[Garbz]

Mornin All

I read on here somewhere a long time ago about some cable termination. This termination used some type of resistor and was supposed to affect how the source and amp see each other's load or something. It could also have had to do soemthing with how the signal is reflected over the cable, which would make the termination method similar to the SCSI bus. I'm very sketchy on the details.

I seem to recall that Kevin Gilmore's name was mentioned, and some type of resistor was used, but after searching around a bit all i found was the difference types of plugs used to "terminate" cables.

I also remember that it significantly affects volume if that helps.

I'm about to make some interconnects, standard RCA - RCA and i was wondering where I could find info about this. That is assuming that whatever was being talked about at the time wasn't some hypothetical or patented approach, that would land me on the short end of the stick of someone else. The RIAA is bad enough atm

Any info?

 

[drewd]

Boy, I hope that this isn't some audio religious issue, like double blind tests or cryogenic cables or something like that.

Terminating cables is something that has real value for high frequency transmission lines - like in the megahertz region. The idea is that you would like the entire transmission path to have the same impedance from the driver to the receiver, then terminate the end of the path with that same value. The reason is that when the wavelength of the signal is significantly shorter than the length of the transmission path, some of the signal can be reflected back from the end of the cable and interfere with the next series of transmissions. Think of what happens when you drop a rock in a pond...if the waves reach a hard surface, they'll bounce back and interfere with the incoming waves. That hard surface has an impedance mismatch with the water.

But that analogy only goes so far. With high frequencies, you can terminate the transmission line with a lower resistance and still get the same reflection damping effect, but then you're pouring excess energy into the end of the cable. Remember, all of this assumes that the impedance of the transmission path is constant from the driver to the receiver, across any cabling, printed circuit boards and connections.

So, back to audio. In the range of a few tens of kilohertz, the wavelengths that we're dealing with go as short as 5000 meters (30KHz, assuming a propagation speed of .5c). Unless your interconnect is on the order of ten or twenty kilometers, you're not going to have a problem with reflections, so you don't have a problem with terminating the cable with a particular resistance.

It's far more important to terminate the cable with a good connector and a good mechanical connection to avoid introducing excess capacitance and resistance which can attenuate the signal in their own special way.

By the way, high frequency (50MHz to 400MHz) signals is what I work with in my day job. That's where significant problems with proper resistive termination come into play in a very significant way!

Hopefully I haven't stuck my nose into more audio religion!

-Drew

 

[ftjandra]

Zobel network maybe? But, that is only used on speaker cables and consists of a resistor and a cap...

--Ferdi

 

[strohmie]

I'd searching for the word ProSink with eric343's name as the poster, or visiting the Audiogeek (eric343's) website.

 

[Punslayer]

I believe that Garbz is referring to the "prosink" termination used by eric on the Nitrogen cables. Search for prosink, or visit the audiogeek website from the sponsors list at the top of the page to find out more. And i believe that it is based on high-frequency transmission line practices.

 

[Garbz]

ahhh thanks

without persuing any audio religion

I don't know anything about this it's just something i read

Does it have any merits? I mean the general concensus seems that it's good.

But if what i'm reading is correct and in the spirit of DIY wouldn't it be best to terminate the souce and the amp itself rather then the cable?
This seems like a pretty interesting topic i don't see why it isn't more widely discussed (other then it may not bring any sonic benfits

)

Anyway if that's the case a 50ohm resistor in parallel (sig to gnd) before the pot in my amp, and a 50ohm resistor in series (sig to sig) before the plugs on my CDP would provide these benefits. Oh and before u ask i pulled that number out of my ... .

 

[Glassman]

the whole path has to have the same impedance, included all connectors.. get some good 75Ohm coax cable, 75Ohm BNC panel & wire connectors, put 75R resistor in series with the signal leaving the source's output stage, put 75R resistor from signal to ground in your amp or even better make yourself a stepped attenuator with 75R total impedance and here you go, the best analog interconnect you can get..

 

[00940]

check that thread http://www4.head-fi.org/forums/showt...hlight=measure

here is what K. Gilmore said in there:

Quote:

Take two of the most different sounding cables available and properly terminate each one. Now listen to them again. It will be almost impossible to now tell these cables apart. So all the marketing speak that goes with all these cables just got thrown out the window. Bottom line is that after properly terminating the cable, all that is now left is the real quality of the cable itself, meaning quality of shielding, and consistency of impedance over the length of the cable. Many cables with various twists and other geometry BS test very poorly for consistent impedance. ... Series resistor at the source end (in many cases this can be eliminated) Parallel resistor at the other end. Only 2 ways to correctly measure the impedance of complex cable geometrys. One is a TDR. The other is a fast square wave generator a decent oscilliscope and a dual 150 ohm pot. One section wired in series with the input end of the cable, the other wired in parallel to ground. Tune pot for best square wave at the destination. Then measure the pot with an ohmmeter. That is the real characteristic impedance of the cable. The number will be a minimum of 46 ohms, and a maximum of about 120 ohms. alpha-core cables may be less than 46 ohms, i have not measured them.

 

[Garbz]

ok so the only question left is if a standard 75ohm coax cable sounds better then world+dog cables why aren't they used as interconnects everywhere?

Anyway i'm going to do a bit of experimenting.

