Microcontroller Based RC Receiver Project

[Daniel Rubin]

Size my friend... I don't need the resource required to control the
servos... I am planning on using PWM output from the PIC to vary the
speed of 4 small motors directly. I am shooting for something under 2
grams with 1000 ft. range.

 

[TexMex]

Outside of the bit of pique, what the Natural Philosopher said is
correct. You have to amplify the radio-frequency signal at the
antennae, filter out the signals you aren't interested in, and
demodulate the resulting FM signal before you can do much digital
processing. What you have to do to get a useable signal to your PIC
is three-fold…. : )

The first part, the amplifer/filter, is called the "front end" of the
receiver. You need a very low noise transistor(s) to do the
amplification, since all semiconductors (resistors too) generate some
background noise all by themselves. Once the signal is boosted to a
higher level, the filtering can be done in a lot of ways- passive and
active filters, and various combinations of both. The demodulation
process can be done many ways too- phase locked loops, digital
filters, etc. Each has it's advantages. Regardless, you're pretty
much locked into having to use a few passive components.

For your project I would think one of National Semiconductor's FM IC's
would fit the bill. Cheap, small, and light. Get their IC application
book and take a look at some of their sample schematics. Just keep in
mind that RF receiver/amplifier design is a complex project, and
things like temperature drift, noise, and shielding come into play-
things that aren't always in the digital designers toolbox.

One idea to initially circumvent some of these difficulties would be
to get hold of an RC receiver and schematic and using a scope, find
the decoded signal point. Wire from there to your PIC. Then you can
program your digital processing with a known good signal. Once you've
got your code working, you can delve into the design of your front
end. It's tough to de-bug both components of a prototype analog and
digital system at the same time, as you'll have a lot of
interdependent variables to sort out.

I applaud your curiosity, and even if the result isn't as good as a
multi-million dollar corporation's offerings, you'll have gained a
wealth of practical knowledge, and the personal satisfaction of having
done something not many of us can do!

Good Luck,
John

 

[The Natural Philosopher]

Outside of the bit of pique, what the Natural Philosopher said is
correct. You have to amplify the radio-frequency signal at the
antennae, filter out the signals you aren't interested in, and
demodulate the resulting FM signal before you can do much digital
processing. What you have to do to get a useable signal to your PIC
is three-fold.... : )

I think a DSP is fast enough to do the detection from 455khz tho. AND
maybe do digital filtering as well. So the last half of the IF strip is
digitally possible IYSWIM

 

[Johnboy]

What I expect the processor RX to be is 1or 2 dual gate mosfet rf amps
a few passives, and maybe one coil. Then the processor takes over
with the use of its single master crystal. User selects any 72 or
75mhz
frequencies via dip switch, then the processor provides phase lock
loop,
IF strip, narrow banding, FM/PCM demod, and up to 24 channel output.
Lets go way beyond the antique National IC design. As we all know the
decoder is super simple and not worth even an additional ten cent IC.
Also go to a high number of channels. Granted that the plane and car
users don't need all the channels, but the Scale Boat builders can use
them. A trace cut on the board can be used to lock out the unwanted
band. This concept is exactly what the major radio suppliers do not
want
to see, since they make most of their profit selling 19 cent crystals
for
$15.00 each, and selling 20 different receivers. One receiver will
fill the
bill for ALL possible RC activities except the 27 and 53 bands. These
bands are too far from the 72-75 range to use the same front end. A
few
passives and coil change would be needed for the lower bands. I may
be way off base expecting this much performance from a processor
based RX but is there anyone out there wanting to give it a try?
Johnboy

 

[TexMex]

I think a DSP is fast enough to do the detection from 455khz tho. AND
maybe do digital filtering as well. So the last half of the IF strip is
digitally possible IYSWIM

Yes, depending on the S/N. Programing a DSP to handle low S/N in a
mission critical situation is not for the faint of heart, or
hobby-level funding. At a minimum, going to need an In Circuit
Emulator, high-dollar recording digital scope, a digital signal
generator/digital noise source, and the usual lab equip. A circuit
sim. program is nice to learn on, but I haven't seen one for under
$10K that will handle mixed mode analog/digital processing in a
non-idealized environment (Spice on a PC ain't gonna get it done.)
Cell phone grade dsp is a REALLY bad idea when a 25lb, 150 mph Extra
swinging a 15" meat cleaver is involved. And yeah, the National rf
IC's are old, but you can set 'em up with a 50 MHz analog scope and a
Radio Shack multimeter.

I suspect that most of this discussion is moot, though, as there is
none so blind as will not see. The premise of the thread smacks of a
pianist, who having mastered Eine Kline Nachtmusic on his Hammond,
demands to be shown how to play the Brandenberg, III. At Charte'.

 

[Tom Burke]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Older remote controls used either PWM or PPM (Pulse position Mod)...
Both effectively amplitude shift keying (ASK)... We designed a
project once that used a single (8 pin?) motorolla IC that was
intended to demod FM signals with some minor support circuitry. We
used all SMT caps & coils (they all looked like SMT resistors), and a
PIC to extract the signal. It worked well enough, but there was a
real problem with multipathing in our environment. We operated in
the 50MHz regime.

YOU cannot use a PIC to amplify low level RF, period.




That has many other components apart from the large IC, which may,
ormay not, be a PIC. a PIC talks high level digital signals, and
maybe low frequency high level analogue. You HAVE to have an
anlogue circuit somewhere. Using a microcontroller to synthesize
the local ocsilllator is possible, but you still need a front end
and mixer of some sort, and doing without some form of filtering
there is dangerous in the extreme.

As far as that reciever goes, its a two IOC design with preset
crustal solderd in. Its not possible to see what IC'ds are used,
but one is probably an integrated RF/IF/detector/Mixer/oscillator
chip, and the other is probably a pic doing decode - instead of
teh standard CMOS 4017 IC.


My guess is there are a few coils on the back, and probably a few
ceramic resonators.

-----BEGIN PGP SIGNATURE-----
Version: PGP Personal Privacy 6.5.3

iQA/AwUBP1yguNPjBkUEZx5AEQLyAwCg6LeSTHPIizfI4VPqsJXBF+Ws3KwAnj40
kanTBdxvYktBdF/Sv7sazF8j
=E9ep
-----END PGP SIGNATURE-----

 

[J]

Well because you may want it to look like a real airplane, in which case
ARTF is a no go and a lot of the kits are not much better, or you may want
it to serve a certain purpose, e.g. a load carrier, in which case most kits
are no good. With a RX though, you could even buy a micro 7 channel model,
for about $30 which will still do everything a homebrew could, for less
money, less wasted time and it will be much more reliable

Justin.

 

[J]

AH ha, building a roswell flyer are we?