The input op amps can be changed to AD797 with a mini to round adaptor. A small cap must be added from output to the comp. pin as well, for lowest distortion. The residual noise can drop to almost the limits of measurement on the meter, AND be as good as the AP that everyone loves.
Soon as you said AP, well! the defacto standard! Wether it deserves it or not. And if you can equal it, for less, well, why not?
It's a great dicussion though, two (almost-sorta) different groups, both gently accusing the other of being deluded.
Well, I clearly and easily hear differences in cable. Be that due to termination, or design, is..indeterminate, but is obviously in the human condition. No snake oil, but consumers DO buy them and why?
Because they get to DO SOMETHING. When they want to, is the key. Anal, twitchy AND pro-active. If they wern't doing that, they'd be cleaning something, or chewing their nails. So, the differences heard are realtively small under the umbrella of linear weighted measurement systems.
I've said it before, and I'll say it again, as folks seem to miss this all important fact. The ear is like a diode in how it works. It, to the very large part, only deals with positive waveforms, and once again, to the largest extent..the leading peak edge of a given transient, in whatever level or point it occurs. So, in essence, the ear is ONLY looking at the leading positive peak or edge of a given transient. Be it micro or macro. So..peak transients, harmonic transients and the time between those given peaks. We live on, oh..about 10% of the signal. The rest..she don't matter. (it does to some degree -as a transient carrier positioning system- but I'm fishing for grandiose statements here!!,OK?) If this wasn't true, you can plainly see, that horn speakers would be absolutely impossible for the ear to decipher, as they get EVERYTHING wrong, except that all important positive transient peak.
If you wanted to measure something,and correlate it back to what we hear, then record it in terms of very high frequency digital sampling, and then remove all the unimportant bits (via weighting in analysis) ..and compare that to what you hear.
You WILL get useful results, that are weighted to what the ear actually hears.
You will find that you will go from 0.00XX% distortion, to much, much higher levels. After all, you will go to measuring the points that are actually the most distorted, and weighing that against the original signal, for your distortion analysis.
Re-weight those measurements, and you'll get somewhere. 🙂
It's a great dicussion though, two (almost-sorta) different groups, both gently accusing the other of being deluded.
Well, I clearly and easily hear differences in cable. Be that due to termination, or design, is..indeterminate, but is obviously in the human condition. No snake oil, but consumers DO buy them and why?
Because they get to DO SOMETHING. When they want to, is the key. Anal, twitchy AND pro-active. If they wern't doing that, they'd be cleaning something, or chewing their nails. So, the differences heard are realtively small under the umbrella of linear weighted measurement systems.
I've said it before, and I'll say it again, as folks seem to miss this all important fact. The ear is like a diode in how it works. It, to the very large part, only deals with positive waveforms, and once again, to the largest extent..the leading peak edge of a given transient, in whatever level or point it occurs. So, in essence, the ear is ONLY looking at the leading positive peak or edge of a given transient. Be it micro or macro. So..peak transients, harmonic transients and the time between those given peaks. We live on, oh..about 10% of the signal. The rest..she don't matter. (it does to some degree -as a transient carrier positioning system- but I'm fishing for grandiose statements here!!,OK?) If this wasn't true, you can plainly see, that horn speakers would be absolutely impossible for the ear to decipher, as they get EVERYTHING wrong, except that all important positive transient peak.
If you wanted to measure something,and correlate it back to what we hear, then record it in terms of very high frequency digital sampling, and then remove all the unimportant bits (via weighting in analysis) ..and compare that to what you hear.
You WILL get useful results, that are weighted to what the ear actually hears.
You will find that you will go from 0.00XX% distortion, to much, much higher levels. After all, you will go to measuring the points that are actually the most distorted, and weighing that against the original signal, for your distortion analysis.
Re-weight those measurements, and you'll get somewhere. 🙂
yes, induce a nice and warm fuzzy feeling that you spent your money on something that REALLY improves your highend ....
