[jhill32]

OK. I understand the definition of impedance. And I know the various permutations of ohm's law (E=I x R, etc.). I won't even quibble about the need to use 75ohm cable for video signals. But what I don't understand is how do you actually measure this value for an unknown cable. What does a cable maufacturer do differently between 50ohm RG-59 and 75ohm RG-59? What is actually different between 50ohm and 75ohm connectors?


I realize that I'm generalizing between the terms impedance and resistance, and maybe that's the whole issue. But, I'm still thinking that this is something I should be able to measure with my ohmmeter, but I can't figure out how. Just something I've always wondered.

 

[brothermaynard]

I have been curious about this too... Perhaps it is a measurement of resistance over 1000 feet or some other arbitrary standard. If the copper conductor was physically larger on the 50 ohm cable, it might be the case.

 

[mark haflich]

Guys. Impedance in the case of cables is AC impedance. Thus a 50 ohm cable or a 75 ohm cable. Most video cables are 75 ohm and that is the reason for the need to terminate them with a 75 ohm connector. This avoids internal reflections from impedance mismatch. For most consumers, any mismatch (using 50 ohm RCAs on a 75 ohm cable instead of a 75 ohm BNC) will be unnoticeable. DC resistance varies with cable length and will be very low. The DC resistance of a cable connector (either 50 or 75 ohm connector) will be essentially nil. You can't measure AC impedance with just an ohm meter.

 

[jhill32]

Quote:

Originally posted by mark haflich Guys. Impedance in the case of cables is AC impedance. .... The DC resistance of a cable connector (either 50 or 75 ohm connector) will be essentially nil. You can't measure AC impedance with just an ohm meter.

OK. That helps a little. I can see that trying to apply DC principles to an AC signal is apples to oranges


I know. I should never have dropped that AC signal analysis class in college. But I just couldn't get polar to rectangular conversions!


But to go back to the original question. How do you actually measure AC impedance? Is it something I can see with my oscilloscope? I mean, there must be something different about the actual cable, right?

 

[Selden Ball]

For relatively high frequencies, an appropriate approximation for impedance is Z = sqrt(L/C)


Belden has a technical paper describing the details. See http://bwcecom.belden.com/college/te.../ciocahalf.htm

The equation above is their equation #5.


In other words, you need to measure the cable's inductance and capacitance. Meters are available from various test and instrumentation companies, like Keithley and Hewlett-Packard. They're usually rather expensive though: typical used prices are over $1K.
http://www.testequipmentdepot.com/us...zers/index.htm shows some typical prices for used HP meters.

 

[jamin]

An additional link is http://www.epanorama.net/documents/w...impedance.html


This also shows how the geometry of the transmission line is involved.

 

[mysphyt]

The impedance of a cable is its "characteristic" impedance. It is the value you would measure, using the above mentioned methods, if you had an infinitely long cable, a cable so long that the signal (at a fixed frequency) never made it to the other end, but was instead dissipated by the impedance of the cable. In other words, if you use the methods above to measure a fixed length of 75 ohm cable, you will measure something considerably less than 75 ohms.


By matching the output impedance, cable and input impedance of a signal transmission circuit, you get maximum power transfer,

 

[jhill32]

Thanks, guys!


Well, as sometimes happens, I didn't really have the right subject. I really was just interested in what properties (manufacturing techniques) affect the impedance of a cable; I didn't actually want to measure it myself.


But the good news is, I got my answer anyway Thanks to the links provided by Selden and jamin I got the following quote that pretty much sums it up:
Quote:

The length has nothing to do with a coaxial cable impedance. Characteristic impedance is determined by the size and spacing of the conductors and the type of dielectric used between them. For ordinary coaxial cable used at reasonable frequency, the characteristic impedance depends on the dimensions of the inner and outer conductors, and on the characteristics of the dielectric material between the inner and outer conductors.

Of course, now when I re-read mark haflich's post, it makes more sense (and sounds a little familiar). It's simply another situation where you can't apply DC characteristics to an AC signal.


Thanks for your patience.

 

[J. L.]

