Digital that sounds like analog

[LL21]

As to the silver wire, my friend and I have no idea why. it doesn't show up in measurements either. but my friend has built a fair number of dac and we find there is a significant difference. I was just hoping you could shed some light on this as the cost for silver wire for grounding is just a couple of dollars. Not a problem if not.

Interesting. Audio Note makes a big deal of this in many of their products (particularly their latest NOS DAC)...though that could just be marketing.

 

[LL21]

As for trying silver wires - I'd need at least a preliminary hypothesis for why silver might be better than tinned copper - do you have any suggestion as I can't think of one? I usually only do experiments to test falsifiable hypotheses, that's the essence of science.

Conductivity differences of the two metals? speed/ease with which one conducts electricity makes it a better grounding material perhaps?

 

[LL21]

Conductivity differences of the two metals? speed/ease with which one conducts electricity makes it a better grounding material perhaps?

"The conductivity of annealed copper is 59 x 10^6 siemens/meter. The resistance of a 24-gauge copper wire 1,000 feet long at room temperature will be about 26 ohms.
The conductivity of silver is about 7 percent higher than for copper, or 63 x 10^6 siemens/meter. A silver 24-gauge, 1,000-foot-long wire would measure about 24 ohms.

Because its conductivity is the second highest of any metal and its cost is low, copper sees use in most wire, connectors, printed circuit foils and related electrical parts.

Silver's higher conductivity and cost make it a niche product. It's used as wire and solder in specialty audio electronics."

 

[opus111]

Conductivity differences of the two metals? speed/ease with which one conducts electricity makes it a better grounding material perhaps?

Yes in which case a shorter ground wire is a better optimised engineering solution than using silver. It doesn't need to be very much shorter, just a few percent as you point out the conductivity difference isn't huge. I'll go for the shorter copper wire if that's the hypothesis

 

[LL21]

Yes in which case a shorter ground wire is a better optimised engineering solution than using silver. It doesn't need to be very much shorter, just a few percent as you point out the conductivity difference isn't huge. I'll go for the shorter copper wire if that's the hypothesis

Yes, i have read oftentimes that people try to make the connections/wiring in DACs 'as short as possible'...makes sense. in all out assault, then presumably both as short as possible plus silver would be ideal?

 

[opus111]

As to the silver wire, my friend and I have no idea why. it doesn't show up in measurements either. but my friend has built a fair number of dac and we find there is a significant difference. I was just hoping you could shed some light on this as the cost for silver wire for grounding is just a couple of dollars. Not a problem if not.

My tentative hypothesis is that there's some other design problem with the DAC which is being modified by the change to silver. Silver might turn out to be more lossy at RF for example due to different skin depth - I get changes from fiddling around with RF effects, like adding ferrite beads. But without something to learn from this, I'll put my dollars into other mods - I don't know how far to change the grounding into silver for example. Does the effect extend to the earthing wire in a mains cable? Or does the DAC's chassis need to be silver too, not just the wire going to it? Without a hypothesis to test its too hard to set up a controlled experiment.

 

[opus111]

Yes, i have read oftentimes that people try to make the connections/wiring in DACs 'as short as possible'...makes sense. in all out assault, then presumably both as short as possible plus silver would be ideal?

Well then we do have a sensitivity to a particular parameter which probably could be reduced by modifying the design. So that changing the ground impedance has only an unnoticeable effect. This is one reason my DAC is ballanced - to get away from as many of these odd grounding effects as I can.

 

[LL21]

Well then we do have a sensitivity to a particular parameter which probably could be reduced by modifying the design. So that changing the ground impedance has only an unnoticeable effect. This is one reason my DAC is ballanced - to get away from as many of these odd grounding effects as I can.

Interesting. Thanks.

 

[Phelonious Ponk]

"The conductivity of annealed copper is 59 x 10^6 siemens/meter. The resistance of a 24-gauge copper wire 1,000 feet long at room temperature will be about 26 ohms.
The conductivity of silver is about 7 percent higher than for copper, or 63 x 10^6 siemens/meter. A silver 24-gauge, 1,000-foot-long wire would measure about 24 ohms.

