Cat5e Line Level Standards?

[zigzagflux]

Originally posted by infinia
Why? I think source resistors are quite useful, esp. using op-amps to drive reactive loads like twisted pair or even coax. You have to be extra careful using zero ohm sources to drive audio unless it's designed to.

You are talking about two completely different issues. It is true, op amps often need some added source resistance, but it has nothing to do with transmission line affects. With an op amp, you are dealing with a very low ohm source impedance as a result of negative feedback wrapped around a gainblock of many orders of magnitude. When driving a capacitive load, the amplifier can become unstable; padding with some resistance will mitigate the tendency to a degree. Even then, you are rarely talking about the use of 75 ohm resistors; 1 or 10 ohms is typically sufficient. Further, the selection of this resistance has no bearing whatsover on the characteristic impedance of the line, nor on the terminating resistance at the load. Apples and oranges.

Granted you have a 3 dB power loss but at line level it's not a huge concern.

Sure it's a concern. At line levels, power loss should be nearly zero. The goal of a line level source is to present as much of the signal, uncompromised, to the load. Once at the receiving end, it will typically receive voltage amplication, followed by an impedance conversion (in simplistic form power amplification). No power, ideally, is transmitted over this line. If you would opt for 75 ohm source and terminating resistors, you NEEDLESSLY dissipate power, AND lose 50% of your signal.

I see nothing wrong with using twisted pair for line level audio. In my mind its the right way.

I agree, I would use shielded twisted pair rather than coax. But, the souce will have zero (or as I said previously, nearly zero) impedance, and the receiving end should have as high an impedance as possible. 10kohm would be a minimum starting point, higher would be better.

I believe, as a general rule, coax is properly applied when dealing with very high frequencies, waaaay beyond that of audio. In that application, POWER transfer is very important, and the line impedance must be matched with identical source and terminating resistors. You must keep this separate from audio applications, which is a plain old misapplication regardless of the cable construction.

Jordan: sorry if this was an unwelcome divergence from your original question. While shielding provides many benefits, if the cat5 is existing, I am sure you could use it with fairly good results. Stick to the twisted pairs, of course; calee4nyaboy is right on the money with his advice.

[infinia]

Originally posted by zigzagflux




Sure it's a concern (3dB power loss). At line levels, power loss should be nearly zero. The goal of a line level source is to present as much of the signal, uncompromised, to the load. Once at the receiving end, it will typically receive voltage amplication, followed by an impedance conversion (in simplistic form power amplification). No power, ideally, is transmitted over this line. If you would opt for 75 ohm source and terminating resistors, you NEEDLESSLY dissipate power, AND lose 50% of your signal.



I agree, I would use shielded twisted pair rather than coax. But, the souce will have zero (or as I said previously, nearly zero) impedance, and the receiving end should have as high an impedance as possible. 10kohm would be a minimum starting point, higher would be better.

You never explained why your goal is to minimize power loss at line levels. This is the heart of our disagreement. You say to preserve the signal but we are talking about driving a 60-100 ft of small gauge wires with no shields. This is essentially the problem of telephony and it's a hard learned path that is well worn.
Since we both agree most op-amps do not like driving reactive loads. We have to add some small series output resistance, so it's not really an ideal voltage source. Now then it's not breaking the rules to simply double your 10-20 ohms to get in the neighborhood of Zo. Using 50-100 ohms or so to buffer the output of any op-amp that goes off the PCB is generally good practice anyway. (look at Linkwitz circuit designs) Notice also it's not a big deal of adjusting the op-amp gains at either end slightly higher to make up for any losses in the transmission line. Some really good op-amps like a little extra gain anyway . Now at the receiving end as you lower the impedance of the termination does three things in order of concern here 1) extends to the cutoff frequency of the Cat-5 cable 2) lowers the input noise source at the receiving amp. 3) swamps out signal reflections that will degrade the signal quality.

Another concern not even hinted at in this thread is channel cross talk. Just saying it's rather dogmatic to simply preach "use zero ohm sources and high impedance terminations".

[zigzagflux]

There is nothing dogmatic about speaking the simple truth in an attempt to dispel incorrect information. This will be my last post in this thread, and you friend have much to learn:

http://www.sound.westhost.com/cable-z.htm

Suffice it to say, after being enlightened by your posts, I will now buy a bunch of 4 ohm resistors. I will then proceed to install said resistor right at the output of my amplifier, in order to properly drive my speaker cable "transmission line". I will gain tremendous benefits of 1) extended frequency response 2) lower noise 3) no more signal reflections

[infinia]

We are talking about long wire runs at line level right! Not for speaker cables (power electronics).
Since this was your last post I guess you won't/can't explain your rule.

definition of dogma.
1 a: something held as an established opinion ; especially : a definite authoritative tenet b: a code of such tenets <pedagogical dogma> c: a point of view or tenet put forth as authoritative without adequate grounds

[paulb]

We're using CAT5 here at work to distribute audio, and I have a friend who is doing the same thing at home (at my home I'm using individually shielded twisted pairs). It works very well, although we haven't done any measurements.
I'm not sure whether you want to run the sense lines down the cable - it seems to me you could get problems with the sense lines picking up interference. I'd suggest allowing for both local and remote sense and try both.
As to the source/receiver impedance issue: I don't see how lowering the receiver impedance would extend the cable's frequency response, and that I very much doubt reflections are an issue at these frequencies. I'm using the TI drivers and receivers and they don't mention anything about matching impedance. See Fig. 2 in
http://focus.ti.com/lit/ds/symlink/drv134.pdf

[infinia]

Originally posted by paulb

As to the source/receiver impedance issue: I don't see how lowering the receiver impedance would extend the cable's frequency response, and that I very much doubt reflections are an issue at these frequencies. I'm using the TI drivers and receivers and they don't mention anything about matching impedance. See Fig. 2 in
http://focus.ti.com/lit/ds/symlink/drv134.pdf

They use 600 ohms for Zo. Maybe they dont talk about it, but its there (built in those special purpose devices above). BTW that is a good source for application of twisted wires but they want you to buy thier stuff.

U can use good general purpose audio op-amps to do something similar but the terminating imp. (lower is better) will be related to the max load and THD of the devices you pick. edit> most likely on the order 1-2K ohms or so.

Reflections are an audio problem as in telephones but not severe for short runs like this. that's why it was listed last or #3.