least read the three 'storylines'.
jan
Last edited:
D Self Preamp
This reminds me of a similar argument regarding transformers I had many years ago with one of these HiFi gurus who could hear from DC to light.
He moaned about the output transformer in the final stage of a tube amplifier and said that he hates sound that must travel through an transformer.
My answer was that in a good old analog recording studio the signal passed through... now let's count the number of transformers.
1) The Mic has a balanced drive xfr
2) The input to the console has an input xfr
3) The output of the 24 track console has 24 output xfrs.
4) The 24 input Studer has 24 input xfrs
5) Hell the tape heads are essentially xfrs.
Now that the signal was on 24 track tape it had to be mixed down to 2 track. So here we go again.
1) The 24 track Studer was played back and we have the playback heads
2) The signal passed through the 24 output xfrs into the console
3) The console has 24 input xfrs.
The mix down is done and converted to two channel
1) Now the signal must pass from the two channel format through the console and out to the 2 track Studer through two more output xfrs....Yikes!
2) Now the signal is passed through the recording heads (more xfrs) and we finally had a 2 track master tape.
Now had to make the vinyl "mother"
1) So off again the Studer is in playback and signal comes off the heads = xfr.
2) Passes out of Studer through the output xfrs to the cutting amplifier.
3) Cutting amplifier has input xfrs.
4) Cutting heads are big *** coils.
Total count of signal passing through copper + iron = 14 and this does not take into account the signal ever passing from balanced xfr line outs to other gear (Compressors, de-essers etc all with more iron)
So what does one more output transformer mean? if it is of good quality - nothing.
Same issue with the opamp story. The signal in the recording chain goes through a "mudbath" of ICs, bad capacitors, bad cable and the gurus moan and groan about a few opamps in a pre-amplifier.
I admit my preference is tubes and in my sold state designs I prefer discrete.
The NE5532/NE5534 are great parts vs their cost.
A NE5532 costs me $0.18 and we use tens of thousands of them.
The other low cost IC I like is the NJM2068M
Steve Mantz
Zed Audio Corp.
This reminds me of a similar argument regarding transformers I had many years ago with one of these HiFi gurus who could hear from DC to light.
He moaned about the output transformer in the final stage of a tube amplifier and said that he hates sound that must travel through an transformer.
My answer was that in a good old analog recording studio the signal passed through... now let's count the number of transformers.
1) The Mic has a balanced drive xfr
2) The input to the console has an input xfr
3) The output of the 24 track console has 24 output xfrs.
4) The 24 input Studer has 24 input xfrs
5) Hell the tape heads are essentially xfrs.
Now that the signal was on 24 track tape it had to be mixed down to 2 track. So here we go again.
1) The 24 track Studer was played back and we have the playback heads
2) The signal passed through the 24 output xfrs into the console
3) The console has 24 input xfrs.
The mix down is done and converted to two channel
1) Now the signal must pass from the two channel format through the console and out to the 2 track Studer through two more output xfrs....Yikes!
2) Now the signal is passed through the recording heads (more xfrs) and we finally had a 2 track master tape.
Now had to make the vinyl "mother"
1) So off again the Studer is in playback and signal comes off the heads = xfr.
2) Passes out of Studer through the output xfrs to the cutting amplifier.
3) Cutting amplifier has input xfrs.
4) Cutting heads are big *** coils.
Total count of signal passing through copper + iron = 14 and this does not take into account the signal ever passing from balanced xfr line outs to other gear (Compressors, de-essers etc all with more iron)
So what does one more output transformer mean? if it is of good quality - nothing.
Same issue with the opamp story. The signal in the recording chain goes through a "mudbath" of ICs, bad capacitors, bad cable and the gurus moan and groan about a few opamps in a pre-amplifier.
I admit my preference is tubes and in my sold state designs I prefer discrete.
The NE5532/NE5534 are great parts vs their cost.
A NE5532 costs me $0.18 and we use tens of thousands of them.
The other low cost IC I like is the NJM2068M
Steve Mantz
Zed Audio Corp.
Hi Steve,
I know this argument well. I've seen schematics for a few recording and broadcast consoles and they were full of 5532s, TL072s and cheap electros. (I have heard that some of the most highly regarded ones sonically (eg Neve) contain more expensive components though - don't know if that's true). It's hard to dispute if you believe 100% that THD, FR and noise measurements completely define whether equipment could cause sonic degradation. However if you say the same thing but with the assumption that every part of the recording chain IS degrading the signal, the statement then becomes: "the signal has already been degraded so much during the recording process that the equipment used for reproduction cannot matter" - which is perhaps not such a rational argument?