The "impedance" of a coaxial line is determined by the ratio of the diameter of the inner conductor vs. the inside diameter of the outer conductor.


One reference I have gives the formula as:


Zo = 138 log (b/a)


Where b = inside diameter of outer conductor and a = outside diameter of inner conductor and Zo is the impedance.


The velocity factor, (how fast signals travel through it compared to speed through empty space) is determined by the insulating material between the inner conductor and outer conductor.


The "loss" at a given frequency would depend on several factors: the DC resistance of the conductors, the AC resistance of the conductors (typically, this is how well the individual wires making up both the shield and inner conductor act together at a given frequency), the dialectric losses of the insulator (how much it heats up in an RF field), and the losses through radiation (commonly known as "leaky coax" when a braided shield conductor was not 100%).


With this in mind, larger conductors (less resistance) combined with less dialectric material (air-filled foam vs. solid) will result in less loss. RG6 has less loss than RG59, both are 75 ohm impedance as they have the same ratio of inner and outer conductor diameters.

 

[Rob Dingen]

Hello


An easy way to see the impedance of a cable is to apply a square wave signal to a cable of about 10 Meters/30 Feet on the end of the cable a pot from about 100 Ohm and connect it to a oscilloscope

Look at the square wave and adjust the pot until you have no overshoot and it look the most as the original signal

Remove the pot and measure it with a Ohm meter and you have your impedance


Rob

 

[John Moschella]

The characteristic impedance depends on the geometry (ratio of inner and outer conductor) and the dielectric constant of the insulating material. For instance, if the geometry of a teflon cable insulated cable and a polyethylene insulated cable are the same they will have different impedances. The same is true for connectors, its the geometry and dielectric. Because connectors are standarized you can't necessarily make any connector the impedance you want. BNCs were designed for 50 Ohms. You can make 75 Ohm BNC connectors and the difference is that the 75 Ohm version removes the plastic inside the connector itself to increase the impedance. You CAN'T make real 75 Ohm RCA connectors, but you can get close as Canare does with theirs.


Here are the formulas sans the greek characters.


Z = sqrt(L/C) as previously stated.


L = (Uo/2pi)ln(b/a)


C = 2piErEo/ln(b/a)


Z = (1/pi)sqrt(Uo/ErEo) ln(b/a)


Uo = 4pi x 10(-7)


Eo = 8.85 x 10(-12)


pi = 3.14159


Er = dimensionless number, depends on the insulator, air=1


b = inner radius of outer conductor


a = outer radius of inner conductor


Z = 120/sqrt(Er) ln(b/a) in Ohms


EDIT: I corrected a factor of 2 error.

 

[AnonymousCoward]

"Resistance" means DC resistance.

"Impedance" means AC resistance.


Puttung a pot at one end (which can only ever change DC resistance) won't be accurate. It would have to be a capacitive/inductive network. When this is attached to the other end and is adjusted to the generator's and cable's impedance, it will present a perfect AC resistance. All the source signal will be absorbed by the network, and there will be NO standing waves (signal not completely absorbed, so reflected back an forth, reducing the strength of legitimate waves)


Yes, a cable's impedance is determined by its conductor and dilectric sizes, but the composition of dilectric also matters. Most co-ax uses foam, but super-slim cables (RG316) use teflon for dilectric, so have the same impedance as fatter cables, though with less freq response.

 

[John Moschella]

Quote:

Originally posted by AnonymousCoward Puttung a pot at one end (which can only ever change DC resistance) won't be accurate. It would have to be a capacitive/inductive network. When this is attached to the other end and is adjusted to the generator's and cable's impedance, it will present a perfect AC resistance. All the source signal will be absorbed by the network, and there will be NO standing waves (signal not completely absorbed, so reflected back an forth, reducing the strength of legitimate waves)

I'm not exactly sure what you mean, but you can terminate a trasmission line with a matched resistor. It's done all the time and it will eliminate reflections.


John Moschella

 

[frostlich]

Well, this is all fine and good, and I've discovered some rather interesting ways of measuring cable impedance. However, I guess this question was posed from someone seeking a better or the best possible video cable as it applies to HT?