Hey John, Thanks for the pointer to the Analog Devices part. I'll get some and few cap values and maybe a variable one to null the residual. I also own an Ap System 1. It is good for automating measurements, but I have a soft spot in my heart for Sound Technology AND there's a generator level control you can put your hand on and not rely on quick mousing if the magic smoke starts escaping
I am sure I don't entirely follow you here, but you seem to be implying that the ear cannot distinguish a sine from a sawtooth. Anybody can very easily. What are you trying to say? Are you talking about the fact that many different instruments are hard to distinguish when you are deprived or prevent from hearing the attack of the note?
What are the unimportant bits? How would we compare without using an ear to listen to it?
Look into cochlear implants, especially the testimonials of people who have experienced both natural and artificial hearing. It is clear that the progression of cilia within the ear behave very much like a spectrum analyzer. Yes, there is much more to the picture, but this is a good rough start.
KBK said:
Hmmm ... if cable residuals are below 150dB, no matter how much you "weigh" your measurement, at the end of the day a cone in your loudspeaker will simply fail to move with a signal that low.
If speaker doesn't reproduce it (and resolution of a speaker is much lower than required 150db!) how is your ear going to detect it ? Weighted or not.
I design phono input stages, as well as preamps and power amps. The lowest input that I design for is 50uV, which is the operating level of some phono MC cartridges. This is boosted about 10,000 times to reach the final output. If a cable had a distortion mechanism at this level, it would be boosted accordingly. Maybe then, it might be important.
john curl said:
But distorsion is relative to the signal level, so it won't matter what voltage the signal is. If we assume hypothetically that the phono cable would cause distorsion at, say -120 dB, then these distorsion products would still be at -120 dB after amplification since both signal and distorsion are amplified by the same amount. And if cables caused distorsion, wouldn't it seem most plausible that, just like most other components, the distorsion in relative meausure increases with signal level, in which case it would probably be more likely to cause problems with line level signals and speaker level signals than with phono level signals?
Of coure, if you are talking about some type of noise that is uncorrelated to the signal itself, then it is a different matter, but then it is rather noise than distorsion, I would say.
Christer, at least you gave me 120dB, instead af 150 dB. That is a start. Now consider that the distortion mechanism is ACTUALLY a form of crossover distortion, rather than a typical nonlinearity. It appears to be this, because the distortion goes away proportionally at higher input levels, just like crossover distortion would do. Now, consider LOWER input levels than the -30dBV that I tend to measure at. It could be inferred that the measured distortion will actually get worse, rather than better. Then, it may go from -120 db to perhaps -90dB. I don't have the equipment to measure this far down in input operating level, so I can only predict this to be true at this time, from experience.
Now, think about 2 closely spaced high frequency tones, let's say 19,000 and 19,100 Hz playing back from an MC phono cartridge at normal operating level. This will give, for example, at least one beat tone of 100 Hz. Now, play this back through RIAA equilization, and you increase the -90dB to -50db or so. This gives distortion products at an easy level to detect. Far fetched? Yes. impossible? No!
Please consider this possibility first, before condemning it in advance.
Now, think about 2 closely spaced high frequency tones, let's say 19,000 and 19,100 Hz playing back from an MC phono cartridge at normal operating level. This will give, for example, at least one beat tone of 100 Hz. Now, play this back through RIAA equilization, and you increase the -90dB to -50db or so. This gives distortion products at an easy level to detect. Far fetched? Yes. impossible? No!
Please consider this possibility first, before condemning it in advance.
john curl said:
I'm trying to get a handle on this proposed"crossover" distortion and its ramifications in cables. Are you proposing some form of inter-granular rectification mechanism, perhaps along the lines of dissimilar metal potentials? If so, why wouldn't this problem confound radio astronomers who, after all, spend hundreds of thousands of dollars building cryogenic receivers with (squids, bolometers, josephson junctions, Schottky point diodes, whatever) in the hopes that, when counting photons, the detector is able to respond in a linear fashion and transfer the resultant electrical analogue to an amplifier.
Since these signals are many orders of magnitude below 50 uV from a phono cartridge (like picoovolts or less) and have to travel from the detector to said amplifier, I can't understand why they wouldn't have addressed this issue long ago. Answer is, they would have, and it would have been documented. In my 6 years at NRAO, I never once heard any mention of a cable distortion problem, yet we were spending tremendous resources on ever more sensitive detectors and cryogenic equipment, as well as improved front ends/mixers etc. If intergranular diodes were a significant effect, they would most likely be used as an inexpensive detector, rather than all the effort building the exotic detectors currently in use. Not to mention I've yet to see any sort of advantageous use of said effect, since it could perhaps be harnessed to upconvert signals, etc. as a non-linear process.
