B-WL5M-SUBG1 and NUCLEO-WL55JC radio performance

[RCool.1]

I have 3 Nucleo-WL55JC and 2 B-WL5M-SUBG1 systems all running same code and located within a 1 foot of each other.  Additionally I have custom sensors based on the STM32WL5MOC located 4 to 8 feet away.  The Nucleo-WL55JC modules report rssi values between -50 and -60 dBm.  The B-WL5M-SUBG1 modules report rssi values between -100 and -110 dBm.

Physical range test suggest these reads are real.  As I move the sensors away the B-WL5M-SUBG1 lose the ability to receive the packets much sooner than the Nucleo-WL55JC.  

Note the reads above are based on unmodified modules. 

There is a footprint on the B-WL5M-SUBG1 modules for a pi filter in the antenna line.  The factory configuration is a single serial 0 ohm resistors which is not performing very well.

The datasheet for the for the STM32WL5MOC states that for 915MHz operation:

"Max output power 21 dBm is obtained for Z load = 63 + j×22".

I have tried adjusting the impedance seen by the STM32WL5MOC on the B-WL5M-SUBG1 using the pi filter and have been unable to significantly improve the performance - I can't get anywhere near matching the Nucleo-WL55JC rssi.

Can the B-WL5M-SUBG1 be configured to work as well as the Nucleo-WL55JC?  If so any suggestions on how to make that happen would be very helpful.

[STTwo-32]

Hello @RCool.1 

Are you configuring the same parameters for both Nucleo-WL55JC and B-WL5M-SUBG1 (For SF, TX_Power...)? If no, please try with similar parameters. 

Best Regards.

STTwo-32

[brianlogan_darcy]

I've also experienced the poor range with the B-WL5M-SUBG1, though I use 868MHz. It turned out to be because I was using the PingPong sample from the NUCLEO-WL55JC instead of the PingPong from B-WL5M-SUBG1. The difference is that the RF switch control lines work differently such that RFO_LP and RX are transposed. The result is that the antenna is connected to RFO_LP instead of RX when receiving. So, presumably the limited range is due to just enough signal being leaked to get anything at the antenna at all. Similarly, in transmit, the antenna is connected to RX, so only leakage transmit power finds its way to the antenna.

With a NUCLEO-WL55JC running PingPong, communicating with a B-WL5M-SUBG1 with the correct PingPong programmed, I get an RSSI on each end of about -23dBm, compared with about -57dBm with the incorrect PingPong in the B-WL5M-SUBG1.

With two NUCLEO-WL55JCs running PingPong, I get about -8dBm, so there's still room for improvement on the B-WL5M-SUBG1. The B-WL5M-SUBG1 has no pi matching done, wherease the NUCLEO-WL55JC does have pi matching fitted. I don't know if this totally explains the 15dB difference.