I don't know a single engineer that does not rely on science in order to make a product. I also do not know a single product that does not have that engineer's fundamental sound attribute in the sound. In other words, as transparent a device may be, it still has that X designer's sound to it...as little or a lot it may have, it's there. Same with speakers. You can do anything to make a speaker sound a certain way, but in the end, you can still hear that underlying/fundamental sound that driver makes.
Measurements are always very important, but "reproduction" of anything, will always be just that...something that is in the world of subjectivity. No two ears are the same, even if ears can be so good enough to tell what is what or they can give the basics of what is heard (even if 99.9% say it sounds otherwise)...but it doesn't mean what these golden ears say is right/wrong/the truth/etc.
Audio, as with any subjective aspect of life, will ALWAYS and indefinitely be just that, a subjective experience.
Measurements are always very important, but "reproduction" of anything, will always be just that...something that is in the world of subjectivity. No two ears are the same, even if ears can be so good enough to tell what is what or they can give the basics of what is heard (even if 99.9% say it sounds otherwise)...but it doesn't mean what these golden ears say is right/wrong/the truth/etc.
Audio, as with any subjective aspect of life, will ALWAYS and indefinitely be just that, a subjective experience.
They are not...
Being CFB op amps yield a high slew rate , wich surely did impress you much,
but looking through the details ; you ll see that the AD811 has hardly better
linearity in the audio band than the NE553X , while the LM6181 probably
is much worse in this respect , NS doesnt even bother to publish
THD vs FREQUENCY graphs.
As for the more common VFB op amps , like the 627 , just look at their
GBW products graphs, compare it to the one of the 553X and you can
see all the "progress" that has been made since the late 70s....
Hi,
Maybe they met HIS needs.
But based on the results some of his customers got and their reactions to the results, it failed to meet their needs...
Which is why they paid people a lot of money to rebuild the Soundcraft Recording Consoles, including fixing the grounding wiring problems (appallingly bad and equally an appallingly basic thing to get wrong) and changing to higher performance Op-Amp's among other measures...
Ciao T
Maybe they met HIS needs.
But based on the results some of his customers got and their reactions to the results, it failed to meet their needs...
Which is why they paid people a lot of money to rebuild the Soundcraft Recording Consoles, including fixing the grounding wiring problems (appallingly bad and equally an appallingly basic thing to get wrong) and changing to higher performance Op-Amp's among other measures...
Ciao T
When I was building my speaker's xover I first used resistors and caps that were of typical quality and did FR graphs to get it right.
After I was satisfied I put in Mills resistors and a couple of teflon caps and I had used in other projects, and measured the results.
No difference in FR.
But the music sounded cleaner and clearer.
My simple FR mic (Dayton Audio EMM-6 Electret Measurement Microphone) and software couldn't capture the difference.
http://www.diyaudio.com/forums/multi-way/167960-inexpensive-gedlee-bouska-waveguide-tweeter-midbass-idea-5.html
I'll always use FR in designing xovers, and add in "audiophile" parts in later, as they won't change the engineering part of the design.
After I was satisfied I put in Mills resistors and a couple of teflon caps and I had used in other projects, and measured the results.
No difference in FR.
But the music sounded cleaner and clearer.
My simple FR mic (Dayton Audio EMM-6 Electret Measurement Microphone) and software couldn't capture the difference.
http://www.diyaudio.com/forums/multi-way/167960-inexpensive-gedlee-bouska-waveguide-tweeter-midbass-idea-5.html
I'll always use FR in designing xovers, and add in "audiophile" parts in later, as they won't change the engineering part of the design.
Hi,
They are, please measure them...
I have no idea what you are talking about.
It is the Philips (and on-semi and others) datasheet that shows no distortion information whatsoever. While LM6181 datasheet on the TI site does not show the HD graphs, the datasheet for the dual part does though.
More critical perhaps is that the AD811 and LM6181 (et al) are characterised for distortion at 100 or 150 Ohm load. At 600 Ohm load or 2K load they perform much better and the performance of the 553X at 100 Ohm load is one they never bothered to publish (wisely so).
The OPA627, as well as AD811 and LM6181 are also quite aged parts (if not already discontinued) and not recent.
And why focus on GBW, why not (for example) distortion?
