My local police department dispatcher position has migrated from an aging rat's-nest connecting three older radios (all different Motorolas) to four new Kenwoods, all alike. This enables them to go narrowband when that becomes unavoidable, and also moves them from a cramped corner into a new room. I was given the task of mounting the radios into a cabinet built by someone else (with some suggestions from me), and replacing the rat's nest with something a little more organized. They also wanted a second control position for occasional use. Since the radios are in the same room, the radio speakers did not need to be remoted; both positions are able to hear them well. The primary operator, sitting right by the radios, would have easy access for frequency changes, although only the 4th radio would normally be subject to change; the first three are dedicated to Police, Fire, and State Net. So the main needs were for combined PTT points and a mike for each dispatcher, connections to the paging and siren encoders at the primary position, and balanced lines from each radio to a multichannel logging recorder, mixing the transmit and receive audio. The result is shown at right.
The next photo shows the central logic box. The gooseneck mike for the main position plugs in below the label "DYN MIKE." The secondary position uses a common computer style desk mike (shown in photo 1) that plugs into the desktop box. The main position can also accept the same kind of mike, but they wanted to keep the much more substantial mike they are using. Power is supplied from an IBM brick that I had on hand, which provides ±12v and +5v. The main paging encoder can be switched from radio 2 to radio 1, a safety backup to match the existing arrangement. The weather siren encoder is wired through to radio 4, since it uses a different frequency. This eliminates changing the channel on the fire radio to send those codes, as they did before.
The Kenwoods had to be modified to be compatible with the electret-level mike output. (Kenwoods use a dynamic mike input, and also accept a line-level "data" input: one is too low [5mv], the other too high [.2v pp].) Joining the "data" input with the "TXAFI" input (they go to the same IC pin through different resistors) gives a sensitivity of about .12v pp, which worked well. Unfortunately, Kenwood also does not provide a SQUELCHED receive audio output, so I mounted a 2N7000 mosfet switch in each radio, sent the unsquelched DETO (Detector Out) through a resistor to an available accessory pin, and let the transistor shunt-switch that pin, using the speaker mute line to control the transistor gate. I have years of experience with Motorola radios, which use electret mikes and which do provide squelched Rx audio to the accessory connector. So this system should be actually easier to connect to Motorola radios, as long as they maintain that standard. (addendum: it turns out that DETO is pre-unscramble, so when they switched over, the recorder was recording scrambled audio, not very useful. So I removed the mosfet switch and took Rx audio from the speaker output. This has three advantages: it is squelched, it is de-emphasized, so it sounds more natural, and it is unscrambled. But it is also subject to the volume control setting, fortunately not normally an issue.)
The PTT desktop units are actually quite simple. The remote unit connects using a standard ethernet patch cord, and simply carries four +5v PTT levels (when the buttons are pushed) and the mike audio. The main PTT unit is connected with a common DB15HD VGA patch cord, as it carries four more +5v lines that serve to give it dominance. That is, if the remote unit has a radio keyed, the main operator can key the same radio and take over; the radio is switched to the main mike audio. Or if the main unit has a radio keyed, the remote cannot break in. Whenever any button is pushed, an LED indicator lights above that button on both units, so both operators can see which radio is in use.
The last photo shows the system in place (on the right). The power supply fits easily inside the cabinet, under the radios. A small fan circulates air for cooling. They had planned to use the blank face area for equipment but instead put it under the window. They have since posted many more notices on the cabinet slope, so it will be a billboard.
A new design (in the breadboard stage of development) uses interlock. That is, when a button is keyed on either box, the corresponding LED on the other box lights red, which says "this key is in use and is blocked." The rule here is "first one to key keeps the line." This redesign also makes the boxes identical inside, and they both use ethernet cables. This also means that the original design's master-slave relationship could have been implemented without the four extra lines, so the master's cable can be reduced to 8 conductors. The "First One Wins" (F1W) relationship uses different line levels to distinguish one unit from the other: less than 2 volts tells it to light the red and cut the switch, while more than 2 volts tells it to light the green and key the radio. (Zero volts says do nothing.) I'm sure some whiz could design this with microprocessors, but I did not want to learn PICBasic just to do this. And a blown µchip would be much harder to replace if a way to flash the program back in was not right at hand.
An even simpler design is percolating, and will eventually be built. It is master-slave, not F1W, so it will superficially be identical to the unit in service, except that the key boxes will be interchangeable. Also the recorder buffers will work off single-supply +12, so that power can be drawn from the radios, eliminating the brick.
In the logic box, the mike audio is combined with the PTT line as is standard for Motorola radios, which is another reason this system is Motorola-compatible. After switching, they are then split back apart for the Kenwoods. To attach a Motorola, just connect to the unsplit line. No significant modification required. As for the logging recorder lines, these are basic op-amp buffers with mixer inputs and balanced outputs (hence the ±12v power), sent out to the recorder on a single CAT5 cable. Unfortunately, the Kenwoods threw me one more time when I found that the "TXAFO" line they provide is tapped off before the data/TXAFI injection point, so it carries only the front-panel mike audio, not the accessory input modulation! Since I wanted to record Tx audio both ways, with the hand mike plugged in and not (as a safety redundancy, should the PTT box go down), I had to provide the mix buffers three inputs: RXAFO, TXAFO, and TXAFI. Since all are trimmable for level, this led to a pile of trimpots (12, to be exact). Fortunately, the VGA cables to the radios could support all this. It seems that many VGA cables short all the VGA grounds together (and of course, most vendors do not tell you if this is true of theirs or not), but as it worked out, I did not have to custom-build the cables for the radios.
The cables unfortunately short to ground three of the accessory lines Kenwood offers for undedicated options, which meant that my plans for automatic channel activity monitor had to be shelved. This would have provided a brief "unhook" signal to the radio before PTT, so if uncoded traffic was present, the dispatcher would hear it before commencing to transmit. Should that protocol become mandatory, VGA cables without the ground shorts would be required (if such exist) or custom-made cables made (six paths occupied). The simpler designs mentioned above would use custom DB15HD-to-RJ45 cables (no DB15HD's on the logic box at all) so this function will also then be available.
The radios are mounted in a line on a bar. The lid of the cabinet can be removed and the four-radio assembly lifted out as one unit without disconnecting them. The cabinet supports the bar at each end and in the center. This rapid-access construction has already proved very helpful.
More to come in time...
Created on ... July 31, 2011