I will look and see what the OPA627 can do, but as a single opamp chip, I would probably consider it on a different category of equipment. I do recall it being mentioned in "the best sounding opamp thread"....
soongsc, I wasn't joking about OPA627. If you replace whatever OPAs you use with OPA627, I GUARANTEE you and your Golden Pinnae listening panel AND also any true golden pinnae will hear significant differences.
If you can't afford OPA627, I can sell you some Balancing Paint that when painted on even coaxial or RCA cables will fix any imbalance and the sound deterioration due to it. Send me US$100 in used bank notes (no Confederate money please) for a small sample. You will not be disappointed.
Send me US$170 and I will send you a set of 2.6M RCA interconnects of my design without additional shipping charges. Currently working on some new packaging. Should be out again soon. A supplier was trying to convince me to use a specially processed version of wire instead, when he sent me a sample and I did some measurements, they were different, but it only meant I would have to change our own process to match that specially treated wire which meant additional cost without true benefits.
Mr Soong,
Characteristic impedance of any competently designed and fabricated cable will not vary appreciably in the audio range, at best only minuscule amounts. It's entirely a matter of geometry and materials.
If you send someone these 2.6m cables, will you coil them up in order to put them into a shipping box? If so, you will definitely alter their behavior by damaging the geometry.
Mr Lee is correct. The OPA627 is an amazing op-amp with a well-earned reputation. I use them a lot. Fortunately I don't pay retail.
Characteristic impedance of any competently designed and fabricated cable will not vary appreciably in the audio range, at best only minuscule amounts. It's entirely a matter of geometry and materials.
If you send someone these 2.6m cables, will you coil them up in order to put them into a shipping box? If so, you will definitely alter their behavior by damaging the geometry.
Mr Lee is correct. The OPA627 is an amazing op-amp with a well-earned reputation. I use them a lot. Fortunately I don't pay retail.
BrianL,
I have seen data showing differently from what you say, and also my measurements show that it does vary a lot depending on the load impedance. The lower the load impedance, the less variation.
Yes, I will ship them coiled in a bag, which will then be enclosed in a box. Box is normally not available in retail stores. When ordered through mail from us, I do offer a 7 day listening period, and accept returns if you are not satisfied with them provided that you pay for return shipping in the original package that is trackable and they are not damaged. This is not offered by retail stores because I would encourage listening to them before you buy, and different stores have different policies. In the past few years, I have had one return. The reason why I do this is because we never know what the input impedance of customer equipment looks like, and as a customer myself, that is the way I would like to be treated if I purchase something on faith to find that it is not what I expected. For people whom do not like long cables, we also have 0.65M, but the same price.
I will keep the OPA627 in my database. Right now I am focusing on smaller systems of audiophile entry level. There is a lot to be learned throughout the process.
I have seen data showing differently from what you say, and also my measurements show that it does vary a lot depending on the load impedance. The lower the load impedance, the less variation.
Yes, I will ship them coiled in a bag, which will then be enclosed in a box. Box is normally not available in retail stores. When ordered through mail from us, I do offer a 7 day listening period, and accept returns if you are not satisfied with them provided that you pay for return shipping in the original package that is trackable and they are not damaged. This is not offered by retail stores because I would encourage listening to them before you buy, and different stores have different policies. In the past few years, I have had one return. The reason why I do this is because we never know what the input impedance of customer equipment looks like, and as a customer myself, that is the way I would like to be treated if I purchase something on faith to find that it is not what I expected. For people whom do not like long cables, we also have 0.65M, but the same price.
I will keep the OPA627 in my database. Right now I am focusing on smaller systems of audiophile entry level. There is a lot to be learned throughout the process.
Do you have data to back up this claim? If you do, why do you need me to measure the input impedance for you? Sounds like you already have the data.
Did I miss something? I don't recall you posting any measurements. Do you have a link?
Tom
I have not measured the modulus, cable measurements are here:
http://www.diyaudio.com/forums/vend...r-achieving-0-0004-thd-n-196.html#post4460976
You mean this? A random graph with no labeling on the Y axis, no description of how it was obtained, no description of what it shows and how it's relevant. I'm assuming the X axis is a frequency axis...
An externally hosted image should be here but it was not working when we last tested it.
Tom
Last edited:
Y axis explained here.
http://www.diyaudio.com/forums/vend...r-achieving-0-0004-thd-n-201.html#post4462226
The absolute value and curve will vary depending on measurement equipment input impedance, which I believe was over 20KOhm for these measurements, which is in the range of normal audio equipment. Of course X axis is frequency.
This links to post #1960 of this thread which I quote here:
This shows a graph "impedancevariationexamp.gif" that has a bottom scale that is labeled logarithmically from 10 to 40k (I assume Hz, but many of us know what "assume" means), and a linear vertical scale, but the vertical scale has no markings. Who knows what this represents? Soongsc maybe, but the graph should say if it is to mean anything. Labeling both axes of one's graph (and having a title saying what the graph is, and labeling each trace/curve to say what IT is) is a basic requirement of science and engineering, else one is not communicating well enough to be useful.
The five curves (which all have roughly the same shape) all go UP from 10 to 100, then are roughly flat to 1k where there's a falling slope, then back up to a sharp peak at 40k. I can guess this is impedance magnitude, but it should really include phase as well.