Oh I get it.
We cherry pick specifications that are either not given or not directly comparable or fail to show advantages and then make a big song and dance about it claiming "look, what he said is wrong".
You may find it more useful to actually measure said devices...
Ciao T
They are, please measure them...
I have no idea what you are talking about.
It is the Philips (and on-semi and others) datasheet that shows no distortion information whatsoever. While LM6181 datasheet on the TI site does not show the HD graphs, the datasheet for the dual part does though.
More critical perhaps is that the AD811 and LM6181 (et al) are characterised for distortion at 100 or 150 Ohm load. At 600 Ohm load or 2K load they perform much better and the performance of the 553X at 100 Ohm load is one they never bothered to publish (wisely so).
As for the more common VFB op amps , like the 627 , just look at their GBW products graphs, compare it to the one of the 553X and you can see all the "progress" that has been made since the late 70s....
The OPA627, as well as AD811 and LM6181 are also quite aged parts (if not already discontinued) and not recent.
And why focus on GBW, why not (for example) distortion?
Oh I get it.
We cherry pick specifications that are either not given or not directly comparable or fail to show advantages and then make a big song and dance about it claiming "look, what he said is wrong".
You may find it more useful to actually measure said devices...
Ciao T
They both have higher THD in the audio band than the old dude....
At such low loads , surely that such high current op amps
probaly do better......
You are quite paradoxal since GBW product will inherently be a
parameters of first importance for perfs once we apply NFB,
do we , to those circuits , in the aim to reduce....THD !!..
At 1 khz , the NE5532 OL gain is still 100 dB while the OPA627
has collapsed to 85 dB and the OPA134 at 80db.
Now , at 20KHz , the numbers are 65 dB , 59 dB and 53 dB respectively.
Curiously , when restricted to audio frequencies , the NE5532 has higher
GBW product than those two well regarded op amps , it s only below 100 HZ
that theses "audiophile" products have higher gain and thus higher GBW product ,
exactly where it is not needed at all since at those frequencies
there was already enough gain and thus NFB available...
Basically , those two more recent op amps have higher DC gain
but lower AC gain and their linearity (read low THD) is due to better
manufacturing processes than the one available at the time of
the NE553X wich has notoriously slow lateral PNP transistors,
yet , it s still unsurpassed to these days for audio purposes, and when it is , it s only marginally..
"only marginally". Of course. As one gets closer and closer to perfection, in any endeavor, the margins get smaller and smaller for the effort/cost involved. That's the old 90/10 rule. The first 90% takes the first 90% of time/cost whatever, the last 10% takes the other 90%.
I do remind some of the folks here that the basis for this thread is there is some effect that is not currently understood that causes differences in the audible perception of an amplifier. Something is effecting the dynamic transfer function. This is not something we have today on the data sheets. If it was, John, Nelson, and many others would have eliminated it long ago. (they may actually have, just on a price range I can't play in) What ever it is, it is very small and not everyone can either hear it, or if they hear it, be bothered by it. I would ask that some of the lucky majority who can not hear the differences between decent amps would not be so forceful in claiming because they can't hear or measure a difference, there is not one.
I sadly agree high end is a fading market. It seems portability and ease outweigh quality in not just audio, but in most consumer products. In audio, most of the current generation is half deaf from iPods, car stereos, or metro background noise. Unfortunately, most who have served on active duty do not have all their hearing. They see no reason to waste money on subtle differences they can't hear, and rightly so. They will need the cost of a decent amp for their first hearing aids.
I do remind some of the folks here that the basis for this thread is there is some effect that is not currently understood that causes differences in the audible perception of an amplifier. Something is effecting the dynamic transfer function. This is not something we have today on the data sheets. If it was, John, Nelson, and many others would have eliminated it long ago. (they may actually have, just on a price range I can't play in) What ever it is, it is very small and not everyone can either hear it, or if they hear it, be bothered by it. I would ask that some of the lucky majority who can not hear the differences between decent amps would not be so forceful in claiming because they can't hear or measure a difference, there is not one.
I sadly agree high end is a fading market. It seems portability and ease outweigh quality in not just audio, but in most consumer products. In audio, most of the current generation is half deaf from iPods, car stereos, or metro background noise. Unfortunately, most who have served on active duty do not have all their hearing. They see no reason to waste money on subtle differences they can't hear, and rightly so. They will need the cost of a decent amp for their first hearing aids.
