[sdiy] wireless patching

[mskala at ansuz.sooke.bc.ca]

On Thu, 25 May 2017, rsdio at audiobanshee.com wrote:
> FSK and ASK both assume digital input. If you want to send analog data, then you should be able to use standard FM, which can transmit analog signals. So long as the pilot frequency and the signal frequency survive the transmission, you should be able to reconstruct the analog signal on the receiving end.
>
> I do think that multiple CV would require multiple transmit frequencies and multiple receivers, though, to keep the signals distinct. That would be true with ASK/FSK or straight FM.

I'd be inclined to look at how old-style analog RC models (planes and cars
and so on) work.  The transmitter is just keyed on-off by a repeating
sequence of of pulses that consists of one sync pulse that is fixed-length
and relatively wide, followed by some fixed number (typically eight) of
data pulses that are variable-length and relatively narrow.  The
receiver/decoder is something like what we in the synth world would call a
"sequential switch" - it routes each data pulse to a different output,
stepping to the next output after each one and using the wide pulse to
reset itself.  It's easy to imagine encoding the protocol with monostables
(555-based, even!) to generate variable-length pulses and some
multiplexing logic.  The smart thing to do, but much less "analog," would
be to do the whole works in software with an appropriate microcontroller.

The outputs after demultiplexing typically connect to "servos" that
control mechanical things in the model like the throttle and the airplane
control surfaces.  These are packaged modules which combine a motor and a
control circuit, and they're designed to accept variable-width pulses to
control their positions, not caring much about the repetition rate.  The
bottom line is that it's all possible to do with fairly low-tech
electronics, and although it is digital in the sense of using pulses, it
is analog in the sense that the pulses are of continuously variable
length.  No quantization unless you're faking the analog protocol with a
microcontroller, which is probably what the RC modellers actually do
nowadays if they even use this protocol anymore at all (it is decades
old).  I don't have direct experience with it - I read about it in the
context of ham radio.

The RC model protocol normally runs at about 50 frames per second,
modulated onto a carrier in the high HF to low VHF (like 27MHz or 49MHz),
but in principle there's no reason it couldn't be done at a much higher
repetition rate using a carrier in the UHF range.  One could use
frequency-shift keying to make it more robust instead of the traditional
on-off.  Compliance with the relevant radio regulations is left as an
exercise for the student.

-- 
Matthew Skala
mskala at ansuz.sooke.bc.ca                 People before principles.
http://ansuz.sooke.bc.ca/