Hello RAK team,
We are designing a battery-powered node with a RAK3172-T-8-SM-I and our own firmware (STM32CubeWL + LoRa Basics Modem, EU868). The power source is a 3.0 V Li-MnO2 primary cell (open-circuit voltage about 3.1 V). Its internal resistance is high enough that VDD will sag to roughly 2.2-2.7 V during TX pulses.
Two existing threads are relevant:
- Measurements with a stiff supply (Otii Arc) on a RAK3172-T: Supply Voltage to Current Consumption Scaling for RAK3172 . TX at 14 dBm draws about 110 mA at 2.5 V, but the board crashes at 2.25 V and below (TX at 5 dBm crashes at 2.0 V), although the datasheet specifies VCC = 2.0-3.6 V.
- RAK3172-T Low Level shows that the TCXO is an NDK NT2016SA-32M-END4263A (Vcc 1.8-3.0 V). The low-level package sets TCXO_CTRL_3_0V. DS13105 (Table 58) says the STM32WL TCXO regulator needs VDD > VTCXO + 200 mV, so a 3.0 V setting cannot be regulated when VDD is below 3.2 V.
Questions:
- What is the real minimum VCC for TX (for example +13/+14 dBm) on the RAK3172-T? What causes the crash below about 2.5 V: the brown-out reset level in the factory option bytes, the TCXO supply, or the PA? Which BOR_LEV do shipped modules use?
- Has TCXO_CTRL_1_8V (instead of 3.0 V) been validated on the RAK3172-T? Is there any impact on frequency accuracy or phase noise, given that the TCXO datasheet only specifies the 0.1 ppm voltage coefficient at 3.0 V +/-5%?
- Is the RF matching network optimised only for +22 dBm? Which PA settings (paDutyCycle / hpMax) do you recommend for +13/+14 dBm to reduce the supply current, and what TX current should we expect at 3.0 V?
Thank you!