speaker cable myths and facts - Page 106 - diyAudio
Go Back   Home > Forums > General Interest > Everything Else

Everything Else Anything related to audio / video / electronics etc) BUT remember- we have many new forums where your thread may now fit! .... Parts, Equipment & Tools, Construction Tips, Software Tools......

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 3rd July 2013, 05:52 PM   #1051
diyAudio Member
 
jneutron's Avatar
 
Join Date: Jan 2004
Location: away
Quote:
Originally Posted by Bob Cordell View Post
How short?

But I probably don't have the time right now.

If I did it, I'm guessing I would do it for sinusoids of different frequencies using 2 channels of my Agilent 8 Gs/s DSO, which has a horizontal sweep down to 100 ps/div. I would compare the delays of the sinusoids at each end of the cable after calibrating out any residual scope channel delay mismatch. The test cable would be configured in an out-and-back loop so that the same-length probes could be used for measurement at both ends. The loop would have to be configured as large as possible in diameter to minimize any possible crosstalk between different locations along the cable.

Does that seem like a good way to do the measurement you are thinking of?

One could also do it with square waves or very short pulses, but I don't think that would reveal the frequency depedence you seek (but would subjectively reveal time dispersion incurred in the cable).

Cheers,
Bob
The main issue is capturing the delay of say, a 500hz sine at zero crossing down to the microsecond level.

Unfortunately, my computer with the drawings is currently doa. The test setup is to drive the zip cable using a low z out amp, 15 or 20 feet of cable, to a varying load resistance far end. At the amp, use a .1 ohm current viewing resistor inserted into the ground so that you can measure the cable's current without messing the test up too much.

By varying the end load at a constant frequency, you can see the delay caused by the line slowly (relatively speaking) filling up due to transit delays and end reflections. This will show the minima which occurs when line equals load, and the cusp behavior on either side of match. Trying to see the one transit delay at match is doomed from the start, unless you have differential ins. I wrestled with that for a while, and had to settle with looking at the current at the amp, knowing that it was at most, one transit delay removed from the load current.

Using a square wave would certainly show the settling time of the cable/load system, but yes, the information while interesting, might not be very applicable.

I'll try to find the pics elsewhere in the meantime..

Bingo...found it..



jn
Attached Images
File Type: jpg cable test 21.jpg (43.5 KB, 212 views)

Last edited by jneutron; 3rd July 2013 at 06:05 PM.
  Reply With Quote
Old 3rd July 2013, 06:11 PM   #1052
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
a PC soundcard would be fine for 500 Hz, 1 us phase resolution
  Reply With Quote
Old 3rd July 2013, 06:29 PM   #1053
diyAudio Member
 
jneutron's Avatar
 
Join Date: Jan 2004
Location: away
Quote:
Originally Posted by jcx View Post
a PC soundcard would be fine for 500 Hz, 1 us phase resolution
I'm afraid single digit resolution isn't good enough. I suspect 10 or 100 nSec resolution would be best.

Cyril demonstrated 10 uSec at 10Khz with a 4.9 meter cable and a VSWR of 2.2:1 at that frequency.

Not everybody has an rf bridge available... and the results as he demonstrated, don't exactly show us what is occurring. Yes, there is a reflection of the 10 KHz signal, yes it's 50 times slower than the cable prop delay, but how does that impact what we hear?

My test is designed to see the actual load current albeit one prop delay away, to be able to later test actual speaker loads. My interest lies in being able to spot the impedance changes caused by eddies of the coil as well as eddies caused by the coil movement in the field. Velocity dependence of the coil's effective resistance (Rs).

