[nickcw]

LoRa runs in the 868 MHz band (in Europe) - a frequency range known for being line of sight only as the ionosphere doesn't reflect radio waves of that high frequency.

However sometimes you can get Troposheric propagation on these high frequency bands. This used to be common when I was a teenager living in the south of England in Kent and we could occasionally receive racy TV channels from Europe.

https://en.wikipedia.org/wiki/Tropospheric_propagation

From wikipedia:

> Tropospheric propagated signals travel in the part of the atmosphere adjacent to the surface and extending to some 25,000 feet (8 km). Such signals are thus directly affected by weather conditions extending over some hundreds of miles. During very settled, warm anticyclonic weather (i.e., high pressure), usually weak signals from distant transmitters improve in strength.

Radio Amateurs love this kind of propagation. It is very variable, doesn't last long, but can give you contacts over great distances on frequencies which are usually line of sight. I think the record for radio amateurs on the 70 cm band (430-440 MHz) is over 4,500 km!

[tlamponi]

> LoRa runs in the 868 MHz band (in Europe)

No, LoRa can run on two bands in Europe, either EU433 from 433.05 to 434.79 MHz or EU863 from 863 to 870 MHz.

See e.g., https://www.everythingrf.com/community/lora-frequency-in-eur... or https://www.thethingsnetwork.org/forum/t/is-there-an-agreed-... or for a more dense, but definitive source https://docdb.cept.org/download/4316 ("EUROPEAN TABLE OF FREQUENCY ALLOCATIONS AND APPLICATIONS IN THE FREQUENCY RANGE 8.3 kHz to 3000 GHz (ECA TABLE")

Albeit I'm not sure what band they used here.

[eggfriedrice]

There is essentially no use of LoRaWAN at 433MHz in Europe, this would have been at 868

[blantonl]

Just to expand upon this, tropospheric propagation happens typically where there is a temperature inversion, meaning, air near the surface is cooler than higher up in altitude. That warmer air aloft acts as a layer that radio waves bounce off of.

[dbrgn]

Could the signal bounce off airplanes as well, or is that effect not relevant in practice? (I think I've heard such stories before from radio amateurs...)

[catbert69]

For doing meteor- and airplane-scatter stuff you need quite a bit of transmission power. The max. tx power of LoRa transmitters is way too low.

[H8crilA]

It can bounce off of meteor trails or the moon. It's been done many times in the ham community. Signals do bounce off of aircraft too, there are even passive military radars that track airborne objects using DVB-T or other strong signal reflections. It could be used if the conditions are just right.

All of this has really nothing to do with LoRa and everything to do with physical properties of the emissions (frequency, power, polarization) and physical properties of whatever happens to exist in the world at the time.

[wl]

The path loss for moonbounce is approximately 390 dB. You're not doing that by accident. In addition to using weak signal modes, you'll need directional antennas and lots of power to overcome that.

[drmpeg]

It's not that high. You have to use the radar equation with the moon being a very lossy but huge target cross section. The path loss at 432 MHz and average moon distance of 384,400 km is 261.6 dB.

[extraduder_ire]

Would it be closer to the other value if you factored in the atmosphere, and whatever else is in the way?

As an aside, would radio via the moon ever be practical? Without pumping out so much transmit power that you pick up the signal anyway, without pointing a receiver at the moon. I think it would be neat, if only for the sake of novelty.

[drmpeg]

No, there's very little loss from the atmosphere at 432 MHz. The 390 dB figure is a misunderstanding of how the path loss is calculated. The poster "wl" took the one way path loss (195 dB) and doubled it. But the moon is a huge reflector and provides "gain".

Here's a C program with the correct equation.

  #include <stdio.h>
  #include <stdlib.h>
  #include <math.h>

  int main(int argc, char **argv)
  {
    double  d, f, lambda, loss;

    if (argc != 3) {
      fprintf(stderr, "usage: moon <km> <frequency(MHz)>\n");
      exit(-1);
    }

    d = atof(argv[1]) * 1000.0;
    f = atof(argv[2]);

    lambda = 299792458.0 / (f * 1000000.0);
    loss = (0.065 * (1.738e6 * 1.738e6) * (lambda * lambda)) / (631.65468167 * (d * d * d * d));
    printf("EME path loss = %f dB\n", 10 * log10(loss));
    return 0;
  }

1.738e6 is the radius of the moon in meters and 0.065 is the reflection efficiency.

Satellites make moonbounce impracticable.

[BenjiWiebe]

It's still not going to happen accidentally.

[dbrgn]

Yes, of course this is related to the physical properties of the emissions, but LoRa uses specific frequencies of the ISM band (868 MHz in Europe), and this discussion was about the reasons why that specific LoRa transmission worked even though there wasn't a line of sight. Airplanes could be an explanation, although your sibling comment thinks there isn't sufficient transmission power for it to be feasible.

[H8crilA]

Ah. It's most likely tropospheric ducting, I thought the top comment explained it.

[KeplerBoy]

Sure, that's how you get doppler shifted frequencies and enter the realm of radars.