 

[Publius]

To comment as the devil's advocate, to chime in with others, while impedance-matched cables may in fact test better, there's no particularly cogent explanation as to why they would sound better, and no DBTs have been done to suggest otherwise. Unterminated cabling WILL attenuate the audio signal, but only on the order of a thousandth of a dB or so. Of course, if you're not going to do DBTs then terminated cables may be a pretty interesting avenue to explore.

The lack of termination? Feh, what about the prevalence of unbalanced cables, even in audiophile systems? That arguably has at least as much of an impact on sound quality as termination does, and it becomes vastly more important for longer runs.

 

[drewd]

To continue as the devil's advocate, testing a cable with a TDR and a square wave generator is pretty far from the reality of audio. Yes, it certainly will reveal things about the nature of the cable and proper termination will most definitely result in a significant improvement in the step response of that cable, but it doesn't have a direct correlation with the audio performance of the cable. We're driving the cable with frequencies that are far too low to be impacted by electrical termination. A square wave contains extremely high frequency components and proper electrical termination is critical to the propagation of those high frequency components.

At high frequencies, the cable acts as a transmission line. Its behavior changes from that of a wire to something more like a series of resistors and capacitors. Thus, it's important that the final component of that cable is a resistive element that matches the characteristic impedance of the transmission line. But at low (audio) frequencies, a cable, even a 75 ohm coaxial cable, is a wire. That's because the RC time constant of the cable is so much smaller than the period of the highest frequency signal - the capacitance is charged nearly all of the time, versus a high frequency signal that may be charging and discharging the capacitance almost as often as it is fully charged.

I don't buy into the notion that you can hear a difference between terminated and unterminated cables. I've tried to hear it and I can't. But I do buy into the notion that a poorly mechanically terminated cable or a cable made with substandard components can have an adverse effect on sound.

I think that there is a lot of hocus pocus that goes on in the cable world (let alone the rest of the audio world). I think that the real deal is that at audio frequencies, a wire is a wire. A crappy, oxidized, poorly mechanically terminated wire will certainly sound worse than a high quality, properly mechanically terminated wire.

Just my two cents' of devil's advocacy!

-Drew

 

[Garbz]

kewl i'm liking this discussion more each day.

I'm part of the group buy for the silver wire here and i was thinking of interconnects. Now i suppose whacking some cardas badboys on the end will probably be the best solution, i'm just trying to get a full understanding ov all avenues.

Devils advocates are not unwelcome in these threads. Infact i'd rather hear from them since it'll probably save me some money

.

On the topic of square wave generation, you say it has little to do with audio quality but i see a lot of amplifier datasheets grade opamps on how well they repoduce a 1Mhz square wave. Wouldn't it make sense that if a cable can perfectly repoduce a 1Ghz square wave that it will be good for other frequencies too? (note to devils advocates, don't go overboard with physics )

Anyway i'm begining to think my $20 cheapies are aged compared to the rest of my system and i'm looking for soemthing new. Something DIY

 

[drewd]

It's not so much that it has little to do with audio quality, but that there is not necessarily a correlation between its square wave performance and its "goodness" as an audio cable.

Step response is definitely one way of measuring the performance of an opamp, but we're talking apples and oranges here - the square wave of an opamp is measuring the speed of the output devices. If we measure a MHz or GHz square wave into a cable, we're looking at completely different qualities.

But I'm not completely pedantic!

I understand the question - if a cable can reproduce a MHz or GHz square wave accurately, then shouldn't it be able to reproduce an audio frequency accurately? The answer is...I don't know. You'd have to measure its behavior at those frequencies - and by measure, I think that listening is as valid a measurement technique as electrical tests.

By the way, the pedantic, physics-laden, devil's advocate answer would be "no". Just as an example, the USB cable connecting your mouse to your computer responds excellently to a MHz square wave, but I suspect it would be a lousy audio cable...but it wasn't intended to be one, so that's not a fair comparison.

I suspect, though, that you could design an audio cable that gave great results at GHz frequencies and was a great audio cable, too. But I also suspect that you could design a great audio cable that was a pretty bad GHz performer, too. So I don't think that there's much of a conclusion to draw.

If you're looking toward a DIY cable to replace your old ones, look at the Canare Star Quad cable sticky. Most of my cables are based on that design and they work exceedingly well - and they look pretty darned spiffy, too! But they probably would look terrible on a TDR with a square wave

-Drew

 

[Garbz]

thanks drewed.

I think i'll just use the standard method at this point. Grab some silver wire when it arrives, glue some rohdium plugs on the end. Slap on some stickytape and it's good to go

 

[13405]

Quote:

A square wave contains extremely high frequency components and proper electrical termination is critical to the propagation of those high frequency components.

In the context that you take a square wave and apply fourier analysis to it? Surely square waves do occur in some music, such as electronica etc, that is of course presuming that 44100Hz sampling from a CD and then through a DAC can produce a square wave containing these.

Also, I presume there is always an imperfect digital 'element' left from the outputs of DAC's (where the 'smoothing' routines/methods have not completely been eliminated) In which case there are also traces of a square wave present here too. Could these harmonics also be affected by a cable which has not been terminated properly.

In anycase, when the cables are terminated properly there seems to be an argument that they DO sound differently. And coming from Mr Gilmore I doubt that this is a case of being 'in the head'

Thaughts?