The results are to determine the line to load match boundaries necessary to limit group delay of all frequencies during actual music.

jn
  Reply With Quote
Old 3rd July 2013, 06:45 PM   #1054
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
for a crude heruistic - a second gets you ~1000 zero crossing to average for sub sample time resolution of 500 Hz sine phase

you can use extended observation time, high dynamic range to obtain extremely high time resoluiton

FFT does exactly the right thing in calulating the phase with the complex value
  Reply With Quote
Old 3rd July 2013, 06:51 PM   #1055
diyAudio Member
 
jneutron's Avatar
 
Join Date: Jan 2004
Location: away
Quote:
Originally Posted by jcx View Post
for a crude heruistic - a second gets you ~1000 zero crossing to average for sub sample time resolution of 500 Hz sine phase

you can use extended observation time, high dynamic range to obtain extremely high time resoluiton

FFT does exactly the right thing in calulating the phase with the complex value
Sounds good. But it's beyond my abilities. I'm a simple country doctor, Jim..

It also allows the storage of say, a 50 hz response for later subtraction of a two tone signal used to measure velocity dependence of a speaker's Rs.

I really do like your suggestion..

jn
  Reply With Quote
Old 3rd July 2013, 07:13 PM   #1056
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
its really not as hard as improvising a laser to cut your way out of a jail cell...


Audacity lets you play your test waveforms and record, if you don't like any math SW then LTspice imports .wav, can cruise the FFT with the cursors to get the relative amplitude, phases of any 2 bins

some soundcard freeware like Visual Analyzer may have the capability too - some already know how to calculate complex impedance


...just don't step into the transporter wearing a Red Shirt
  Reply With Quote
Old 3rd July 2013, 07:20 PM   #1057
diyAudio Member
 
jneutron's Avatar
 
Join Date: Jan 2004
Location: away
Quote:
Originally Posted by jcx View Post
its really not as hard as improvising a laser to cut your way out of a jail cell...
You do test my memory. Don't I need a communicator for that??

Quote:
Originally Posted by jcx View Post
Audacity lets you play your test waveforms and record, if you don't like any math SW then LTspice imports .wav, can cruise the FFT with the cursors to get the relative amplitude, phases of any 2 bins


some soundcard freeware like Visual Analyzer may have the capability too - some already know how to calculate complex impedance
I'd really just like to see the amplitudes. if I want to see imaginary, I'll buy a Q chip..

Quote:
Originally Posted by jcx View Post
...just don't step into the transporter wearing a Red Shirt
They're always the ones to bite the big one, aren't they?

jn
  Reply With Quote
Old 4th July 2013, 08:43 AM   #1058
Elvee is offline Elvee  Belgium
diyAudio Member
 
Elvee's Avatar
 
Join Date: Sep 2006
Quote:
Originally Posted by jneutron View Post
As I said, we're speaking about short speaker wire lengths terminated terribly.
I am perfectly aware of that. This was just to answer your question about the 45° phase missing at low frequencies.
Note that having a complex and variable characteristic impedance is not without consequences, even for short lengths: these effects will alter the impedance seen by the amplifier


Quote:
When I said 200 segments, I meant 200 to model the length given, in Cyril's case it was 4.9 meters.
Yes, the 200 number I took is just a coincidence.
Anyway, as I said in another discussion, modelling real lines with lumped elements, even in very large numbers is a poor and crude method.
For physical simulation, it might still be acceptable but for soft simulation there are better options

Quote:
Ah, 24 guage, 40 meters long. Not exactly appropriate for the discussion, eh?
It's important to model what is being discussed rather than trying to shoehorn the available model to different circumstances. You're losing HALF the power to the line, certainly not reality..
Not appropriate if we want to discuss specific numerical results, but perfectly adequate to illustrate theoretical concepts like the variation of characteristic impedance with frequency.
My point is that the usual transmission line concepts like matching, etc, are essentially useless at audio frequencies.
If you build a 8Ω speaker cable, it will stay at that value down to a few KHz, at the very best, and go all over the place below: matching makes no sense, you would need to present a load having a phase of 45° over many octaves, together with a variable magnitude, which is completely unrealistic
Quote:
Despite all this discussion, what is being bypassed it the fact that ACTUAL hard measurements of load to line discontinuity caused reflections has been demonstrated despite statements that the line is too short at these frequencies.
Transmission line effects are completely agnostic, and manifest themselves wheter the line is 1,000λ or 0.001λ. The underlying theory remains the same.
With a good VNA working in virtual time-domain, you can see incredibly small details.
Quote:
Originally Posted by Bob Cordell View Post
Good work, Elvee. However, it would be even more interesting if you plotted it out to 10MHz. A lot of the fun does not begin until beyond 100kHz.