Maybe I'm just missing something here...
auplater
I think the astronomy analogy has a factor that needs to be considered - signal processing. Many moons ago (I could not resist the pun) I took a graduate level physics course on signal processing, and a component of that course was using Bayesian statistical methods to identify and characterize signals that were buried well below the noise floor. In audio we are listening to a signal whose only processing is the addition of amplification and conversion distortions, whereas in astronomy and physics they use extensive and elaborate signal processing to identify those miniscule signals.
Just a thought.
Cheers, Terry
Just a thought.
Cheers, Terry
OK John, fair enough. I missed the possibility of "cross over " or "quantization" type distorsions. If we were talking of that type of effects, then yes, what you said before would make sense. However, that assumption also strengthens the counterarguments that such distorsion would likely cause problems for, and be known by, astronomers an others.
audibility
Yes indeed, alot of autocorrelation, statistical signal processing does go on in radio astronomy, but that by no means negates the argument... the signal still has to leave the detector and reach the analytical equipment...
Have you ever heard of "cross crystal distortion" other than in the esteemed Dr. van den Hul's commercial "explanations"? Is there some peer reviewed metallurgical or IEEE presentation of supporting evidence for this effect? Grain boundaries indeed have significant and profound effects on material properties, but I'm not sure this is one of them.
If this were indeed true, then it would seem that amorphous structures would be the ultimate cabling structure, as there would by default be no grain boundaries. "Single crystal" or "directional crystallized" cables seem at best a band-aid approach to the "grain boundary distortion / nonlinearities" presented as the logical cause of extremely low levels of distortion which may, in fact, be artifactual.
Perhaps a cable constructed of a semiconducting material, wherein the conduction band carriers can be purposefully modulated by design could be constructed as a "proof of concept"? As in the proverbial "strait wire with gain"....
l8tr
John L.
metalman said:
Yes indeed, alot of autocorrelation, statistical signal processing does go on in radio astronomy, but that by no means negates the argument... the signal still has to leave the detector and reach the analytical equipment...
Have you ever heard of "cross crystal distortion" other than in the esteemed Dr. van den Hul's commercial "explanations"? Is there some peer reviewed metallurgical or IEEE presentation of supporting evidence for this effect? Grain boundaries indeed have significant and profound effects on material properties, but I'm not sure this is one of them.
If this were indeed true, then it would seem that amorphous structures would be the ultimate cabling structure, as there would by default be no grain boundaries. "Single crystal" or "directional crystallized" cables seem at best a band-aid approach to the "grain boundary distortion / nonlinearities" presented as the logical cause of extremely low levels of distortion which may, in fact, be artifactual.
Perhaps a cable constructed of a semiconducting material, wherein the conduction band carriers can be purposefully modulated by design could be constructed as a "proof of concept"? As in the proverbial "strait wire with gain"....
l8tr
John L.
auplater said:
I'm still struggling to get a clear picture on the correspondence between the instrumentation of astronomy and audio. The field of radio provides a example of my hurdle. By similar reasoning the common FM Walkman can be seen as an extraordinary audio instrument. It's frequency range spans almost 10 MHz, approximately 500x that of human hearing. It operates on signals from well below the uV level to, with the proper line sample, 'pick a number'. At the facility I left at three this morning a tuner is used to directly monitor a 50,000 kWatt feed into a multi-user antenna. At a glance this remarkable $20 device would in some ways appear to be leagues above the requirements of high performance audio. Everyone here I hope agrees a $3 amp chip crushes it in that regard.
The difference is tuned circuits. Another quick example to illustrate it. AM broadcast transmitters work by superimposing an audio signal on a single frequency carrier. Ideally that carrier is a perfect sine wave at the broadcast frequency (in the 1 MHz range.) One transmitter I formerly maintained 'approximated' that carrier sine wave with a square wave. How did they get away with this? The second harmonic of 1 MHz is 2 MHz. The AM bandpass of interest is about 10 kHz. Simplifying somewhat, the output of the transmitter simply passed through a passive tuned network which eliminated the harmonics. Almost all older FM transmitters came with passive output filters.