But it doesn't look line any cable I could imagine. The actual audio-band impedance of any cable less than thousands of feet long would be purely capacitive reactance, which would fall as a straight line (on a log-log graph) with increasing frequency and with the phase angle staying very close to 90 degrees.
If this were an actual cable impedance curve, I would avoid the cable at all cost.
What I mainly hear (and I hesitate to veer from the tech talk, but I'm making an exception here) is the old theme song from "Rawhide."
https://www.youtube.com/watch?v=2KPplYp7K7M
I guess junior needs a 'my first jacob's ladder
. But 'X doesn't cause audible breakthrough' is not a quantifiable measure off goodness. Would be nice to have some real measurements for the file.
Is there confusion occuring between imbalance between the twisted pair (which the receiver handles) and capacity imbalance to shield which causes differential mode pickup? for the latter as the papers show its a simple case of picking the right cable. Something designed for the task, not some audiophool mash up?
Could really throw the cat into the birds by putting starquad into the discussion under the heading 'severe overkill for domestic installations'
Could really throw the cat into the birds by putting starquad into the discussion under the heading 'severe overkill for domestic installations'
Actually, the THAT 1200 series addresses exactly that by implementing the bootstrap bias patent described in Whitlock, Bill: A New Balanced Audio Input Circuit for Maximum Common-Mode Rejection in Real-World Environments.
Easily answered: it will cause variation in sound quality once the imbalance becomes severe enough...
A comparison between balanced implementations with and without the abovementioned bootstrapping method is provided in: Israelsohn, Joshua: A matter of balance: line-receiver CMR in real-world environments (alternative link: EDN).
So the effect of mismatched impedances on the CMRR is known.
Do you offer an alternative source of variation in sound quality?
As in: granted it's caused by (small scale) impedance mismatch, but not differential and/or common mode distortion?
Last edited:
Modulus-86: Composite amplifier achieving <0.0004 % THD+N.
In the quoted articles that I could quickly browse through, I see nothing that takes the impedance drop at high frequencies of cables into consideration. If the impedance drops to 1K Ohm at 20KHz, probably not all source drivers can drive that load, what happens at both differential outputs? Additionally it assumes 100Ohm source impedance in the report, how universal is this at the high frequency end of the audio spectrum?
In the quoted articles that I could quickly browse through, I see nothing that takes the impedance drop at high frequencies of cables into consideration. If the impedance drops to 1K Ohm at 20KHz, probably not all source drivers can drive that load, what happens at both differential outputs? Additionally it assumes 100Ohm source impedance in the report, how universal is this at the high frequency end of the audio spectrum?
Last edited:
I think the articles do: by revealing that CMRR gets worse at high frequencies.
You apparently can't have everything.
That would depend on the particular output driver circuit.
But I don't see how this relates to the THAT 1200.
How would a (linear) receiver possibly correct for errors that happened before transmission?
What circuit are you discussing?
I refer to the Modulus-86, which isn't a line driver.
It doesn't act as a source to a shielded twisted pair, but certainly stays below 100Ohm at 20kHz.
Can you give a particular example of a widespread shielded twisted pair cable type that you use as a data point for your argument?
Last edited:
Could it be that for any competent cable and amplifiers this is a non-issue? I assume your big concern is over http://audiosystemsgroup.com/TransLines-LowFreq.pdf
This is very incorrect.
It's a common error to use the simplified radio frequency formula to calculate Characteristic Impedance. Using the correct (much longer) formula, you will see that:
The Characteristic Impedance of a cable will vary by nearly two orders of magnitude through the audio spectrum.
From the Jim Brown paper:
"Transmission Lines at Audio Frequencies, and a Bit of History"
The behavior of cables at audio frequencies is widely misunderstood.
This tutorial attempts to shed some light and bring some sanity to the discussion.
http://www.audiosystemsgroup.com/TransLines-LowFreq.pdf
I think the articles do: by revealing that CMRR gets worse at high frequencies.
You apparently can't have everything.
That would depend on the particular output driver circuit.
But I don't see how this relates to the THAT 1200.
How would a (linear) receiver possibly correct for errors that happened before transmission?
What circuit are you discussing?
I refer to the Modulus-86, which isn't a line driver.
It doesn't act as a source to a shielded twisted pair, but certainly stays below 100Ohm at 20kHz.
Can you give a particular example of a widespread shielded twisted pair cable type that you use as a data point for your argument?
Well, I was reading about the discussion related to input filtering and balancing, and could not see the benefit spending time on such since the interconnect creates a filtering effect and imbalance of it's own. I would assume one would want to integrally take cable variation into consideration so that as long as you have cables within certain characteristic, any cable should not cause to much audible difference where a strong preference will occur. Right now none of the commercially available products provide sufficient data to make the proper match.
Can you quantify or provide an example where this is a problem? It's pointless to discuss the mere possibility of a problem. Of course problems are possible.
For every choice of a certain characteristic you can find an argument against that particular choice.
But there are findings i.e. from communication networks that suggest shielded twisted pair of average quality doesn't have a problem with imbalance and rising impedance that can't be dealt with in the audio band.
Other than that, I understand that you wonder why most manufacturers don't publish more data. But THAT Corp. did!
For every choice of a certain characteristic you can find an argument against that particular choice.
But there are findings i.e. from communication networks that suggest shielded twisted pair of average quality doesn't have a problem with imbalance and rising impedance that can't be dealt with in the audio band.
Other than that, I understand that you wonder why most manufacturers don't publish more data. But THAT Corp. did!