I mention this because I think so much of my record collection is badly recorded. For example, how many pianos have you heard in real life from a normal distance that sound like you have your head inside the lid and the bass strings are 3 metres away from the treble strings? Yet for some reason most pianos seem to get recorded that way. I could go on...
I also can't understand why the preamp passing line level signals with essentially no or very little gain has such a big effect on the sound of the system as a whole, but it does seem to - possibly equally or more so than the power amp. Rationalise that!
I agree the the 5532 is a great opamp - even cost aside. I don't think it's any coincidence that it sounds good while also producing the lowest distortion and noise in these sort of applications amoung the cheap opamps. But interestingly recently I have found that different brands of 5532 seem to sound different. Happily I don't think anyone commenting here has poo-pooed them either other than suggesting trying some different types to see if the sound is different. I certainly wouldn't have embarked on building a preamp littered with them if I thought they couldn't sound good. I think they must be essentially harmless sonically if used in certain ways, hence why I think what I'm hearing with this design comes down to the particular circuit configuration. But I'm talking about small differences that would not bother someone without a passion for hifi - ie audio rationalists (is that a sensible term to use?)
Cheers,
Owdeo
I know this argument well. I've seen schematics for a few recording and broadcast consoles and they were full of 5532s, TL072s and cheap electros. (I have heard that some of the most highly regarded ones sonically (eg Neve) contain more expensive components though - don't know if that's true). It's hard to dispute if you believe 100% that THD, FR and noise measurements completely define whether equipment could cause sonic degradation. However if you say the same thing but with the assumption that every part of the recording chain IS degrading the signal, the statement then becomes: "the signal has already been degraded so much during the recording process that the equipment used for reproduction cannot matter" - which is perhaps not such a rational argument?
I mention this because I think so much of my record collection is badly recorded. For example, how many pianos have you heard in real life from a normal distance that sound like you have your head inside the lid and the bass strings are 3 metres away from the treble strings? Yet for some reason most pianos seem to get recorded that way. I could go on...
I also can't understand why the preamp passing line level signals with essentially no or very little gain has such a big effect on the sound of the system as a whole, but it does seem to - possibly equally or more so than the power amp. Rationalise that!
I agree the the 5532 is a great opamp - even cost aside. I don't think it's any coincidence that it sounds good while also producing the lowest distortion and noise in these sort of applications amoung the cheap opamps. But interestingly recently I have found that different brands of 5532 seem to sound different. Happily I don't think anyone commenting here has poo-pooed them either other than suggesting trying some different types to see if the sound is different. I certainly wouldn't have embarked on building a preamp littered with them if I thought they couldn't sound good. I think they must be essentially harmless sonically if used in certain ways, hence why I think what I'm hearing with this design comes down to the particular circuit configuration. But I'm talking about small differences that would not bother someone without a passion for hifi - ie audio rationalists (is that a sensible term to use?)
Cheers,
Owdeo
I don’t think so; I see every step/device in the process/chain of reproducing sound as a musical instrument in itself. When looked at it in this way, the output of each stage is an 'original' reproduction. So, whatever the influence is of the next stage, the components/process of the next stage will influence the reproduction of its input (previous stage) and this may be measurable/hearable/detectable in any/many way(s), by human or machine
The result may be pleasing or dis-pleasing, but there will be a result, and most of the time it will not be neutral/undetectable.Thanks Jan, very interesting and is going to take a few reads to absorb fully.
I note the following paragraph:
"I surmise that, since the amplifier’s distortion was never negligible, mak-ing it constant across the audio band makes it fly under the psychoacoustic radar more easily. My
own subjective experience would support this. To my ears, amplifiers with the normal 20dB/decade
behaviour but whose distortion is not negligible at the end of the audio range have glassy mid-highs, a “superglue stereo image” as KK once put it and the illusion of spectacularly, unnaturally
tight and impossibly controlled bass. Some love this, and seceded into a subculture of ultra-beefy
amplifiers. I don’t and when forced to make a choice I’ll take higher but consistent distortion across
the band."
What's this - a reference to how we perceive distortion and the author's speculation based on his own unscientific subjective experience? Shock horror! Naughty naughty, what will the hardcore objectivists say?
I agree with the description but my own subjective experience has been that all the "Lin" configuration amps I've heard sound this way. An obvious disprover of his theory re rising THD causing it is the Leach amp, which has rapidly rising THD at HF from low levels at LF and yet exhibits none of the "glassiness" etc.
Speaking of Leach I think some of what is covered regarding TIM and GBW etc is covered very succintly and more simply in the Leach amp FAQ section where he explains what "low feedback" actually means and where various pundits are telling him he should have used various other techniques in his amp to give more gain or BW and he explains why this would not make any difference when the GBW product is fixed by design.