Because its conductivity is the second highest of any metal and its cost is low, copper sees use in most wire, connectors, printed circuit foils and related electrical parts.

Silver's higher conductivity and cost make it a niche product. It's used as wire and solder in specialty audio electronics."

So I'll worry about this the next time I'm running 1,000 feet of wire. Seriously, do the differences narrow as the lengths get shorter? And by the time we get to any length of wire that might actually be used in practical application have those differences all but vanished?

If I acted on what audiophiles hear I'd be demagnetizing plastics and pasting pet rocks to my walls. If I report what I hear I'm simply told that my system must not be resolving enough. There's no arguing religion. What's the science?

Tim

 

[bbb]

I don't know how far to change the grounding into silver for example. Does the effect extend to the earthing wire in a mains cable? Or does the DAC's chassis need to be silver too, not just the wire going to it?

You are getting ahead of yourself. As I said earlier, I have no explanation why, so I thought I would try to tempt you to experiment with very little cost. Since you have indicate your reluctance due to the lack of theorectical evidence, I accept that and I think we should leave it at that. Anyway, the change we found was a much lower noise floor but don't ask me why. This is across a few platforms including a SD type that we are currently working on. And yes, current segment still got it beat easy even though we put much more work into it.

 

[opus111]

Thanks for sharing the experimental result of lower noise floor - was that subjective noise floor - like the perceived darkness between notes? I have found that changing grounding topology makes that kind of difference - going from the rather haphazard layouts I've found in many products I've reverse engineered, adopting rigorous star earthing pays great returns in that dept. This I surmize is down to reduce coupling between circuit elements through common ground impedances. Did you have common ground impedances when you introduced silver into the mix?

 

[bbb]

Yes it was subjective in the sense that it does not show in measurements. We are pretty sure it has nothing to do with common ground topology. My friend actually makes his living manufacturing parts for pro audio so he does know a thing or 2 and we are stumped. We did this initially because like you we doubted the claims of certain quarters and also because we just could not get the noise floor of the SD dac good enough for us (irregardless of what the measurements say) so we thought what the heck. We have actually taken it much further than just grounding and it is the same result as well. I don't really want to take this any further here because I don't have any explanation for it and don't want unwanted noise to drag this thread into the abyss.

 

[LL21]

So I'll worry about this the next time I'm running 1,000 feet of wire. Seriously, do the differences narrow as the lengths get shorter? And by the time we get to any length of wire that might actually be used in practical application have those differences all but vanished?

If I acted on what audiophiles hear I'd be demagnetizing plastics and pasting pet rocks to my walls. If I report what I hear I'm simply told that my system must not be resolving enough. There's no arguing religion. What's the science?

Tim

no idea...was just providing the answer to the question someone asked about copper v silver...and i was guessing it would have to be conductivity not knowing much about any other differences. Does a 7% difference in materials matter in short runs? Up to the engineers to answer that one.

That said, i am someone who has found differences in power cables and interconnects/speaker cables, so i would expect wiring does make [some] difference.

 

[DonH50]

Unfortunately most the reasons I could postulate for why cables sound different do not much depend upon the material, particularly at audio frequencies. "Silver is harder than copper, therefore silver cables must be less microphonic and less reactive to sound waves inducing micro-tremors in the cables that corrupt the sound." Might make good ad copy but makes very little practical technical sense. There seem to be lots of questionable claims about cables and wiring in general. The few times I have delved more deeply it boiled down to interaction among wire and components.

@bbb: First, thanks for the kind words. Next, phase, polarity, and time can be related but that discussion should be held elsewhere. Polarity is generally considered an absolute, it is "positive" or "negative", 0 or 180 degrees. In fact, changing polarity means applying a 180-degree pahse shift at all freqencies. Speaker crossovers (among other things) add phase shift that is frequency-dependent so it is possible for a speaker to be "in-phase" at one frequency and "out-of-phase" at another. It is often a design goal that all drivers are in-phase at the crossover point where both drivers (say woofer and midrange, or midrange and tweeter) are working together when they overlap instead of cancelling each other out. As an aside, and to bring us back on-topic, the type of filters opus111 is building provide optimum phase response and thus optimum time response.

opus111: It might be interesting to folk if you were to graph frequency (Bode) plots and time (step or impulse) response for say your elliptical filter and the same order in a Butterworth design. I would help, but am swamped with work and performances this week/weekend.

