Measuring signal strength on the RF12

I have a jeenode on top of the compost bin - the signal is very weak.  So dug around and found someone was using the software rf12_getRSSI function to create a crude 7 level signal strength.  I have got so far with my new found knowledge - able to get the following:

The data has been truncated from 20 to 10 items 'cos I am not really interested in all the junk.

Clearely the 8,10,14db is wrong - it is simply reporting the RSSI and not translating it - and this is where i have the problem.

Member for

14 years 5 months

mark_baldwin

13 years 4 months ago

I've been a little busy elsewhere lately and to tell the truth not really read your post in any depth but the & B1000 is just 'ANDing' with drssi with binary value 1000 i.e. it is checking to see if bit 3 is on (assuming you count from bit 0)

To remove the <cr><lf> (and any other supious control bytes) run dos2unix on the file

e.g. if the file is example.txt then type dos2unix example.txt from the command prompt on the HAH

Member for

15 years 11 months

kim

13 years 4 months ago

Very helpful - I can continue playing.

Kim 

Member for

15 years 11 months

kim

13 years 4 months ago

xfx viewer data

Now showing sensible crude signal strength.

You can see node 2 and 3 are quite strong.  But node 4 is right down in at the same level as the noise.

Next I shall try making node 4 report more quickly - then I don't have to keep waiting approx 3 mins every 

time I adjust the receiver position.

Member for

17 years

derek

13 years 4 months ago

I had always thought that the RSSI piece on these modules needed some external analogue components in order to work. This piece is really interesting. It would be good to know if doubling the antenna length on the remote Node made a worthwhile difference.

A 'setup' version of the code could assist with the optimal positioning of the Nodes.

All good stuff.

Member for

15 years 11 months

kim

13 years 4 months ago

I did try lengthening at the jeenode end and the basenode.  It made things worse!  I think there is impedance mismatch once you go away from the quarter wave ground plane length.  The basenode is definitely best with the antenna (bit of wire) at right angles to the circuit board.  For the airwick you can't do that but I have made a small hole next to the ldr and pulled the wire out of the top - then waterproofed it all in a freezer bag (with a hole for the wire of course).

As you have mentioned before the length 82mm is important. I broke off one wire at the base by moving it around too much and tried to resolder it onto the board - about 2mm shorter.  Didn't work as well.

I might try to play with bandwidth, data rates - as suggested here http://jeelabs.org/2013/06/05/what-if-youre-out-of-wireless-range/

Meanwhile I am happy that the compost is over 50 degrees centigrade!

Thanks for your interest.

Kim 

p.s. This looks tempting http://jeelabs.net/projects/cafe/wiki/NRfMon_-_nano_Spectrum_Analyzer_with_the_RFM12B

Has anyone tried it?