Building snake cable.

kristian

New member
OK so i have Belden 1222B multipair cable to order. Im going to make 2 snakes. One Balanced(LXRF->XLRM), as far as i see there are only hot and cold no ground so i just use those for hot and cold? i dont get that. As for the unblanced one(1/4"TR->RCA), do i just use one for ground and one for hot. wel thats pretty self explanatory, but i just htought it weird no ground is there. each pair is individually shielded, and then there is a common shield. any ideas?
 
Well, this gets into a discussion of shielding and grounding methodology pretty quickly. Nerdy shit follows... Here's how I'd do it, starting from first principles:

I'm a great believer in single-point ("star") grounding for audio. What this basically means is that a ground reference is established at one single point, and then used for the whole system. For the central point, I always choose the mixing board, since essentially everything talks to it. What I do next may sound like it is more than a little bit at odds with the _letter_ of the National Electrical Code, but please bear with me before throwing those tomatoes...

I ground the chassis of the mixing board to safety ground with a sizeable ground bond (I actually use a woven braid strap about 1/2" by 1/16"). This creates my single ground reference point. I then ground the chassis/safety ground bond point of *each individual piece of gear* to this central ground point with a separate, sizable piece of wire (12AWG is my favorite: it has to be bigger than the power supply lead to act as a proper safety ground). Do not daisy-chain: each piece of gear gets its own completely separate reference run.

I then float the ground connection at the power plug for each piece of equipment with a cheezo 3-2pin power adapter. Yes, that's right: what I am doing is substituting my own safety ground mechanism for the manufacturer's mechanism, so that I can have a bombproof safety ground that is also completely under my control.

What I'm trying to avoid is the creation of any sneak ground loops between pieces of equipment: a current loop from ground at one wall outlet, through the in-wall-wiring to ground at another wall outlet, through the output connector shield, through the input connector shield and then back to the wall outlet again. Depending on how the building is wired, the "ground" available at any arbitrary pair of outlets can differ by many volts. The bottom line is that safety requires grounding, but building wiring is generally not controlled enough to provide decent *audio* performance. So we have to do both.

Once I have the safety/reference ground issue settled, then I can worry about shielding and signal return paths. The thing to keep in mind is that "shield" is a completely different concept than "ground". The purpose of the shield is to collect noise along the length of a cable, and to dump it somewhere other than your signal path. The "somewhere" is the connection between shield and ground located inside the output device.

Having a single ground reference established _separate from the audio signal paths_ allows a very clean setup to be done. The rule of thumb is that the shield connects at the output end, and does _not_ connect at the input end. This is very much like extending the chassis of the box with the outputs right up to the input jack on the other box, but not connecting them. Your separate star ground handles that for you, and *in only a single place*. Thus, no loop. This is called "telescoping the shield".

For your XLR-XLR snake, this means that you would ideally connect the shield to pin 1, hot to 2, and cold to 3 at the output end only. At the input end, you would connect hot to 2 and cold to 3, leaving pin 1 floating. The signal can be thought of as "sent" through the hot, and "returned" through the cold. The ground reference is handled entirely separately, by your new ground run, and no loop is created. This applies to any equipment that has that new ground connection to the center reference of your star. Make sense so far?

The unbalanced setup is more difficult, because the signal return is also necessarily the shield: you only have two conductors to work with. The inability to really isolate shield noise from signal, and isolate signal return from ground, is one reason that the majority of pro equipment is balanced: single-ended stuff requires a lot of skull sweat to really isolate, and in the final analysis you may have to use 1:1 transformers to *completely* isolate and break the loop, and keep the cooties out. But let's not sweat that for the moment...

The best single-ended solution is when you can drive an unbalanced _input_ with a balanced _output_. In this case, you drive tip on the input with hot on the output, and ring on the input with cold on the output. You connect shield at the output end, and leave it floating at the input. And the ground reference is established with your external ground wire: once again, no ground loop, but a good solid signal return path. This works perfectly in most cases.

You have unbalanced-to-unbalanced in your setup. Then you'd connect tip to tip, and connect shield *only at the output*. The signal returns would all be provided by the external ground reference to the center point of your star. Gnarly, since you're returning signal via the ground path. Yucko. In other words: nonoptimal, but functional for _most_ equipment. And still a damned sight better than having a loop through the power supply grounds and the cable shields/returns...

Sometimes, due to the vagaries of equipment design, you end up with problems *anyway*. Then you start having to play with 1:1 matching transformers to get signal-return isolation without coupling ground-to-ground via the shield. That's moderately rare, but it can happen with single-ended gear.

And then there's the phase of the moon. Sometimes your equipment shorts chassis to chassis through the rack mount ears in addition to your ground bonds, giving you a sneak loop there that you weren't expecting. Argh. Sometimes, you tear out a lot of hair getting the last of the crud out of a rig. But going with a star ground scheme can minimize the hair-pulling, and the results are unformly worth it in signal cleanliness... Ain't this fun?

Damn. Too many words, which means that nobody will probably read this. Hope it helps a little bit for anyone who does, though. If you're going to do this, remember that the safety grounds become your responsibility: they *have* to be there, and they have to be *bombproof*. We can still co-opt them for our own purposes, though. They don't have to make the audio suck.

[Edited by skippy on 10-31-2000 at 10:46]
 
Back
Top