Ha ha I'll second that one.
Numerous fitbits and other partially working sh1t - I've had my fill personally. And chasing those damn Blutooth orders for '100's of millions of pieces'
Yeah. All bs.
I posted my thermocouple experience up a bit earlier. I don't know what the effects are at audio, but at DC or very very low frequency they are certainly there. But the physics is clearly understood and it's not micro diodes - we've been using it to reliably measure temperature for over 100 years or so.
There are opamp app notes that talk about this stuff for very low level DC measurements as well - see Natsemi for example.
There are opamp app notes that talk about this stuff for very low level DC measurements as well - see Natsemi for example.
Last edited:
Someone remind me, as we seem to have been led around the houses, what was the point of the original alleged experiment with the bit of wire, that morphed into a wire with connectors, etc. I seem to have forgotten the hypothesis? Or is this a phenomenon searching for a hypothesis? Or, just a time sink?
Probably the last. Mix together vague terms ("directionality") and an "experiment" described poorly, season with faux-indignation and insults when the details and definitions are asked for, then bake at 350 degrees for 15 minutes.
Serves four as an appetizer.
Ah, so you're a part of the Tall Ones, too.
BTW 20 years ago I would have told you that digital is the best and SACD solve all of our audio problems. It just hasn't worked out that way.
Is that because you found out that perfect digital sound can be produced by equipment costing less than hundred bucks?
But we didn't know how the fibres were arranged - that was part of the work. If they were mainly near the surface then they could have been re-arranged, or any contamination could have affected acoustic coupling. My point was that I didn't know what effect manual handling might have, so I would have been cautious. She didn't even think of the issue. I guess a mathematician knows even less about experiments than a theoretical physicist.Derfnofred said:
But nothing like 7.5ohms characteristic impedance across the audio spectrum. Well, maybe looking a bit like 7.5ohms from perhaps 10kHz upwards. As I said, 8R at RF is fairly easy; 8R from 20Hz to 20kHz is rather hard.RNMarsh said:
LSI-11 was surely the original PDP-11 on a chip (well, actually two chips on one carrier?). Also know as PDP-11/03 when nicely packaged into a system?
JC complained that we were always dissing him, even when he had support from clever people like VdH, MH and ES. I pointed out that MH has a history of talking nonsense, and ES has a history of doing bad experiments. ES then promptly and very kindly proceeded to illustrate my point.gpauk said:
You don't say why exactly. I will take a rough stab at it. Take a coax cable for example. Even if resistance is low, the formula used to calculate characteristic for coax assumes the wave is confined to the cable and not influenced by external structures such as other cables or ground planes, etc. However, due to the low permeability of copper, skin depth at audio frequencies in coax cable is insufficient to confine the wave to only the cable structure. Therefore characteristic impedance would not be well defined or well controlled. However, that wouldn't explain why only low-Z is hard. Maybe I should think about it some more. Interesting question though.
Last edited:
Actually the peanut gallery is trying to obscure any issues that they don't like. I suggested a simple experiment anyone can try on an audio interconnect. The biggest bit of snow has been the noise pickup. Anyone with practical experience would run a baseline without signal to establish the noise floor. You will note the claims of the peanut gallery have no actually experience or measurements.
These tests can be done with what today is simple gear. But I do use a bit better.
Reminds me of the time I was showing a fellow his crossover network inductor saturated. You could hear it and I was showing him the current draw on an oscilloscope. He turned his head away to avoid looking at it.
Now I did not say what the results would be, just how to do a directionality test, but that was enough to get the religious defenders of the faith to start screaming. Don't worry next they'll tell you describing a test and asking others to try it isn't fair and open. But quite simply they have no clue.
As a normal simple twisted pair runs between 100-150 ohms all you do is parallel a bunch of them. Many use telephone cable with many pairs to do this. As telephone cable has a different twist for each pair it doesn't quite combine as nicely as could be. If you got special cable made where all of the pairs to be paralleled had the same twist it would work a bit better. Then for a mixed cable you could use pairs with a different matching twist for a different driver.
I am not sure that people who know how to do nanovolt level measurements accurately are that thick on the ground, even here. So claiming 'anyone' when you need the right gear and a deep understanding of how to do it right and not confuse noise with signal limits is a little confusing. So to me the peanuts have a point.
I didn't say because it can be found in any good textbook, or found online. The formula for characteristic impedance is Z0=sqrt{(R+jwL)/(G+jwC)} (see Wikipedia, for example). At RF frequencies L and C dominate, so the simpler approximation sqrt{L/C} can be used; as often seems to happen in engineering, some people then assume the simple RF approximation is the whole story and use it where it doesn't apply. At audio frequencies in a normal cable R and C dominate, so characteristic impedance is given roughly by sqrt{R/jwC} - it is frequency dependent and not resistive (roughly equal amounts of resistance and capacitance). Actually, it could be argued that in this frequency domain the cable no longer supports wave propagation but signal diffusion.Markw4 said:
To get a resistive impedance at audio frequencies you need to do some combination of reducing R, increasing L (e.g. add series inductance), increasing G (e.g. add shunt resistance) and reducing C (e.g. widely spaced cables). Unfortunately increasing L and reducing C raises the impedance, which may be why telephone lines were nominally 600ohms. Adding lumped L (the only serious method available) reduces bandwidth.
So, putting lots of RF cables in parallel does not create a low characteristic impedance audio cable. It merely creates a low impedance RF cable. This may still be useful, but not for the reason which some people imagine.
You suggested a poorly-described bad experiment which is likely to lead some people to false conclusions.simon7000 said:
That is not a directionality test. It is merely a good example of how not to do an experiment.
A poor description of a bad test giving misleading results to those daft enough to try it is not a useful way of teaching people anything. Apart, perhaps, about the foolishness of doing experiments with hopelessly insufficient background knowledge to test what you intend to test in a reliable manner.
People will have to form a judgement about who has no clue.
Okay, then it sounds like you are saying that inductive reactance is so much smaller than resistance that the effects of inductance are negligible and resistance predominates. This still assuming the cable as modeled fully defines boundary conditions for wave equations, and that loss of support for propagation is simply a function of, or a definitional point of view regarding, the rate of attenuation.
That would still seem to leave the issue of line level 120 ohm cable as workable over reasonable distances, if not for long distance telephony. If so, why then wouldn't paralleled cables work to lower net impedance as some would predict? Or are you saying 120 line level doesn't actually work either without loading coils over even modest distances? How about many 600 ohm cables in parallel?
- Status
- Not open for further replies.