Capacitance doesn't figure - in this model it is shunted by source impedance and load. The TL model works fine, and discrepancy between LCR and TL prediction is negligible anyways. Single cell LCR should be most accurate for audioband risetime, but AFAIK you still use fast step changes, the cause of discrepancy here. If you observe a dependency on C/length in this scenario there is something wrong with your model or you are using fast rise times.
But assuming spice handles it, I'll waste an hour of my life confirming a large scale discrete LCR mimics a TL, and which in this scenario for audioband risetimes must produce near identical results as TL and single cell LCR.....
Yes, I know. In your model.
No, the fast step changes do not cause a discrepancy. The tl model is EXACTLY the same as a 201 element LCR model.
It has nothing to do with fast rise times. TL and 201 element agree absolutely. And, as Davis said, adding the capacitance will speed the system up.
Spice does handle it, I believe that is what Scott used. And he confirmed that a very fast rise time had nothing to do with the accuracy, they superimposed.
Which is what I've been saying all along.
jn
A big discrepancy arises because, with a fast rise time, your choice of where on the resultant waveform to measure 'settle time' affects the outcome. You choose 95% level, which gives a different outcome to say 69%. So, once again, your choice of conditions affects the result. Whereas with audioband risetimes, wave shape is in practice well conserved, and latency effectively uniform throughout the risetime. Only at the 69% mark does fast rise latency match audioband risetime latency.
And you'd be very worried if a large discrete LCR model did not match TL results. For audioband risetimes all three models should match well in any event. This must include the effect of source/load impedances shunting cable capacitance. BTW, this renders variation of the rf characteristic impedance of the TL irrelevant, as it should be of course.
Anyways, I'll run the large discrete LCR spice model tomo and see.
Scott, I've never experienced this myself, I've never played with it, at all - the cable goes in, as is - my last fancy stuff was 30 years ago.
There may be an effect, depending upon everything, as usual; Dan has come across something, and it's working for him trying the options - so far, I've got what I wanted without worrying about such issues, it hasn't come up on my radar ...
Edit: the directionality thing would be all about the fact that every electrical part is made of up of real materials, which do not have the "perfect" properties so beloved in textbooks. There are always secondary behaviours, which 99.99% of the time contribute sweet FA to the device working correctly or not - audio is one of those funny areas where unfortunately it does appear to have some impact.
Last edited:
For some strange reason I'm led to using the field of winetasting as analogy, 🙂 - which is very simple: you compare competent wines, for various qualities. And what is a competent wine? Well, one without flaws or faults, a very straightforward concept - very easy to look up what this means.
And if you get a incompetent wine, no-one is going to ask the judge to do a double blind test to check whether that is the case or not - he likely spit, wine 😀, in your face ... there are clear "signatures" to the wine not being 'right', part of the learning about wine is to recognise these.
And exactly the same applies to audio ...
Unfortunately some of the world's greatest wines are technically flawed by academic standards.
Which still may be acceptable in the bigger scheme of things, because there is a special quality in the wine which is worth the "price" of that flaw. Here a balancing is taking place - technical incompetence offset by the quality of the raw "ingredients" which went into the wine ... there are plenty of audio systems that could be said to fit into this space ...
My comment about audio needs being "funny" still stands - human ears are incredibly adaptive, they can ferret out an irritating flaw in what they're hearing, instantly judge whether what they're listening to is up to scratch. That's why I talk about having the volume up when listening: most systems fail, and fail badly, on this test - all the little defects are now so obvious, it's impossible to listen to for any extended period; for pleasure, at least. The sign of a system working correctly is that there no obvious audible problems when this is done - there is nothing in the sound that offends or disturbs one, you don't have to "put up with it being loud", it's intense and pleasurable for all the right reasons.
Last edited:
No, I leave that to folks over on your side of the big water - you lot are masters at that sort of thing ... 😛.
I for one would love to see such information on audio cables from a reputable source....
you never do come up with anything just more audiophool nonsense, quite often it is the scientific facts that tell you there will be no audible difference, something else you often want to dismiss.... Also don't forget a lot of us work in the world of electronics, in my case over 7 years on Mil communication(especially hearing protection) equipment, so maybe we also actually see experiments, work with acoustic engineers and do experimentation.
Provide some proof, science based.....
As Scott said and as I have asked endlessly, why is this only something that happens in esoteric audio? This is one of the more stupid audio myths that distracts from the real issues and makes the hobby look silly...
Where is the proof?
As expected, building and running a 201 element LCR model shows no differences with a single cell LCR, nor the spice TL model, for an audioband risetime.
Attached is a simulation of a 25uS risetime signal into a 4m cable with approx. parameters same as cable 7 from the Davis paper, a Beldon parallel pair, at JN's request. Total L 3.456uH, C 432pf, R 0.04R. The load in this image was 8R. It's easy to see the latency is c 0.432uS, which is negligible.
One might choose a lower impedance load and obtain more latency - 1R provides c 3.4uS, all as expected, and no significant difference between the 3 models.
I wasn't expecting an epiphany, and didn't obtain one. What was I supposed to discover using the 201 element LCR that the other models didn't reveal again?
Attachments
For an audio cable to be directional would require that some material used in its construction had significant coupling between electric and magnetic responses - for directionality you have to sense both voltage and current. Such materials exist, but they are rare and have to be specially made from the kinds of things (e.g. certain ceramics) which are not a normal part of audio cables. Hence normal audio cables cannot be directional.fas42 said:
You mean that careful simulation shows that a single cell model gives identical results to a 201 cell model in the audio band? How can this be? It was only a theory, and everyone knows that theory is just the ignorant speculation of boring engineers with poor hearing. How could theory happen to be true?luckythedog said:
Why cables may be directional doesn't interest me - what does is whether they effectively form a path of good conductivity throughout their length, between the electrical parts they link. And often they don't, which is why a major tweak I virtually always carry out is to hardwire the complete system - having become very sensitive to the distortion artifacts introduced by poor metal to metal contacts I have a hard time listening to systems where this is an audible factor. I would be very confident that many of these directionality possibilities simply disappear by doing this mod ...
You should hire yourself out to CERN or similar, you are more sensitive than some equipment I have seen in use for measuring these effects, its incredible.....
How does a poor metal to metal contact manifest itself in your world.....
- Status
- Not open for further replies.