I’ve re-written my rf24 wrapper to tweet its findings, although as you can see the DHT11 and BMP085 seem to have gone insane, so possibly low voltage or cold issues. I’m thinking I’ll have to move to three AA batteries and a 3.3v regulator:

Altitude: 645.97m, pressure: 14.65psi, temperature: 13.60/12.31/15c, humidity: 49%, dewpoint: 4.37c, date: Tue 29 Oct 2013 15:53:02

The code is below:

#!/usr/bin/env python
from time import strftime, localtime
import sys
from twython import Twython

# twitter auth
CONSUMER_KEY = '******'
CONSUMER_SECRET = '******'
ACCESS_TOKEN = '*******'
ACCESS_SECRET = '*****'

# open file for reading
file = open("/var/tmp/rf24weather.txt", "r")

# read the line
line = file.readline()

# split on colons into array
data = line.split(':')

# print results
tweet = "Altitude: %.2fm" % (float(data[0])/100)
tweet += ", pressure: %.2fpsi" % (float(data[2])/100)
tweet += ", temperature: %.2f/%.2f/%ic" % (float(data[1])/100,float(data[6])/100,int(data[3]))
tweet += ", humidity: %i%%" % (int(data[4]))
tweet += ", dewpoint: %.2fc" % (float(data[5])/100)
tweet += ", date: "
tweet += strftime("%a %d %b %Y %H:%M:%S", localtime(float(data[7])))

api = Twython(CONSUMER_KEY,CONSUMER_SECRET,ACCESS_TOKEN,ACCESS_SECRET) 
api.update_status(status=tweet)

Update: I’ve confirmed the DHT11 problem is down to the low voltage of the coin cell, testing it with a regulated 3.3v supply works fine, 2xAA batteries does not work either. So I’ve either got to ditch the DHT11 or redesign the circuit to incorporate a 3.3v regulator and use 3xAA batteries or maybe a boost converter on 2xAA batteries. Power requirements of the components are:

I’ve soldered on my waterproof/wired DS18B20 and written a Python wrapper that prettifies the data written by the C++ daemon. So it converts a text file containing “8110:2396:1449:23:49:1170:2343:1383008407” to:

BMP085 altitude: 81.10m
BMP085 pressure: 14.49psi
BMP085 temperature: 23.96c
DS18B20 temperature: 23.43c
DHT11 temperature: 23c
DHT11 humidity: 49%
DHT11 dewpoint: 11.70c
Date: Tue 29 Oct 2013, 01:00:07

The entire code can be downloaded as a tarball

Here’s a photo of the Arduino transmitter, I’ve ordered a plastic project box to make it waterproof:

I’m still playing with my nRF24L01+’s, I’ve now got the RPi running the receiver as a daemon and the Arduino transmitter in low-power mode. The 3.3v 8MHz ATmega328p-pu now takes a reading (well at the moment its still a dumb counter!) and transmits it, then turns off the radio and goes to sleep for 30 seconds.

This is why its important that the Pi is constantly listening (well, sleeps every 5secs to reduce CPU load) and only overwrites the file if it receives a good reading. The Pi appends the unix time to the reading, which can then be parsed with Python’s time.localtime() function.

Further to my previous post I’ve moved onto using the Raspberry Pi as the receiver.

There’s a lot of incorrect documentation, and even incorrect comments in the libraries, so I thought I’d better note down what I’ve done.

First we need to get the Pi version of the RF24 library from GitHub, so on the Pi:

git clone https://github.com/stanleyseow/RF24.git
cd RF24
cd librf24-rpi/librf24
make
sudo make install

I couldn’t get the hardware SPI version to work, so stick with librf24 not librf24-bcm.

I’ve finally gotten around to having a quick play with my BeagleBone Black and all I can say is that its hideously badly documented! An easy to read pinout would be nice, or correct documentation regarding UARTs and the USB networking. For those of you who actually want correct information here goes…….

1. USB networking – this doesn’t automatically bring up a network interface on your host PC, the key here is to run ifconfig as root:

ifconfig eth1 192.168.7.1/26 up
ssh root@192.168.7.2

I know, root SSH, yuk (and no fecking password!) I found that the BBB is really fussy about USB cables, it only seems to work (power+net) with the one supplied with it.