-190dB if I do my calcs correctly is about 0.3ppb. I just dont know how you can measure something that small. I'd be interested to see the test set up that produced these results. I would imagine 0.3ppb is swamped in noise, unless some clever technique is used (maybe cryo chamber and averaging over many thousands of reads, or some Kalman filtering . . . ;-) ). Anyway, who am I to argue - better let some of the better qualified guys give their views on this. at 0.3ppb, I am certainly out of my depth.
(Little 6 legged plastic package devices are probably matched transistors pairs - lot of suppliers offer these, some with Hfe matching to 2% between devices; you can get current sources and mirrors also)
(Little 6 legged plastic package devices are probably matched transistors pairs - lot of suppliers offer these, some with Hfe matching to 2% between devices; you can get current sources and mirrors also)
-190dB!! Ouch - I think a -190dB noise floor is well beyond the capabilities of the best 24bit A/D converters, even when the processing gain is accounted for. Maybe they did employ some sort of coherent (i.e. in-phase) averaging?
L.
I remember the one about speaker cables. For many years there were people claiming that they could hear the difference between speaker cables and their were some that even claimed that if they moved their cables that the sound would change.
Then there was the skeptics that scoffed and laughed at the golden eared lunatics.
Anyway, someone bothered to setup a test circuit several years ago and the results showed clear variations in the harmonic distortion spectrum when the cables were moved or when they were left unstressed. There were even bigger differences between cables.
There are still people claiming that the only thing that matters with speaker cables is the DC resistance and that there is no audible difference between different types of cables if they have the same resistance.
Speaker cables are a different can of worms (though not unrelated). If you have actual data demonstrating audibility of well designed chip circuits versus well designed discrete circuits, let's see them.
What is interesting is I found out some time later that Tektronix did something very similar to me that predated what I had done.
Heard of the Frazier lens? IIRC the scientist that reviewed the idea for the Australian goverment said it was theoretically, let alone physically impossible to build a lens with the attributes he was claiming. Panavision has the patent on it because some idiot with a degree and a whole lot of letters after his name didn't know what he was looking at.
In life there are doers and skeptics. Which camp do you fall in?
With all the trolling I for a moment got deflected from the true purpose of this threadBack on target now...
IC opamps these days don't break a sweat at sub -120dB distortion. So your open loop stage exceeded this distortion by accident or by design? If by design do you consider such low distortion to be important to sound quality?
There have been lots of papers written on pyschoacoustics and distortion.
If you read some, then did some experiments with this sort of stuff you may learn something.
You wouldn't have any info or documentation about that 'someone' and what he did & found....?
jan didden
They're not mutually exclusive. Skepticism is a key factor in distinguishing progress from wheel-spinning.
I think it may have been in Wireless World (about 5 years ago give or take a few. My memory is not good at some things). I thought it would have been known about around this forum.
Speaker cables have resistance, inductance and capacitance. The latter two are bound to change a little when the cable is moved or flexed. The capacitance will be a bit non-linear, as the cable insulation will have been chosen for reasons other than its goodness as a dielectric. Putting together Setup A in a lab so differences between speaker cables can be measured should not be too big a problem. It almost certainly would require a highish driving impedance, in order to see any effect. RF pickup may complicate things.
Setup B for listening tests will almost certainly be different from Setup A. This immediately means that transferring results from one to the other involves assumptions and extrapolations.
As I have said before more than once, if copper cable is non-linear then much of modern electronics simply cannot work, but it does. Therefore copper cable is linear. I think the same argument can be applied to good op-amps in well-designed circuits.
Setup B for listening tests will almost certainly be different from Setup A. This immediately means that transferring results from one to the other involves assumptions and extrapolations.
As I have said before more than once, if copper cable is non-linear then much of modern electronics simply cannot work, but it does. Therefore copper cable is linear. I think the same argument can be applied to good op-amps in well-designed circuits.
The only ABX test regarding speaker cables I remember is the one Tom Nouisane did a few years ago - it's well described in his article 'Wired wisdom', which is worth reading. Needless to say, Nouisane found that - under controlled conditions - none could distinguish the expensive cables (990$) from the 18-cent-a-foot zip cord.
Never seen a (serious) test which resulted in the opposite conclusions.
L.
Never seen a (serious) test which resulted in the opposite conclusions.
L.
Copper cable is non-linear (Does not follow ohms law!) It is not that hard to measure, see Linear Audio Vol. 1. But the level of deviation in normal use is very small.
Try measuring and listening to zip cable vs 10pr telephone cable...pretty much any amp and any speakers and the alert listener will discern a describable difference.
Eric.
No, but such advances are only detectable by instrumentation, and they're certainly not going to come from somebody who expects to be able to hear the difference.
This was decided when blind testing years ago disclosed that many amplifiers were audibly indistinguishable once carefully matched for output volume. It's well enough established (and ignored) that nobody even bothers to do it now.
For a generalisation or a stereotype to stick it has to be built on consistencies. Nothing you have said so far convinces me that you don't conform.
This is hardly fair Craig. All I'm doing is pointing out the fact that this thread is propagandising a world view that all the evidence suggests is false-to-fact. The whole thread is predicated on the presupposition that there is a difference to be heard between opamps. Post enough threads like this and pretty soon you've got a coterie of the otherwise uninformed believing this crud.
Yes, It's my experience. You'll pardon me if I point out that in my experience the best way to 'minimise individual semiconductor and thermionic vacuum tube parasitic effects' is to use an IC.
You discrete opamp guys. You all think 'merda taurorum animas conturbit'.
w
I couldn't see anything there about copper non-linearity. OK, I accept that nothing is completely linear but copper is sufficiently linear that it can be regarded as linear for almost all practical purposes - certainly much more linear than any loudspeaker drive unit. Probably more linear than the speaker cabinet, and maybe more linear than the air in the cabinet.simon7000 said:
This reminds me of a story I heard about a Rupert Neve mixing desk.
One of Neve's clients said that his desk didn't sound right out of one of the channels. So some of Neve's engineers went sent to investigate. The engineers told Neve that they couldn't hear anything wrong with it. Neve told his engineers something along the lines that he trusted what his client was saying. Anway the engineers put a scope and sig gen on the channel and run a sweep through it. Turns out the channel did have a fault. It had a small bump in the frequency response at 70kHz but was otherwise flat. They replaced the channel and the client was happy.
This is a perfect example of some people can hear things that others cannot.
One of Neve's clients said that his desk didn't sound right out of one of the channels. So some of Neve's engineers went sent to investigate. The engineers told Neve that they couldn't hear anything wrong with it. Neve told his engineers something along the lines that he trusted what his client was saying. Anway the engineers put a scope and sig gen on the channel and run a sweep through it. Turns out the channel did have a fault. It had a small bump in the frequency response at 70kHz but was otherwise flat. They replaced the channel and the client was happy.
This is a perfect example of some people can hear things that others cannot.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.