I had a situation where I had a Raspberry Pi 3B+ in a location where there were two usable wifi networks in range, but neither were particularly reliable in terms of Internet access. The wifi networks themselves would stay up, but the ADSL connections used by the routers would sometimes drop and stay down for a while (in the case of one, until someone rebooted it). So I wanted the Pi to switch from one network to the other in the event of losing Internet connectivity. So I wrote a bash script that can be run periodically by cron.
There are a number of ways to achieve this, including using additional packages like nmcli, but I wanted something that should just work on a plain version of Raspbian.
I also decided to test it on older Pis without built-in wifi and using a doingle instead. I found that a few tweaks were required to get the script to work reliably.
I also encountered one situation where the Pi under test ended up not connected to any wifi network at all. Annoyingly I have not been able to reproduce that outcome to identify what went wrong.
The script is set to switch between two networks, SSID1 and SSID1 - obviously you'll need to change these names to reflect your own environment.
To use the script you'll need two different copies of your wpa_supplicant file - which contains the network name and pre-shared key (password) required to authenticate - , one for each SSID. Give them sensible names like wpa_supplicant_1.conf and wpa_supplicant_2.conf.
The script logs to a file wifi.log - this can be useful for debugging. It also echoes to the terminal too, to make it easier to see what's going on when testing. If you don't want any of this, just remove or comment out the 'echo' lines.
There's no reason that this shouldn't work on other Debian distros or more widely on Linux, but I have not tested it.
The script needs to be run on a Pi with sudo. If you want to schedule using cron, then type sudo crontab -e and then add a line to run as frequently as you like. For example: 0 */2 * * * /home/pi/wifitest.sh Don't forget to make the file executable chmod +x /home/pi/wifitest.sh
One of my birthday presents this year was two 900MHz Adafruit feather M0 LoRa 800MHz boards. I hadn't had a chance to try them out but just before going on holiday I realised that our destination in the Alps might provide a nice testing ground (or mountain) so added them to my hold luggage.
I know that LoRa tech is capable of impressive line-of-sight distances when there are no obstructions, but as I hadn't had time to pack my HAB gear, I needed to find some other way of testing the range.
The Feather boards are a nice size and I was just using a length of wire as an antenna. Example client/server Arduino code from the Radiohead RFM0x library was good enough for my purposes. Each board would blink its LED when it received a packet so I could see if data was being exchanged.
For the tests, one board stayed back at base in the chalet - shoved out the Velux window for maximum height.
It was very sunny and hot so I decided that the base Feather ought to have some shade.
The second board came along with me on my travels, powered by a battery pack.
Over a few days I took the Feather out with me, first of all just cycling around the town, then for a grander test, heading up the mountain in the Grand Massif tele-cabin.
On the ascent, I was surprised to not be picking up any packets, even though I had been able to get good reception if I wondered about in the car park at the bottom. Thinking that the cabin itself might be acting as a bit of a Faraday cage, I poked the Feather out the window, and immediately started seeing the blinkin' red LED.
The contours of the mountain meant that there was actually quite a lot of rock and trees in between the two boards right at the top of the tele-cabin, and despite some wandering around to try to get better reception, the LED stayed off until we'd trekked some way further up the mountain.
Fortunately you can also get a chair lift even higher, so we continued to ascend to 2100m (the chalet itself is at 700m).
At the top the signal was very strong and with binoculars we could actually see the chalet down in the valley.
I recorded the various locations using the GPS Waypoint app on my phone and then exported these via KML into Google Earth. Red markers are places where there was no reception, green are where packets could be received.
This allows me to calculate the straight line distance between the two boards - the maximum distance was 5.86km,
Even on flat ground, with trees and buildings in the way, I was still able to exchange packets at 2.60km, and after that point I rounded the edge of the opposite mountain which probably blocked any further data.
So I was very impressed with the boards. They'll certainly be up to relaying data from my new shed....