Hornet Chasing receiver on 148 MHz

[SensorsIot]

I want to use an SX1276 to receive a beacon signal on 148 MHz. This comes from a transmitter glued to a captured hornet (invasive, kills honey bees). This hornet flies back to its home, and we want to use the receiver on a drone to find it.
I can receive signals in FSK mode, but only at rxBw down to 50 kHz. Afterwards, I only get RSSI -56.0
I want to get down to the smallest bandwidth to get the highest sensitivity. So far, without success. Does anybody know how to configure the chip? The signal is a 0.1-second peak and 0.9-second pause, BTW.

[HeadBoffin]

I'm sure you know from High Altitude Ballooning, that you can get incredible range on 433MHz if you have the height - which with a drone (presumable electric, not the honey bee sort) should be no problem at all - on more than on occasion I've located a payload using direction finding that was just tx'ing LoRa via RSSI and a grid search.

What power source are you using?

Somewhat off topic, but some of the Fanstel BLE modules have decent sized PCB trace antennas that give 500m+ range and will run for months on a slim coin cell. With a decent sized antenna on the receiver you should be able to get a very useable range.

If there is room for a very smaller piezo buzzer, activating that remotely for the last 50m can really help with human based sound direction finding as the RSSI values fluctuate.

Thankfully we aren't seeing any yellow-legged visitors in the NW of the UK.

[SensorsIot]

The transmitter weighs less than a gram (including battery). So its power is only a few microwatts, and I need a very sensitive receiver to get a decent range.
Here is a link to one of the suppliers: https://atstrack.com/tracking-products/transmitters.aspx?SpeciesType=Insect&attachmentType=Glue-on-/-Tape-on

[StevenCellist]

Is there any chance you could use an SX126x radio? I've recently tried some FSK (in a completely different setup) and noticed that there appears to be some serious reliability gap between the SX127x and the newer SX126x. It may be worth a try..

[SensorsIot]

Thanks for sharing. The SX126x start at 150MHz, not at 137 as the SX127x (I do not know if this is a hard limit by the PLL). And I do not have a module handy where I could try to change the antenna matching network.

[HeadBoffin]

The transmitter size & weight defies belief - I may buy one just to take it apart!

Could the supplier shift the Tx frequency? Or is it user programmable in some way?

[SensorsIot]

These frequencies seem to be licensed for that purpose. This is not "public" space

[jgromes]

There are a couple of things to try:

1. I would first check the transmitter frequency stability. The cheap SX127x modules usually don't have a great crystal or TCXO, which in itself will cause a couple of kHz frequency error between the SX127x receiver and any transmitter. So if the transmitter is also offst or drifts in time/with temperature, on narrower Rx bandwidth the signal might fall outside the acceptable range.
2. It's not clear to me what does "receive a signal" mean, whether you just want to check presence of that signal or measure it's RSSI e.g. from multiple receivers for triangulation purposes (then again you only mentioned a single drone). If you only want to have a signal presence indication, you could try setting up the SX127x in the direct mode, and detect the CW beeps from the transmiter as changing output frequency from the DIO2 pin. This is the same mechanism that's used to receive Morse code - just running that example could tell you whether the approach is feasible or not.
3. I would also consider using the SX127x in OOK mode instead of FSK, it might improve things. I think the signal you are trying to receive is unmodulated anyway, correct?
4. There's this thread about receiving voice and SSTV on the SX127x which might offer some additional insights.

[SensorsIot]

Thank you for your answer.

1. The stability should be ok. During my experiments, I see a spike if BW are above 50k. If the stability would be a problem, I would lose the spike and only see noise. However, I see a constant -56 RSSI which is full scale (noise is below -100). That is why I think I have a problem with the chip or the software. The sketch is in my repo: https://github.com/SensorsIot/HornetHunter/tree/main/HornetHunterLoRa_V1
2. I tried the morse sketch without success. But I am not sure if I did it right. Ihe sketch is in my repo: https://github.com/SensorsIot/HornetHunter I tried with different frequencies to get the 400Hz tone
3. Our scenario: We want to locate a hornet hive where a tagged hornet stays. For that, we want to hang a device on a drone and fly pattern across the assumed location of the hive. The device has a receiver for the tracker, a GPS and a LoRa transmitter (433MHz). The lora signal consists of position and RSSI of the tracker. The LoRa receiver visualizes this data on an overlay on a map (like a heatmap). With this, we hope to be able to see where the hive is.
   Therefore I only need RSSI of a 0.1 seconds spike with a 0.9 seconds pause. I have built such a device with a Raspberry Pi and an RTL-SDR dongle that is not very sensitive. The SX1276 would be much more elegant (smaller, lighter, cheaper, and hopefully more sensitive)

[jgromes]

Thanks for the context - if only the transmitter was in the SX126x frequency range, then I would suggest usign the SX126x spectrum scan feature. Unfortunately I don't think the device will allow setting an out-of-range frequency value (and RadioLib checks that anyway). Since it's not a huge offset, one might get away with using a different crystal than the 32 MHz the SX126x expects since that is the base for the PLL. But that's rather hacky.

Regarding the SX127x, I can't rule out a problem in the code, this is not a feature that would be used often. It might be worth also trying some experiments with setRSSIConfig to change the number of samples for the RSSI.