When taking measurements for example in the 10 GHz range, harmonic distortion components of the instrumentation appear at 20 GHz. Audio operates over three frequency decades. Since astronomers get to use liquid nitrogen I get to pull an example from a more advanced digital audio format, 24-bit 96 kHz digital. Is there an analagous situation in astronomy which requires performance across almost four frequency decades to a -120 dB resolution limit simultaneously? To superimpose this on the above example, are astronomers worried about a captured signature 120 dB down at 1,000 GHz when making measurements at 10 GHZ?
What I am most probably measuring is PIM distortion. This is PASSIVE INTERMODULATION DISTORTION. Why don't those of you who don't know what this is, 'Google' it and see what you get? I just did and I got a great article called: 'Passive Intermodulation Distortion in Connectors, Cable and Cable Assemblies' by David Weinstein of Amphenol Corporation, Communication and Network products. It effectively matches what I find.
bad connectors
I guess the issue of the cabling construction is now redirected to the connectors themselves as the cause for distortion components.
Since the jist of the Amphenol article seems to be about crappy connector designs and interfaces at rf and how to avoid them, I agree, it's a good article (if not somewhat obvious to anyone well versed in surface properties of materials). Maybe most EE's need a course in surface chemistry and properties of materials, I dunno.....🙄
No argument that typical audio component interconnect designs suck big time when it comes to quality. I'd love to replace all the cruddy plugs and jacks in my system with hardwiring, but then I'd be soldering all day long 7 days a week since I'm always switching / adding / taking apart to try something new and different...
But I'm not gonna upgrade to $$$$ cables with cruddy terminations because some guy says they have less "cross crystal intermodulation" if that makes no sense to me, and is unsupported in the engineering community. YMMV I've just replaced the crappy terminations with ones that work better, and left the cables as is. Bigger fish to fry... 🙂
So if you're saying that typical cables have lousy terminations that can produce undesirable problems (distortion, loss of signal, etc.) I agree. On all but the most transient rf systems we built, we used semi-permanent copper jacketed coax with BNC or better positive force connectors, or completely hardwired the system (using waveguides as necessary for the high frequency areas as appropo.
Cheers!
John L.
I guess the issue of the cabling construction is now redirected to the connectors themselves as the cause for distortion components.
Since the jist of the Amphenol article seems to be about crappy connector designs and interfaces at rf and how to avoid them, I agree, it's a good article (if not somewhat obvious to anyone well versed in surface properties of materials). Maybe most EE's need a course in surface chemistry and properties of materials, I dunno.....🙄
No argument that typical audio component interconnect designs suck big time when it comes to quality. I'd love to replace all the cruddy plugs and jacks in my system with hardwiring, but then I'd be soldering all day long 7 days a week since I'm always switching / adding / taking apart to try something new and different...
But I'm not gonna upgrade to $$$$ cables with cruddy terminations because some guy says they have less "cross crystal intermodulation" if that makes no sense to me, and is unsupported in the engineering community. YMMV I've just replaced the crappy terminations with ones that work better, and left the cables as is. Bigger fish to fry... 🙂
So if you're saying that typical cables have lousy terminations that can produce undesirable problems (distortion, loss of signal, etc.) I agree. On all but the most transient rf systems we built, we used semi-permanent copper jacketed coax with BNC or better positive force connectors, or completely hardwired the system (using waveguides as necessary for the high frequency areas as appropo.
Cheers!
John L.
"cross crystal intermodulation" is what happens to you when each girlfriend realizes you bought the other a diamond. While clearly audible... it might not be audio.
🙂
🙂
poobah said:
Ha!!
Back in my collegiate days, the girlfriend on the way out "dropped off" a hatchet I had leant her at my front door after I failed to answer it @ 11PM on a sat. nite... as I had a new girlfriend over for dinner, etc., and she'd called a few times "to talk"....
Needless to say, when we left my appt. the next morning, I had some 'splainin to do, to paraphrase Desi...🙂
John L.
- Status
- Not open for further replies.