The unlock feature isn't really a hack, it's an actual feature of the ClearCom system, by design.
I'll try to describe what I know about the ClearComs, but really the best explanation I've ever seen, and where I learned most of this, is from this page:
http://www.rcrowley.com/comclone/Project.htm
http://www.rcrowley.com/comclone/CircuitDesc.htm
http://www.rcrowley.com/comclone/default.htm
Basically it's a page on how to construct ClearCom-compatible (he calls them "ComClone") intercoms.
Basically ... it works something like this. The ClearComs use three-pin XLR (balanced audio microphone cable) as a physical medium. There is one master box, which plugs into the wall, and then there are many portable beltpacks, which you daisy-chain off of the master. In terms of topology it's kinda similar to old coax-based ethernet, only in addition to the shared-medium data line, you also have a Vcc and Gnd wire.
But instead of coax, you're using balanced audio cable, so you have two signal wires and then a shield wire running around them. One of the signal wires is used in the clearcom setup for power, another is used for (unbalanced!) audio, and then ground is used for a shared power/signal ground.
The master box feeds DC onto the power wire, and this is how all the devices on the system get power -- this way the belt packs don't need batteries. I think it's like 24VDC or so. The master box also terminates the audio signal wire, with some fixed resistance. I think it's like 600 ohms or something (don't quote me on that, though). And it grounds the third (ground) pin.
Each beltpack transmits audio onto the signal wire, by acting as a variable AC CURRENT source (not a variable voltage source, as you might suspect). Remember that the audio line is terminated at one point, back at the master console. So V = IR, with a fixed R (the termination resistor), means there's a fluctuating voltage signal.
In order to receive audio, each beltpack acts as some very very high resistance in between the audio signal and ground, and basically measures the voltage change. Since each beltpack acts as some really high impedance (up in the megohm range, I think), and the audio is transmitted as a fluctuating current through the terminating resistor, which is much, much smaller than the internal impedance of the beltpack receivers, you can put a lot of beltpacks on a circuit without diminishing the audio signal. The audio is basically right around "line level" (few hundred mV).
Also, and this is fairly important -- all the audio parts of the transmitters and receivers (which I think are opamps) are AC coupled; they're isolated with capacitors. This is important, because the system imposes a DC bias on the audio wire in order to send signals.
There are two types of control signals that the system allows for. One is the "attention" signal, which makes a light flash on the beltpacks, so that you can wake up someone who might have their headset off and get them to come on line. The other signal is the "hangup" signal, which causes the PTT switch on the remote stations to release.
Both of these functions can only be initiated from the master console (the one with the power supply in it). Basically, when you want to send 'attention,' you press a button, and the box imposes a DC bias equal to about half of the supply voltage on the signal line. The belt packs have a bright LED that goes on in response to this. I don't know exactly how it's triggered but there are a lot of ways you could do it (zeners, etc.)
For the other signal, the hangup signal, there's an even higher DC bias imposed, I think. (Maybe just Vcc, assuming that the attention signal is Vcc/2?) It could conceivably be a negative DC bias with respect to ground, or something else (I've never actually measured it), but it's some other kind of DC bias on the signal line.
The beltpacks all have PTT switches on them that are non-mechanical. When you press and hold one, it works like a PTT. When you press it twice, quickly, it "locks" and you can talk without holding it down. They are designed so that if the person at the master console presses the unlock button, the belt packs will unlock the PTT in response to the signal. Honestly I'm not sure exactly how the beltpacks accomplish this, since I've never reverse-engineered one; I'm pretty sure though that the home-made ComClones *won't* do it, so I think it's a fairly complex analog circuit. (The easiest way would be with a latching relay, but I'm pretty sure that this is not how the beltpacks work, I think it's all solid-state.)
Since the audio signal and these DC signaling pulses are on the same wire, whenever the person at the master console uses one of these features -- attention or unlock -- you can hear it in the headset as a "clunking" sound.
That's about all there is to them. There are a few competing designs for simple party-line intercoms to ClearCom's; Telex is the biggest alternative, and I think they may do something that allows for balanced audio (the ClearComs will hum if you get them too close to a power line, which is a problem in the theater where you're using them alongside horrifically noisy SCR-based lighting dimmers) but they're essentially the same idea.
They also make two-channel versions that use 4-pin XLR cable, and basically just have two signal lines, so you can have two "subnets" (say, you can put all your backstage crew on channel A, and all your front-of-house crew on chan B, so the FOH people don't hear the backstage chatter if they don't want to, but the stage manager, sitting at the master console, can talk to everyone or even bridge the two groups if he/she wants).
More modern systems made in the last 5-10 years are digital and/or allow for multiple channels on top of each other by using frequency modulation techniques; wireless ones are also big. However, the 2-wire (plus ground) ClearCom system is the de facto standard in many theaters and production facilities, and in many cases the buildings have been wired for them (plus you can run them through unused channels in XLR "snakes", etc.).
Anyway hope this made sense. I'll probably copy this email and put it in my /. journal, and perhaps some other knowledgeable folks will correct any mistakes I've made.
--
The body of the above message, excepting material quoted or reproduced from other sources, and specifically excluding any and all attachments unless specifically noted, is licensed under the GNU Free Documentation License Version 1.2, with no Invariant Sections, no Front-Cover Texts an no Back-Cover Texts, and may be copied, distributed and/or modified subject to the terms of the License at http://www.gnu.org/licenses/fdl.txt