Big intercom refurbishment

    First fixups

[Images are linked to larger copies.]

One of the major problems was with switches and volume controls on the beltpacks.  The pots had gotten scratchy enough that sometimes received audio would cut out entirely -- and if the line was quiet, the user wouldn't realize it until finding out that they simply hadn't heard something.  The microphone switches were in similarly rough shape.  All of this was easy enough to fix with a shot of DeOxit and working the movement range several times, but every pack needed to be opened up and the volume knobs removed to access the right bits.

The cleanup process had produce quite a pile of detritus -- old tape, pieces of torn earmuffs, cracked plastic from some of the headbands.  Tape goo had to be removed here and there, for which mineral-spirits paint thinner seems to work best -- it doesn't dissolve the stickiness, just softens it long enough to transfer off to a paper towel.  Fortunately the kit included a few spare earpiece pads so I didn't have to buy any more, but learned that they're a little tricky to remove and get back onto the "210" style headsets.

Eventually I could test most of the system all together at once, and start testing all the hookup cables as part of that.  I call this the "money shot", as it represents most of the active inventory other than the rest of the cable.  The key piece here was an MP3 player hooked to the audio line sending in a low 30 Hz test tone, which is the best way to then go through and wiggle things and turn volume pots and listen for any hint of scratchiness on top of that.  With many headsets hooked up and all with their "signal" lights on, I could also do overall current-drain tests and check for any power supply droop or hum.  [It always helps when main filter capacitors don't *fall completely out* of their circuit boards, which had happened inside one of the supplies!]

Each station draws less than 15 mA quiescent, maybe a little more if squawking loudly, and about 30 mA with the signal light lit.  Thus, a power supply only has to be solidly good for about half an amp to drive a fairly large system.  Mind you, that's still a fair amount of current to be shooting down the equivalent of microphone cable. In the background is also the six-pack of new color-coordinated attached-lid crates I ordered to store and ship the gear in.

To distinguish these crates from green crates used by some others in the community, I decided to put a stripe on the ones that would go afield in a way that could still be seen when the lids were flopped open.  This involved testing various different types of black spraypaint I had kicking around, to see which would adhere best.  Ultimately the best step was to lightly sand the strip where the paint would go before spraying, a lesson similarly taken from doing the hood stripe on the car. And there's a definite thematic color match here, as the system's major hallmark has always been this bright green cable that's instantly identifiable.  In the process of re-inventorying I eliminated a few bits of ordinary black XLR, staying with a 100% green theme for the remainder.  That cable is also known to NOT have any stray connections between shield and XLR shell, which a random audio cable very well may have and has been a source of problems in the past.

Once the procedure for masking and striping was solidified, the rest got done in bulk.  Here are some initially drying.

Next I investigated the two-channel headend we usually use, and found that its "link" switch appeared to be a somewhat home-hacked mod with some fairly sloppy soldering.  The previous owner didn't know the exact provenance of this, but I fixed it up to be a little more robust anyway.

    Fleshing it out

The next phase was to make more cables and some Y-adapters.  The inventory included some very long [and green] cables wired in some sort of multi-channel bus, presumably for long shots across a stadium-size venue, but our usual gigs don't have much use for such things and could be handled with a few regular hundred-footers.  I decided to cut one of these up to make a batch of 25-foot cables, as there weren't any and it's a convenient length to hook up stations that aren't quite right next to each other but too close for a 50 to make sense.  As I was heading off for my usual holiday-time travels, I decided to take the whole batch with me as piecework -- the cable stock, whatever XLR connectors I could scrounge up, labels, and tools.  The fact that stations can daisy-chain together tends to make people think that they *have* to daisy-chain, like DMX, leading to some fairly horrific rats-nests of wiring, but that is simply not the case with this stuff.  Y-splitters are just fine, because it's a low-bandwidth shared analog audio bus.  More splitters in system construction actually make it *easier* to debug, as whole chunks can be isolated away by disconnecting a Y at a known point instead of running around trying to figure out which packs are downstream of which others.  A more parallel layout structure also minimizes the number of stations which could go unterminated and start feeding back when someone makes connection changes, as only units downstream of an audio-line break are affected. I made the splitters with mid-line splices, rather than trying jam it all into one of the female heads.  This made for more reliable single-lead connectors and allowed more room to spread out and insulate the interconnections away from each other.  It turned out that making Ys with the female and males all on one side of the splice, like tripod legs, was a little easier than trying to make them linear like a fork in the road.  I also planned to add a couple of "bud-box" types later on, using panel-mount connectors and possibly even throwing in more new-channel switches.  A small system doesn't need a fancy two-channel headend to have two independent lines, just a new terminator dropped in someplace.

The long cable turned out to be genuine Canare Starquad, with a very tight shielding braid that was rather tedious to pick apart.  And of course the "quad" meant that *two* wires had to be soldered onto each of the power and audio pins.

I worked slowly and carefully on this over some days, including having to remediate some issues with the connectors I'd brought along.

Some of what I'd recovered from the long multi-cables were cheap knockoffs of Neutrik heads, but with female-side lugs that didn't adapt to flexure very much and turned up mis-sized to grab a pin firmly.  They all should look like the one on the left, almost closed until a pin goes in, but I found several sort of spread open a bit like the one on the right.  Use the big-pictures here to see the detail I'm talking about.

The female side should open up just a little when actually on a pin, but not past the elastic limit and certainly not as far as I'd found some of them in the empty state.  But I was short on stock and needed to get some of this done; rather than wait around for new connectors I re-crimped several of these just a little so they'd relax to a more closed state and kept assembling my new cables.  I would simply keep an eye on the knockoffs over time -- it's easy enough to test for proper grab with a naked male connector, one pin at a time, checking that insertion resistance feels solid and doesn't let the pin flop around.

Here's how the real Neutrik connector does it: offset contact points at specific places, and slightly more flexible construction to help ensure that it doesn't get permanently sprung more open by an oversize or crooked pin.