characteristic cable impedance

[AndrewT]

Is there a simple DIY method of measuring/checking the characteristic impedance of a twisted pair?
What equipment is needed?

Can the method be applied to find the impedance of any cable?

[sreten]

Hi,

In its simplest form Zo = sqrt ( L / C ) where L and C are per unit length.

The above does not take into account resistive / conductive losses,
consequently it is of little value at audio frequencies / long cables.

Telecoms uses ~ 600 ohms for audio frequencies whilst the actual
characteristic impedance at high frequencies will be nearer 100.

As far as I know it is far easier to model / calculate / estimate than
actually measure. There are lots of data standards for twisted pair
interfaces that yield values in the range 50 ohm to 150 ohm.

Spaced twin antenna lead (think cheap FM aerials) takes it up to 300 ohm.

FWIW most phono plugs mismatch the cable they are used to connect,
yeilding the idea of characteristic impedance a rather moot point.

/sreten.

[wakibaki]

http://www.sigcon.com/lib/htm/TWIST.htm

The simplest way of determining the impedance of a cable is with a network analyzer, effectively by measuring the reflected wave at the discontinuity with a known cable. You might be able to terminate the cable with a known resistor and use a ham radio SWR meter.

Cablle impedances are irrelevant where the cable is electrically 'short', i.e. less than a tenth of a wavelength.

w

[janneman]

Originally posted by sreten
[snip]FWIW most phono plugs mismatch the cable they are used to connect,
yeilding the idea of characteristic impedance a rather moot point.

/sreten.

Sreten,

What would be a good XLR or RCA plug for digital audio interface, like AES/EBU or S/PDIF? Are there any connectors that preserve the characteristic impedance?

Jan Didden

[Ouroboros]

You can use a square-wave generator, a 'scope, a resistor and and a 200 Ohm variable-resistor to make a simple time-domain-reflectometer.

Simple drive a length of 15 metres or so of the cable through a 100 Ohm (or there-abouts) resistor from the square-wave generator. Use the 'scope to monitor across the line after the series resistor. If you do this with a 1MHz square wave, you should see, at about 150nsec after the driving edge, a small reflection from the far end. If you now try it with the far end of the cable shorted, you should see the reflection alter in polarity.

Use the variable resistor to terminate the far end of the cable, and as you vary the resistance, there will be a value where the amplitude of the reflection is at a minimum. This resistance value is close to the characteristic impedance of the cable.

[wakibaki]

Ah, this is a good trick, if you can get it to work. As I remember we had to build a dedicated fast rise-time pulse generator as most of the equipment in the lab didn't hack it.

w

[AndrewT]

Originally posted by Ouroboros
You can use a square-wave generator, a 'scope, a resistor and and a 200 Ohm variable-resistor to make a simple time-domain-reflectometer.

Simple drive a length of 15 metres or so of the cable through a 100 Ohm (or there-abouts) resistor from the square-wave generator. Use the 'scope to monitor across the line after the series resistor. If you do this with a 1MHz square wave, you should see, at about 150nsec after the driving edge, a small reflection from the far end. If you now try it with the far end of the cable shorted, you should see the reflection alter in polarity.

Use the variable resistor to terminate the far end of the cable, and as you vary the resistance, there will be a value where the amplitude of the reflection is at a minimum. This resistance value is close to the characteristic impedance of the cable.

this I can manage.
Could an error here be causing the test signal I input into my amps to appear peaked rather than square/squarish?

If I'm sending a mixture of analoque and digital signals around the house over twisted pair varying between 10m and 100m long, do I need to know my characteristic impedance?

Does this impedance change with the varying twist found in CAT5?

[Ouroboros]

CAT5 is specified at 100 Ohms (I think!) and it's fairly well controlled. If you're sending digital signals over it at those lengths, then the cable should be driven and terminated in 100R, otherwise the edge reflections on the data signal may cause bit-errors.

[AndrewT]

Originally posted by Ouroboros
CAT5 is specified at 100 Ohms (I think!)

or is it 110ohms to match PCB trace over ground plane impedance?

Thanks!!

[sreten]

Originally posted by janneman

Sreten,

What would be a good XLR or RCA plug for digital audio interface,
like AES/EBU or S/PDIF? Are there any connectors that preserve
the characteristic impedance?

Jan Didden

Hi,

BNC's (bayonet connectors) aka oscilloscope inputs match
the cable impedance they are designed for (50 or 75 ohm).
XLR's AFAIK are hopeless.
Video phono's with circular crimping (like BNC's) are better than the usual.
(But what about the phono inputs - same issues apply as plugs .....)

For a DAI there are possible tweaks using matching resistors etc
at either end to improve matters but are no fixed rules AFAIAA.

/sreten.