 

[Kindhornman]

Richard,
I thought that I would add this reference but I am going to assume that you already know of it. http://www.analog.com/en/amplifiers...ign-sim-tool/filter_wizard/resources/fca.html Just wondering if this would work instead of your inductor filter though it is an active circuit.

Steven

 

[microstrip]

So I'll worry about this the next time I'm running 1,000 feet of wire. Seriously, do the differences narrow as the lengths get shorter? And by the time we get to any length of wire that might actually be used in practical application have those differences all but vanished?

If I acted on what audiophiles hear I'd be demagnetizing plastics and pasting pet rocks to my walls. If I report what I hear I'm simply told that my system must not be resolving enough. There's no arguing religion. What's the science?

Tim

Tim,

The difference in sound of cables is due to a very complex mix of factors, and I think you know pretty well that there is no accepted universal theory in cable sounding. I think most of us accept that most of the time the cable sound can not be associated directly to the resistivity, capacity or inductance, but that factors that affect these properties of cables can also affect sound. I am addressing aspects such as non ideal dielectric properties, mechanical susceptibility and even conductor composition and finishing. The interaction of these variances with the equipment is not also easy to predict, but some cables designers manage to create some typical sound signatures.

This timer I will resist addressing your second paragraph punchline.

 

[opus111]

I thought that I would add this reference but I am going to assume that you already know of it. http://www.analog.com/en/amplifiers...ign-sim-tool/filter_wizard/resources/fca.html Just wondering if this would work instead of your inductor filter though it is an active circuit.

Thanks Steven, I had seen it before but had forgotten. I had a play, putting in my parameters. If anyone else wants to try, the parameters for my filter are that the passband edge is 17kHz, stop band 22kHz, Amax 0.2dB, Amin 53dB. It tells me I need >8th order filters, but when I select 'Butterworth >8th order' I get an error. Not very helpful!

I do intend to design an active filter based on my passive filter and I'll put up details when I get to that. Right now I'm building the second prototype DAC so that I can transport it for listening evaluations - the first is just a jumble of boards perched atop one of my speakers.

@Don - yes, agree would be interesting, right now up to my neck in winding filter coils

 

[opus111]

http://www.elec.qmul.ac.uk/people/j...maDeltaModulation-SolvedandUnsolvedIssues.pdf

Just found this interesting paper - from 2008, which explains some of the unresolved issues with S-D converters for audio. It does though contain a significant oversight in terms of how to know when noise modulation has been licked. This from p59 :

With TPDF dither both the average quantization noise and the average
quantization noise power are constant, thus the noise modulation has been
removed. However, due to the finite bits in the quantizer, the results do not hold
for the lowest and highest quantization levels with RPDF dither, and do not
hold for the lowest two and highest two quantization levels with TPDF dither.

I don't see how it follows that because the average noise is unchanging, therefore there's no noise modulation. Rather like saying because the average temperature in my apartment is 20oC therefore it never feels cold

 

[opus111]

Another immensely satisfying piano disc from DG - pure, rich piano tone and no audible noise modulation. Li Yundi (who now incidentally has dropped the Li part of his name, and moved over to EMI) is a ChongQing born pianist who won the Warsaw Piano competition at age 18, back in 2000. This release is full of glorious Chopin, played magnificently

http://www.amazon.com/YUNDI-LI-CHOPIN-BEST/dp/B00338T674/

 

[opus111]

Festive selection

Since its the festive season, I'm currently spinning this vivacious release from Yo-Yo Ma and friends. Splendid energy and vitality here only slightly marred by sibilance on vocals. If you're fed up with the standard versions of the Christmas music, try this for a little variety

http://www.amazon.com/Songs-Joy-Peace-Yo-Yo-Friends/dp/B001BN1V8U/