Project

General

Profile

Remote "pushbutton"

Added by jwp over 5 years ago

Hi,

In my application I have a central node with many nodes scattered around it, that all do exactly the same thing at the same time, as told by the central node: They switch on, or they switch off. That’s it. As they are now all equipped with RFM12B, they can do a lot more, and for a more costly variant of the hardware I will keep using the RFM12B or RFM69W and make use of their features.

But I’m now in the process of trying to make the slave hardware as cheap as possible. I’ve made a lot of progress already on other subsystems, but now it’s time for the wireless module. I was thinking, if “on” and “off” are the only commands the central node transmits (at a (flexible) rate of between 0 - 10 times per second), why use a full blown \$3 FSK transceiver in the slave nodes? Can’t I use a simple RF detector circuit or IC that just gives me a 1 if the central node is in transmit mode and a 0 when it’s not? Not only would it probably save money on the RF receiver, I could also use a simpler and smaller MCU (already went from mega328p to tiny84, but with this a tiny45/85 might do). Or I could get rid of the MCU completely..

I’ve been looking at diode based RF detection circuits and RF detector IC’s (logamps), but I’m not sure if these could be useful for my goal. I would also need to tune the circuit for the frequency the central node is using (868Mhz in my case) and off course would need to write a specific driver for the central node (using RFM12BP currently). It sounds a bit OOK like, but then even simpler.

I wonder if anybody could point me in a direction that would enable to verify my thoughts…

JW


Replies (5)

RE: Remote "pushbutton" - Added by JohnO over 5 years ago

Are you concerned about any transmitter switching your slave nodes? The MCU at offers the possibility of validating the transmitter as the legitimate one.

RE: Remote "pushbutton" - Added by jwp over 5 years ago

Well..only to the extent that my slaves shouldn’t start switching when somebodies mobile phone goes off. If I could filter on a narrow enough carrier frequency it should be fine. How could the MCU validate the transmitter without having to demodulate the signal?

RE: Remote "pushbutton" - Added by JohnO over 5 years ago

RF noise is everywhere and random triggering would be anything from irritating to dangerous. Demodulating the signal with a MCU allows all sorts or security mechanism - rolling codes similar to car remotes etc.

Remote "pushbutton" - Added by martynj over 5 years ago

jwp,

Many of the low end, OOK like solutions rely on high signal strength to get around the poor frequency discrimination that a practical tuned circuit gives at these frequencies.  That’s one reason many keyfob systems have very limited range.  A fixed or rolling pattern detect circuit might be incorporated to further discriminate between channel interference and the expected signal.
I’d suggest the transmitter is still command based (e.g. flash at rate X for Y) rather than just connecting up the load when the transmission is validated - otherwise the duty cycle may easily break the ISM guidelines and/or not leave enough “silences” for transmissions for other subsystems to get through in the same or close by channels.

RE: Remote "pushbutton" - Added by jwp over 5 years ago

Thinking about it, my current implementation does not even transmit the “off” command. The slave turns itself off based on the frequency that it is switched on at. The “on” command is now slightly more complex, containing the “on” frequency itself (in case a slave misses the next transmission) and some other gimmicks that I’m willing to sacrifice. So my reasoning was to skip the whole demodulation part and have the receiver just trigger an interrupt every time a strong enough signal is present. But I guess the tricky part is telling the difference between “strong enough” and “noise” versus acceptable range. Nevertheless, I would still like to do the experiment. I’ve seen some very simple “mobile phone detector” circuits around based on a few opamps and a few supporting passives, I could try getting something out of that. Then there are Schottky RF diodes that are cheap and seem to be meant for the purpose and at last the completely handsoldering-unfriendly RF detector IC’s.

    (1-5/5)