With plots to 10MHz, it would also be interesting to see these three situations depicted for the conditions of the cable terminated in 100 ohms, the cable terminated into a short, and the cable completely unterminated.
Here you are. I have also included a 20Ω load, partly to adress John's criticism about the lack of suitability of the 0.5mm model, and partly to get a more realistic view of the problem: actual speaker cables normally have an impedance lower than 100Ω, and with 20Ω, the Zo to Zl ratio is better respected.
The 100Ω case is duplicated, to zoom on the finer details.


This discussion can be of interest too:
Modelling Audio Cables
Attached Images
File Type: png 40mComp3.png (107.8 KB, 203 views)
File Type: png 40mComp4.png (106.9 KB, 200 views)
File Type: png 40mComp5.png (100.7 KB, 201 views)
File Type: png 40mComp6.png (94.4 KB, 186 views)
File Type: png 40mComp7.png (103.9 KB, 22 views)
File Type: png 40mComp7z.png (105.6 KB, 17 views)
File Type: png 40mComp5z.png (104.2 KB, 22 views)
__________________
. .Circlophone your life !!!! . .
♫♪ My little cheap Circlophone© ♫♪

Last edited by Elvee; 4th July 2013 at 08:47 AM.
  Reply With Quote
Old 4th July 2013, 10:58 AM   #1059
diyAudio Member
 
Join Date: Sep 2006
Quote:
Originally Posted by Elvee View Post
I am perfectly aware of that. This was just to answer your question about the 45° phase missing at low frequencies.
Note that having a complex and variable characteristic impedance is not without consequences, even for short lengths: these effects will alter the impedance seen by the amplifier



Yes, the 200 number I took is just a coincidence.
Anyway, as I said in another discussion, modelling real lines with lumped elements, even in very large numbers is a poor and crude method.
For physical simulation, it might still be acceptable but for soft simulation there are better options


Not appropriate if we want to discuss specific numerical results, but perfectly adequate to illustrate theoretical concepts like the variation of characteristic impedance with frequency.
My point is that the usual transmission line concepts like matching, etc, are essentially useless at audio frequencies.
If you build a 8Ω speaker cable, it will stay at that value down to a few KHz, at the very best, and go all over the place below: matching makes no sense, you would need to present a load having a phase of 45° over many octaves, together with a variable magnitude, which is completely unrealistic

Transmission line effects are completely agnostic, and manifest themselves wheter the line is 1,000λ or 0.001λ. The underlying theory remains the same.
With a good VNA working in virtual time-domain, you can see incredibly small details.

Here you are. I have also included a 20Ω load, partly to adress John's criticism about the lack of suitability of the 0.5mm model, and partly to get a more realistic view of the problem: actual speaker cables normally have an impedance lower than 100Ω, and with 20Ω, the Zo to Zl ratio is better respected.
The 100Ω case is duplicated, to zoom on the finer details.


This discussion can be of interest too:
Modelling Audio Cables
Hi Elvee,

Thank you very much!

It may take me awhile to get this all looked at as we are celebrating Independence Day here in the US.

Cheers,
Bob
  Reply With Quote
Old 7th July 2013, 12:07 PM   #1060
qusp is offline qusp  Australia
diyAudio Member
 
qusp's Avatar
 
Join Date: Oct 2009
Location: Brisbane, Australia
Elvee, jcx, Jneutron et' al; thanks so much, plenty to chew on,
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
myths and facts on electrolytic capacitors in audio circuits krachkiste Solid State 92 8th July 2012 01:10 PM
25 pair cat 5 cable for speaker cable? mpmarino Parts 19 3rd March 2006 04:46 PM


New To Site? Need Help?

All times are GMT. The time now is 10:54 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2