Hello,
I have a problem with modules RAK3172 with TCXO for 868 Mhz - low visible distance, not more than 100 meters on the street. I use a P2P connection, without Gateway, and my custom firmware
My test conditional:
2 modules on RAK3172 with 2.5 dbi antennas with SMA connector
output power for Lora modem - 22 dbm, with TCXO and enabled DCDC for modem
first test - in the room - between modules about 4 meters, directly visible
RSSI = -100
SNR = 5
second test - on the street - direct view- first module near the window on the second floor
170 meters - RSSI = -124, SNR = -17
130 meters - RSSI = -124, SNR = -8
what’s wrong maybe? How can I increase power or change settings for a better range?
between Lora module and SMA socket - about 4 mm
For test I have changed DCDC for module, changed antennas, changed CAD settings - nothing changed…
Indeed I was having a similar problem.
The RAK3172 module is extremely sensitive to PCB layout.
In my setup with a SF of 8 and a BW of 125 KHz, RF Power +22dBm. I was getting some 50 dB of attenuation. Using the U.FL connector with coax with 3 dBi antenna
What I had to do is completely clear the PCB of ALL ground and copper beneath the RF output pad on ALL layers
Indeed I even removed the RF pad on the PCB.
As a temporary fix I removed the RAK3172 module and re-installed (or new one) it with the RF pad not making contact with the PCB. Only soldering ground, power and UART pins. That fixed it. At 1 meter you should be seeing something like -10 dBm.
You can see in the diagram what my PCB looks like.
Why bother with an external antenna connector? You’ll get better performance using the U.FL on the RAK3172 module. And you’ll save the cost of the RF connector too. The signal is from the RAK3172 RSSI which is in -dBm. This can be extracted from the EVT event (+EVT:RXP2P:-30:13:1234) or the RSSI command AT+RSSI=?. Of course you will need a transmitter with known output power. You can use an RF power meter or a spectrum analyzer to measure directly the output RF power. In my case i did a comparative analysis using the Free Space Loss equations and worked it backward. Which pretty matched what I got…The attached chart assumes +22dBm, 2dbi anntena