Anyway, just to throw in my 2 cents.....as has been said earlier, any impedance mismatch will probably cause you problems (reflections, VSWR, other than maximum transfer to load, etc).


So, as long as you use the correct Z cable, connectors, etc...you should be fine.


However, I wonder if there would be any advantages to cutting a cable to electrical length so the cable would appear purely resistive(with no inductive or capacitive characteristics) to the source, allowing maximum signal transfer. Probably little to no visible difference as far as video goes, but I've always wondered.

 

[AnonymousCoward]

Quote:

Originally posted by frostlich However, I wonder if there would be any advantages to cutting a cable to electrical length so the cable would appear purely resistive(with no inductive or capacitive characteristics) to the source, allowing maximum signal transfer. Probably little to no visible difference as far as video goes, but I've always wondered.

Impedance is not related to length. It's the relation of capacitance ti inductance in the cable. I'm not sure what you mean by "electrical length", but impedance is determined by cross-section and materials.


However the longer the cable, the more electrical resistance, which degrades the signal, starting with higher frequencies. (rounding a square wave) So for shorter runs, the new materials and thinner cable at a given impedance is fine. But for longer runs, larger diameter at a given impedance is in order.


And only the very cheapest transmission systems would use resistive termination, barring SCSI. I don't know of any examples where pure resistance is used instead of an LC network, at least in a transmission line designed by an engineer. Not sensible technically nor economically.

 

[drewman]

Gosh, I guess all the broadcast equipment that can be terminated by a 75 ohm resistor stuck into a BNC connector was designed by morons!


I worked at television stations for 12 years as a broadcast engineer, looked at many schematics and can tell you that video in broadcast equipment is terminated solely by a 75 ohm (1%) resistor.


Andrew

 

[AnonymousCoward]

Gosh, I guess they were...


Those surely weren't custom LC terminators... (heh)

 

[Morbius]

Quote:

Originally posted by mysphyt The impedance of a cable is its "characteristic" impedance. It is the value you would measure, using the above mentioned methods, if you had an infinitely long cable, a cable so long that the signal (at a fixed frequency) never made it to the other end, but was instead dissipated by the impedance of the cable. In other words, if you use the methods above to measure a fixed length of 75 ohm cable, you will measure something considerably less than 75 ohms.

Mark and Chris are essentially correct.


You have to study a branch of electrical engineering dealing with what are called "transmission lines".


If you had an infinitely long cable - at any point on that cable - the rest of the cable would look like a 75 ohm [ or

whatever the characteristic impedence was] resistor.


That's why you can terminate the cable with a resistor - the cable "thinks" it's the rest of an infinitely long cable.


You could do the same thing with a "Slinky". If you put a Slinky on the floor and hold one end and shake the other end

once - the wave you produce will reflect off the end you are holding still. However, if instead of holding it still -

you hold it with a viscous damper [ like the shock absorbers on your car ] of the right value - the Slinky will "think"

it is an infinitely long Slinky - and you will get no reflection.


Dr. Gregory Greenman

Physicist

 

[bluedevils]

Heh I was going to use a skipping rope, but a slinky is much more fun!


BTW Rod Dingen is the standard way to find the proper resistance to terminate a line.


Quote:

Originally posted by Morbius Mark and Chris are essentially correct. You have to study a branch of electrical engineering dealing with what are called "transmission lines". If you had an infinitely long cable - at any point on that cable - the rest of the cable would look like a 75 ohm [ or whatever the characteristic impedence was] resistor. That's why you can terminate the cable with a resistor - the cable "thinks" it's the rest of an infinitely long cable. You could do the same thing with a "Slinky". If you put a Slinky on the floor and hold one end and shake the other end once - the wave you produce will reflect off the end you are holding still. However, if instead of holding it still - you hold it with a viscous damper [ like the shock absorbers on your car ] of the right value - the Slinky will "think" it is an infinitely long Slinky - and you will get no reflection. Dr. Gregory Greenman Physicist

 

[AnonymousCoward]

OK Morbius, I accept what you say.


But a pure resistor is still not best practice. It can't optimally absorb at the full bandwidth of the cable, as could an LC network.