No, not really. The effects which cause amplifiers to be audible are all well-understood and well-documented. The questions then become commercial:
1. Do I want to market an amplifier whose output is audibly indistinguishable from its input?
2. If not, which colorations are desired in my target market?
3. If so, how do I differentiate my product from all the other products which are competently designed to that same goal?
Note that I'm talking about boxes of gain. Transducers are far more complicated.
We will have to agree to disagree SY. You criteria are quite valid. But, some amps succeed in whatever it is that causes what we are hearing. Others do not. There is an audible effect that although not inherent as some amps do not have it, but not understood as others do, and in a price range where if the issue was understood, it would be addressed.
As the mail was slow and I did not get the parts I ordered, I can't start again on this on my day off tomorrow. I may just have to sit back and listen to music through the Rotel's I have that do not seem to exhibit the effect and enjoy a few well crafted products from the Anchor Brewing Company of SF. If there is any product that the good old US of A can make head to head with anyone, it is Porter!
BTW, SY, are you familiar with the A/B box test David Hafler did back in the 80's. Very convincing about no difference with in vs out, but at a steady state. I may try this basic setup but with dynamic inputs to see if I can catch anything. My contrived pulse test may make it repeatable enough to get a picture of it. The DH 120 and 220 are amps he built that passed his test, (or was it a test he designed his amps passed) but my wife informs me do not pass hers. I will then repeat the test with my Rotel. Again, my suspicion, this needs to be done with a real speaker load not a perfect load.
One possible reason for the controversy is that different loads may aggravate whatever the issue is more than others.
As an example, very small amounts of high order distortion can play hell with metal cone drivers if the crossover is not well behaved, yet an easier to use cloth dome may not exhibit the issue. So, is this a fault of the driver or amp? I contend it is "yes" as it is a system problem. The only problem I have with his setup is propagation delay. That may mask the problem. It could also be the problem is not detectable with my scope or analyzer. All this is just words until I can get some pictures.
Wonder if I could get away with snagging an HCA 800-II off the web without the "boss" noticing. I don't really have any excuse for another amp other than testing.
As the mail was slow and I did not get the parts I ordered, I can't start again on this on my day off tomorrow. I may just have to sit back and listen to music through the Rotel's I have that do not seem to exhibit the effect and enjoy a few well crafted products from the Anchor Brewing Company of SF. If there is any product that the good old US of A can make head to head with anyone, it is Porter!
BTW, SY, are you familiar with the A/B box test David Hafler did back in the 80's. Very convincing about no difference with in vs out, but at a steady state. I may try this basic setup but with dynamic inputs to see if I can catch anything. My contrived pulse test may make it repeatable enough to get a picture of it. The DH 120 and 220 are amps he built that passed his test, (or was it a test he designed his amps passed) but my wife informs me do not pass hers. I will then repeat the test with my Rotel. Again, my suspicion, this needs to be done with a real speaker load not a perfect load.
One possible reason for the controversy is that different loads may aggravate whatever the issue is more than others.
As an example, very small amounts of high order distortion can play hell with metal cone drivers if the crossover is not well behaved, yet an easier to use cloth dome may not exhibit the issue. So, is this a fault of the driver or amp? I contend it is "yes" as it is a system problem. The only problem I have with his setup is propagation delay. That may mask the problem. It could also be the problem is not detectable with my scope or analyzer. All this is just words until I can get some pictures.
Wonder if I could get away with snagging an HCA 800-II off the web without the "boss" noticing. I don't really have any excuse for another amp other than testing.
Yes, I know the Hafler box; it was designed to promote the XL280, if memory serves. There's far better ways of doing the same thing now- for example, the brilliant DiffMaker software. For ears-only listening comparisons, one can easily rig up a high quality switch box (see Linear Audio Volume 2 for a discussion of this).
There are a lot of amplifiers designed to be effects boxes, and there you get into a whole different set of questions. But the reasons that they're effects boxes and the nature of the effects are straightforward engineering.
There are a lot of amplifiers designed to be effects boxes, and there you get into a whole different set of questions. But the reasons that they're effects boxes and the nature of the effects are straightforward engineering.