Sooooo, getting back to the preamp, perhaps the effect I'm hearing is caused by not enough feedback around the active gain control
Thank goodness we all now understand that too much feedback is barely enough!
I don’t think so; I see every step/device in the process/chain of reproducing sound as a musical instrument in itself. When looked at it in this way, the output of each stage is an 'original' reproduction. So, whatever the influence is of the next stage, the components/process of the next stage will influence the reproduction of its input (previous stage) and this may be measurable/hearable/detectable in any/many way(s), by human or machine
That makes sense for multitracked pop or other artificial recordings, but I think recording an acoustic group of whatever type should be about capturing the sound of the original event, not modifying it. Same goes with reproduction equipment - I'm not looking for just a pleasant sound, I want a system that sounds as close to the real thing as possible.
always love the "reproduce a real performance" claims
which mics, where? - only binaural dummy heads? crossed cardioid, ambisonics? - what seating position?
virtually all commercial recordings are close miced, most live groups in venues of any size are amplified today - have individual instrument pickup, mic feeds going to a mixing board
studio music production can have performers in different soundproof rooms
many live acoustic performance recording use many mics, some in "listening position" others often flown to capture "ambiance" - mic selection, placement are "artistic choices" - modify recorded sound, capture sound from positions no listener is able to get their ears to
which mics, where? - only binaural dummy heads? crossed cardioid, ambisonics? - what seating position?
virtually all commercial recordings are close miced, most live groups in venues of any size are amplified today - have individual instrument pickup, mic feeds going to a mixing board
studio music production can have performers in different soundproof rooms
many live acoustic performance recording use many mics, some in "listening position" others often flown to capture "ambiance" - mic selection, placement are "artistic choices" - modify recorded sound, capture sound from positions no listener is able to get their ears to
Last edited:
Curious then why you'd build an NE5532-based pre which is also an opamp of (broadly) the same topology. Here's an experiment to try - swap all those 5532s out for LM6172s and listen again.
jcx,
Me thinks you are trying to preach to the choir. You are coming from the position of a pro-audio person rather than the thinking that you can reproduce any live performance and get it to sound exactly like the original. Even being in a live situation if you are sitting in a different position than someone else you are hearing a different blending of the sounds. Coming from the live sound field I can agree with you that it is just unrealistic to have any recording that sounds exactly as the live event. Even to say that a piano should sound a certain way when reproduced goes against the fact that every piano sounds different. Even the same piano played in a different room will sound different and then there is the fact that one person tuning the piano may tune it slightly different than another. All we can really do is make our systems sound as good as we as individuals like. I may like a certain sound and you are looking for something different. What one person says is a sparkly sheen in the upper frequencies another might find annoyingly bright. That is the same reason that earlier Owdeo could not put into words what it was that he thought wasn't right with the preamp in the first place. It isn't something that can be quantified. Or even measured. These arguments can go on for years with no resolution but a lot of opinions about what the truth is.
Me thinks you are trying to preach to the choir. You are coming from the position of a pro-audio person rather than the thinking that you can reproduce any live performance and get it to sound exactly like the original. Even being in a live situation if you are sitting in a different position than someone else you are hearing a different blending of the sounds. Coming from the live sound field I can agree with you that it is just unrealistic to have any recording that sounds exactly as the live event. Even to say that a piano should sound a certain way when reproduced goes against the fact that every piano sounds different. Even the same piano played in a different room will sound different and then there is the fact that one person tuning the piano may tune it slightly different than another. All we can really do is make our systems sound as good as we as individuals like. I may like a certain sound and you are looking for something different. What one person says is a sparkly sheen in the upper frequencies another might find annoyingly bright. That is the same reason that earlier Owdeo could not put into words what it was that he thought wasn't right with the preamp in the first place. It isn't something that can be quantified. Or even measured. These arguments can go on for years with no resolution but a lot of opinions about what the truth is.
Good question. I've seen the internal schematic of the 5534 and agree it's broadly similar topology, though it does have nested feedback and a few other things that might make it different. But actually I've only found this "sonic signature" to be the case with power amps of this topology. One of the nicest sounding preamps I've found so far was in fact using something similar to the discrete class A opamp published by none other than Douglas Self in his excellent Small Signal Audio Design book. So I guess I like the sound of this topology at line levels but not as a power amp. Wierd I know.
I just had a look at the LM6172 datasheet - 3MV/us slew rate
Doesn't look like one that's likely to be stable as a drop in replacement. Have you had any practical experience with it? What's different about it - is the topology different?