First ridge test to test the far field of the antenna. We tested two sets of LoRa parameters: one which is fast but unreliable, and one which is slower but more reliable. We also tested multiple levels of attenuation to simulate various distances from the satellite. This is primarily a proof-of-concept for ridge tests, and to help us iron out details for more thorough and precise testing.

Tx: Stanford dish trail, near W6YX ham shack
Rx: Stanford dish trail, near Dish
Distance: 1.25 km

Chris transmitted a message once a second from the satellite antenna, which was connected to a battery and fastened to the end of a long cardboard tube (to hold it at least 1m away from ground/metal/humans). Chris was located at the peak of the ridge near the W6YX ham shack station. Sarah received the message with a 70cm 3-element Yagi antenna, located at the peak of the hill next to the big dish. The receiving script in Mu collects 100 messages, compares each one to the expected message, and counts how many messages were received incorrectly.

The two settings we chose for LoRa Parameters:

	Fast, Unreliable	Slow, Reliable
Bandwidth	500 kHz	125 kHz
Spreading Factor	7	12
Coding Rate	5	8
Preamble Length	8	8

For each of these two settings, we added increasingly more attenuation onto the receiving antenna, inserted between the Yagi and the motherboard. We measured each set of parameters with an added attenuation of 0dB,-10dB, -20dB, -30dB, -40dB, -50dB, and -56dB (the maximum possible with our set of attenuators).

Fast, Unreliable settings:

We were able to add up to -20dB without missing any messages. When we reached -30dB of added attenuation, the signal was received at -81dB (according to the CircuitPython script) and we missed 2/100 messages. When we reached -40dB of added attenuation, the signal was received at -87dB and missed 1/100 messages. However, we also had code issues, where partway through the message collection we would reach a Unicode error and have to start over. The -40dB test took three tries to get a full test without these errors. (We also got this Unicode error the very first time we tried testing without any added attenuation.) Signals with added attenuation of -46dB and higher were too weak to detect reliably enough for a full test.

Slow, Reliable settings:

We were able to add the maximum attenuation (-56dB), for a received signal strength of -97dB, without missing any messages. This is the expected behavior, since these LoRa parameters should be much more reliable. These messages also expectedly sent slower, which made for a significantly longer testing procedure when added onto the 1-second delay for both transmitting and receiving the messages.

For more details: see the results spreadsheet

For the code: see the Rx script and the Tx script

• Figure out what is causing the Unicode error
• Deal with 1-second delay to make testing faster
• Focus on higher end of added attenuation (-30dB and up) since we know that 0 to -20dB is always reliable
• Experiment with more varieties of LoRa parameters

For the second ridge test, we're focusing on testing how robust our signal is to doppler shifts in the received signal, and how this effect changes for different LoRa parameters (focusing on bandwidth and spreading factor).

Tx: Stanford dish trail, near W6YX ham shack
Rx: Stanford dish trail, near Dish gate
Distance: 1.06 km

All testing equipment is the same set up as the previous ridge test, except we're now using our new set of attenuators instead of those borrowed from SSI, and using code that is faster and more robust to errors (i.e. fixed time delays and Unicode errors from last time). We're still saving data by hand into a spreadsheet because we couldn't get the micro SD card to work yet. To test a doppler shift, we added the shift to the transmitted signal. This way we determine if we can still reliably pick up the shifted signal at the nominal frequency of 433MHz.