Hi,
Please show your measurements... I find they have substantially less under identical conditions with 1K loads...
If a circuit has less THD than another (which the OPA627 has compared to the NE553X) then the implication is that it is more linear. If it is that while having less OLG than the other circuit, it simply means it is more linear.
It is not just marginally surpassed, but quite dramatically, once we get real circuits. But please continue to use if you like.
I use 553X in some specific applications (though not the direct signal path) where the relative decent specs and very low price make it an excellent choice. Heck, i have use for the TL081 at that in audio gear (but not to amplify audio)...
Ciao T
Please show your measurements... I find they have substantially less under identical conditions with 1K loads...
If a circuit has less THD than another (which the OPA627 has compared to the NE553X) then the implication is that it is more linear. If it is that while having less OLG than the other circuit, it simply means it is more linear.
It is not just marginally surpassed, but quite dramatically, once we get real circuits. But please continue to use if you like.
I use 553X in some specific applications (though not the direct signal path) where the relative decent specs and very low price make it an excellent choice. Heck, i have use for the TL081 at that in audio gear (but not to amplify audio)...
Ciao T
I find it very hard to accept that he never ever listened to his own designs.
I can accept that during a production run of amps that he would not listen to each one - simply verify that each met the expected set of measurements. No problem with that. Once the design is set in stone there is no point in individual listening tests of individual amps of the same design as long as they measure the same, in fact you're more likely to hear differences that aren't actually there in such testing.
I could even accept that during the design process he didn't listen to them, but was guided entirely by measurements and following specific design principles. This is a bit more of a stretch, but still believable.
But never listened to an amplifier he designed ? I don't believe that for a second. He never used one of his own designs on his own personal home system at any time ? Sounds a bit like actors who claim to have never watched one of their own movies...
Peter Walker was famous for saying that all amps sound the same unless they are broken ie 'wire with gain' (he may have even coined that phrase).
He didn't have any truck with fancy cables either.
However famously at one US audio exhibition one 'audiophile' came up to him and said that while he owned the exact same all-QUAD system as they demonstrated it never ever sounded as good. The only difference where the strikingly orange cables QUAD used at the show so he insisted on purchasing them.
Turns out that when the QUAD people packed the gear they completely forgot speaker cables but found some old Black&Decker mains cable in the skip which is what they used at the exhibition.
Hi,
I serviced the whole SS Quad range while working at Music & Video Exchange (London) extensively, next to them Naim sounded "solid gold" and our cheap cheap Technics Class AA integrated "shop amp" walked so all over them, it was unfunny.
There was worse out there sonically, though, still these sounded such on our JBL Studio Monitors (the Shop Speakers) that I preferred running them into dummy loads for post repair testing, not speakers with msuic as we usually did.
I later worked on restoring a pair of Ex BBC Quad II's. These where not all that bad, quite nice little tube amp's, but they really did not offer anything special, power was not really better than a 300B SE Amp subjectively. Worked okay with Tannoy Reds.
Based on my experiences I find it hard to doubt that PW never listened to his company's amplifiers, neither would I...
Ciao T
I serviced the whole SS Quad range while working at Music & Video Exchange (London) extensively, next to them Naim sounded "solid gold" and our cheap cheap Technics Class AA integrated "shop amp" walked so all over them, it was unfunny.
There was worse out there sonically, though, still these sounded such on our JBL Studio Monitors (the Shop Speakers) that I preferred running them into dummy loads for post repair testing, not speakers with msuic as we usually did.
I later worked on restoring a pair of Ex BBC Quad II's. These where not all that bad, quite nice little tube amp's, but they really did not offer anything special, power was not really better than a 300B SE Amp subjectively. Worked okay with Tannoy Reds.
Based on my experiences I find it hard to doubt that PW never listened to his company's amplifiers, neither would I...
Ciao T
THD of both devices is about the same once we get on a real circuit...
As for linearity , well , the OPA627 use much more internal
local feedback than the 5532 , wich explain an apparently (only)
better linearity in open loop , but in fact it has worse intrinsical
linearity...
Attachments
I.e. when you design bread-carrier, you prefer to build a tank from scratch, instead of building something more like a truck with appropriate shelves for bread?
- Status
- Not open for further replies.