Decoding a modulation tone
William Sitch
I'm specifying an application that modulates one of many tones
onto a 56.8kHz carrier, which is then transmitted via infrared light.
The beacon will use two modulation "tones": 200Hz and 300-2000Hz, in steps
of 100Hz. (ie: one beacon might modulate 200Hz and 700Hz, another might
modulate 200Hz and 1100Hz) Each beacon will only transmit one modulated
carrier at a time.
Building the beacon will be fairly straightforward: a pair of 555s, a 556,
a crystal oscillator or cheap ceramic resonator and divide-by-xxx, etc.
There's a million ways to do it, and all are fairly simple.
Receiving and decoding the signal is turning out to be a little more
complicated than I initially expected. LITEON makes a fantastic IR Remote
Control Receiver Module that sets Vout low when the 56.8kHz infrared
carrier is detected. It's cheap (CDN$3.91), easy (3 pins, two are
power!), and perfect for this application. The output on Vout will be the
modulation tone: a square wave from 200-2000Hz.
The application calls for detecting a range of the tones. One robot may
need to detect all tones above 250Hz, while another might need to detect
from 150-650Hz and from 750-2050Hz. This means that I want to be able to
check a RANGE of frequencies and if a signal is present, drive a digital
input port of the microprocessor high. I don't need to correctly identify
each tone.
The obvious solution would be to use a number of 567 tone decoder ICs -
there's even a low-power CMOS one available from National Semiconductor.
But because these ICs won't cover large ranges of frequencies, one is
needed for each modulation frequency. That seems to be overkill - 19
chips!?. Another problem is the number of signal cycles required for a
positive 'detect'.
I guess we could use frequency-to-voltage ICs, but I don't have any
experience with these. How would I tune them for the frequency range I
was looking for? This application is interfacing with the MIT HandyBoard,
so using one of the analog input ports (ADC, 8-bit) is an option.
I'd like to see the students using a PIC, but the learning curve might be
a bit much - they've all been avoiding the PICs to this point. With a
PIC, one could time the positive pulse width and use standard logic, or
even attempt a FFT. Doing this stuff on the HB would be nice, but it's
going to be busy enough as-is.
Other things I've been kicking around are peak detection, filters and
windowing, pulse gating, etc, but I'm not sure if these sort of things can
be realized with a low-cost and easy-to-interface IC.
Any ideas? I'm beginning to kick myself that I didn't spec DTMF tone
frequencies instead of this customization.
(If you're interested in the application, http://robotag.carleton.ca)
--
William Sitch, B.E.Eng. (1999) [Post from: Carleton U residence]
M.E.Eng. Candidate, OCIECE http://www.engsoc.carleton.ca/~wsitch
DKDeckmann
tones...
In article <8tcfhf$qhk$1...@bertrand.ccs.carleton.ca>, William Sitch
William Sitch
: Hi all:
[snip]
I've moved to a new thread: "Detecting multiple IR beacons", and sent the
followups to comp.robotics.misc.
--
William Sitch, B.E.Eng. (1999) [Post from: Nortel Networks (CRK/12/1A5)]