Full duplex wireless intercom

I had a succesful first test of one-way communication for the RF900DV modules from Laipac. This is the beginning of version 3 of my bicycle intercom. Since I've never posted about this before, a little introduction is in order:

About 3 years ago my sister and I became very active in road cycling, specifically Ultra-Cycling events such as 24 hour races, 500 mile races, and the penultimate Race Across America - a 9 day, 3000 mile race across the USA from the Pacific ocean to the Atlantic ocean.

Training and racing with another person is a lot of fun, but if you can't chat easily while cycling then it's only slightly better than riding alone. For safety we generally don't ride side by side, and shouting gets tiring.

Most of the typical solutions do not lend themselves to easy communication, limiting the users to half duplex conversations.

My first attempt to alleviate this problem (version one) was to take a cordless phone set, a small corded headset compatible phone, and hook the two together with a 12V gel cell. As many sites on the internet can show you, you can form a basic two phone circuit with a resistor and voltage source.

Sarah would take the cordless handset and a headset. I had the gel cell, base station, and headset telephone in a shoebox on the rear rack of my bike. Aside from the weight, size, and being physically attached to my bike, this was a good solution. It would generally last for 4-6 hours at a time, and since it used a generic 12V gel cell I could extend my end by attaching a 12V solar cell to the top of the shoe box.

Once the concept was proven, I determined to improve it (version 2) by removing the circuitboard from the base station and placing it and a smaller battery into a smaller, more portable space:
V2.jpg: VHS case with base station circuit board, 9.6V RC battery pack, wired and wireless handsets

The VHS cassette case fits the circuit board and battery very well. I removed the power jack, cut a few traces on the PCB, and added a resistor so one could plug any phone into the jack at the top - useful with a speakerphone in a following vehicle for instance. The two bolts on the right of the case are to synchronize the handset with the base station. Technically they also charge the handset, but I instead opted to charge the batteries with another charger and simply carry as many as I needed pre-charged for any given event.

The VHS case fits (snuggly) into the back pocket of all the cycling jerseys I've tried it in. This was smaller, lighter, and didn't require me to remember to disconnect the headset before dismounting. It is still very large and heavy considering what it does, and was a power hog. We could generally count on 2 hours of good communications, maybe 4 depending on how well the batteries were managed previously. We would end up carrying extra batteries.

As you can see in version 2, there is a small silvered radio module in the base station. There is a similar one in the handset. These are the transceivers. I had disassembled another phone and started probing the connections to determine how I could use the modules without the remaining support circuitry. I determined that they contained PLLs that were controlled with a generic clocked serial protocol. I stopped that project when, after purchasing another phone (well, several) discovered that the various transceiver modules are quite different and I would likely need to reverse-engineer a new module each time I wanted to create another intercom.

About this time I found that Laipac Technology supplied similar transceiver modules with some documentation. Shortly thereafter they also started supplying easier modules, the RF900DV and RF2400DV, which cut the number of channels from 40 to 16, but made those channels easy to select using 4 pins set high or low rather than setting up the PLLs directly. I purchased a few of these easier modules to start with, hoping to move to the less expensive PLL only modules later and just tonight finally have gotten around to setting them up. Unfortunately the 40 channel modules no longer appear to be available from Laipac.

one_way_test.jpg: 2 RF900DV modules, two protoboards with some circuitry, and two battery packs

I set up a one way transmission for this first test, primarily so I could limit the points I may have to trouble shoot if there are problems, but also because I lack a few parts at the moment. I'm essentially following the audio portions of the application circuit provided by Laipac. The unit on the right is connected to a headset via a 2.5mm jack (same as used for cell phone headsets), and is a single transistor amplifier feeding the audio input of the RF900DV. Note that the RF900DV has a base and a handset unit - I'm using them interchangably and won't specify which is which since it doesn't matter as far as these circuits are concerned.

The unit on the left runs the audio output of the second RF900DV into an LM386 amplifier and then to a speaker. I merely need to duplicate the circuit from one end to the other to make a full duplex unit. They appear to consume around 60mA so a small 3.6V 600mAH battery pack should give a good 10 hours use. The system is quite noisy at the moment, but I plan to improve the audio path a little bit before committing the design to a PCB.

The photo doesn't provide a lot of detail, but I've simplified the radio interface as much as possible. There are four wires leading to the transceivers: Audio out (green), 3.6V (red), gnd (black), Audio in (white). They are hardwired to be powered all the time (one could power the receiver and transmitter seperately to reduce power), and to work on channel 0. The few remaining wires (busy, data out) are left unconnected. I soldered a 3.12" wire to the antenna pin to act as a 1/4 wave antenna. The audio quality, as poor as it currently is, is very usable and doesn't change noticably as I move the receiving end around.

Primary goals for this version are:

  • Channels changable by hand
  • Easy to charge and/or change batteries
  • Volume control
  • Finish by end of March

I'm most interested in completing the design, and so the goals are all easily within reach. Secondary goals are great audio quality and handles 2 headsets per end - with the possibility of linking two or more sets together so 3 or more people could participate in a "party-line" style chat. The extra headset could also be used to input shared music into the system, record conversations (useful for training - the cyclist is never objective when they're in pain, but listening to the training session may prove useful), patch into another radio system or cell phones, etc. The music option would need an auto-mute.

However, a lot of this will have to wait for version 4 or version 5. I have a lot more in store for this system using different technology, but this will be a good small step on the path towards the ultimate communications system.