Recap of 18 months experience with cheap 433 MHz Tx and Rx

saunj

Senior Member
I have been replacing X-10 home automation modules with these Wen Shing transmitters and receivers; their models TWS-BS3 and RWS-371-6 which I get from Spark Fun as their WRL-10534 and WRL-10532. Having made 7 transmitters and 10 receivers I have yet to experience a false control action. Two of the transmitters and receivers use data messages sent once per minute. All use the same technology which I will attempt to describe better than my previous post.
The attachments are typical and are of a display in my living room which receives both data messages and control messages to select which item to display. Space does not permit a description of a transmitter, which is trivial.
The digital output of the receiver, pin 2, is directly connected to the Picaxe pin C.3, for SERIN. It is also connected to an itegrate-and-dump circuit then to pin C.0, the interrupt input. In the absence of a signal the receiver output is saturated with thermal noise. Peaks exceeding the 3V gate threshold of the FET cause the bipolar transistor to shut off, and its collector voltage to rise. However, as soon as the noise input falls below 3V, the bipolar transistor very quickly dumps the charge on the capacitor on the collector. The circuit is designed so the interrupt is triggered infrequently on noise. This is my first trick.
My signals are preceded by a 20-ms pulse, made by setting the transmitter Picaxe Serout pin High, Pause 20, then Low. After 4.4 ms the interrupt occurs. The pulse also sets the receiver’s AGC to an appropriate level for the rest of the message.
For the rest please refer to the interrupt section at the bottom of the program listing. First, the time is measured until the C.0 goes low, using a loop. Almost always a noise-generated interrupt will be less that the 15.6 ms of the true signal. This terminates the interrupt so that the Picaxe is not tied up. This is my second trick.
After that, my messages have a 10ms gap before SEROUT. This is my third trick. It allows time for the Picaxe to set up SERIN.
SEROUT is at 2400 baud not the 4800 claimed by Wen Shing. Incoming data is stored in memory, but only after a key code is received. I use 14L1776, the license plate number of one of my former motorcycles. This is my fourth trick.
SERIN will time out since enough memory is allocated to exceed my longest message. If not, interrupt is terminated.
The first stored character of my message, after ”6” is a control action or a message ID. If this was not changed by SERIN, interrupt is terminated. This is my fifth trick.
The rest is specific to this device and I will not go into it.
 

Attachments

Last edited:

PhilHornby

Senior Member
This looks interesting. Can you post a larger version of the schematic? - the original's not quite legible...
 

premelec

Senior Member
Great to read of this success - are there a lot of 433MHz signals in the vicinity you operate in?
 

saunj

Senior Member
PhilHornby: Sorry, I should not have chosen a gif-the forum made me make it smaller, and this loses lines. It's a jpg now
premelec: I have lots of RF devices myself, but only 1 I know to be on 433, and it's used manually only twice a day. No commercial transmitter would get past my filters anyway.
 

premelec

Senior Member
OK - I thought in some urban areas 433MHz devices are all over - I noticed you didn't use RFIN and RFOUT though you have qualifiers - glad it works. IoT seems to be coming on strong - not sure what frequencies they are going to be using as they are just getting protocols agreed to ... Thanks for sharing your success!
 

saunj

Senior Member
I don't see the usefulness of RFin-out. An extra chip is needed and its only good for 8 bytes - I have messages of 30 bytes.
If I wanted message integrity I would have added a CRC and send it twice, as most commercial products do. (Garage door openers, key fobs)
I'n not against Manchester encoding - I was responsible for it being specified for Flight Data Recorders, and I am an Old Mancunian.
When I was a schoolboy I learned about it on a visit to the first English general-purpose computer at Manchester University with Alan Turing.
 

premelec

Senior Member
The RFIN/OUT are in the firmware on the 14M2.... No external chip requirement: from what you say you ARE old... :) [I'm 80 shortly].... what ever happened to tunnel diodes and ovonic switches as well as gel integraters [Solion]? So much cleverness so little time!
 

bpowell

Senior Member
I've had a 433 transmitter sending data every 15 seconds for about 5 years now...it's working great! The receiver also works...I'd filled that 08M up to the brim...253 of 255 bytes if I recall! I sometimes think about dropping a 08M2 in there as a replacement...but if it ain't broke...
 

Jeremy Harris

Senior Member
I've also had really good results with the cheap 433MHz Rx/Tx modules, just using raw data prefixed with a qualifier. I have a few of these systems around our current house and have never had a problem with them at all, they just seem to work reliably all the time.

However, when I used exactly the same set up to transmit some simple energy consumption data once every ten seconds for a few metres at the new house we're building I've found that the reliability is hopelessly poor, even at 1200 baud. I also have real problems getting the commercial Byron remote light switches to work reliably; the maximum range I can get inside the house is only around 3m, any further and they stop working too. A quick look at the spectrum using a friends spectrum analyser shows that the 433MHz band at the new house location if absolutely jammed solid with signals, probably from the plethora of home automation, weather stations, heating controllers etc that all work on the same frequency. I changed the energy consumption system to run at 868MHz, using a pair of ERF modules and the performance is excellent, probably because there are very few systems that transmit continuously (or near continuously) at this frequency in our area.

Even 2.4GHz isn't good at the new house, as I found I couldn't get my 2.4GHz wireless CCTV system to work at all well there, and have had to get some 5.8GHz replacement transmitters and receivers to get it to work. I haven't checked, but suspect that the problem on 2.4GHz is the same as that on 433MHz, just too many people trying to use the same frequency band at the same time. I'm glad I wired the new house with ethernet cable, so we aren't reliant on wifi, as I strongly suspect that wifi performance may well be degraded because of the high level of local interference.

I suppose the moral of this tale is that you need to be cautious when looking at any radio datalink and choose a band that is relatively interference free in your location.
 
Top