Hi there,
LOL 
A picture of the DUT , setup ? Can’t be anything proprietary , is it in an enclosure.?
can you post the code ?
GL 
PJ 
I would imagine that “since the default state of RF_SW is on the transmit side, it works fine in transmit mode, however it does not switch to receive mode when receiving, so it appears to be less sensitive.”
If the shield is open and pins 4 and 6 can be observed, it would make sense to check if the logic is inverted during transmit and receive.
Checking with the simplest possible setup and firmware may help to identify the problem!
As above, pin 4 and 6 has been observed and confirmed to be correct. These modules are permanently either in Tx or Rx mode. It does not toggle. And our code has been checked, reviewed and scanned thru, and that only initialization part drives the GPIOA-PA4/5 into the respective state once.
Are you at least saying that you observed that the state of pins 4 and 6 are different in the Tx and Rx modules?
Is there any possibility that the radio library could manipulate pins 4 and 6 after setting them by GPIO?
No, we scan thru every code possible, and live-view the GPIOA register. It did not change.
Latest…I have just confirmed, that the chip does not work well in SF6. Switching to SF7 normalised these reading (RSSI and SNR). However, range is still very bad. At 80dB path loss, it start to show very bad lost packets
Received RSSI, is reading -68dBm to -75dBm. But SNR reads -7 to -9dB. On my Spectrum Analyzer, the channel is clean. Something is not right, and we have got no clue.
jFYI again, this is Seeed Lora-E5, running a STM32WL5 inside.
Hi there,
So No Picture? No Code? Your “RF_expert” is out of Ideas?
You try another set of units,
Have you considered another Model then? Others have tested " Theses "
and NO-one has reported and problems, SO I would re-review your settings and timings. Check any solder work or modifications. I would also switch to antennas and try some additional testing. But your information is confusing, without seeing what it is you say is the problem.
GL PJ
@PJ_Glasso You should be able to switch between TX and RF much faster than at 100 ms intervals. The time for echoes to die down in a given RF link is microseconds, not hundreds of milliseconds. Your code in the Wio-SX1262 thread has “leakage” from TX to RX because you use radio.transmit() instead of radio.startTransmit(). Just like with reception, transmission end is signaled as an interrupt on DIO1, which will trigger your receive function. Since SX1262 uses a shared transmit/receive buffer, when you go to read the buffer after this interrupt you end up reading the data you just transmitted as a result. Jgromes recommends using the dual-interrupt approach used in the Radiolib SX1262 Ping Pong example to avoid this. I had issues with TX/RX leakage and after switching to using filtered interrupts for both TX and RX, I have not had this happen once since.
Hi there,
Yes, You are correct Sir… LOL I saw that and was able to get it much Faster, I have same demo code using interrupt ,Just DOI1 is all you get though. This was just some quick code to show the New Hardware works, None had been posted prior, only Meshtangle 
existed on the wiki early on. It’s slowly coming out.
I’ll be testing more with different ant and add the GPS, Stay tuned I’m sure We are just scratching the surface on it. 
Great stuff though forsure
GL PJ
PJ,
I may not understand exactly what are you looking for. If you want picture, let me add some. I cannot share the exact code, but the main library is direct official from ST-microelectronics. This is verified over a few units, very repeatable. why switch antenna, when VNA shows better than -15dB RL?
The setup is simple enough, that I do not understand what is to share. We are still prototyping off Seeed Lora E5 Mini…as cited.
The main test now, as shown on a spectrum analyzer, is 20m, across 4 walls. Showing very strong CNR (or SNR), and yet the Lora module reads -8dB SNR at times, with a RSSI of -68dBm to -75dBm. The SA’s noise floor is also shown, showing no other in-band signal. Even at -80dBm, there is no reason for SNR of -8dB
As shown in the link below, I am already communicating over 20 km using the same device as shown in your photo. If you are having problems with the same device, it is not a hardware issue.
We need more information on the exact information to get closer to your problem.
“Programming LoRa-E5 with Arduino, LoRa-E5 mini Point to Point Comminucation over 24 km”
We are indeed, trying to figure out, what else, did not library go wrong. Everything on paper, on instruments, show that everything is in order. We can get ZERO packet loss when its another 6dB stronger, and then hell break loss even when we are still long way from the noise floor.
We already sure, the RSSI reading in SF6 does NOT work, so definately strange, if errata did not state anything.
One thing in question, how’s your test area’s noise in term of cellular signal.
We do have cellular here at 760MHz, and 885MHz onwards. am wondering if that is the cause.