Hi,
I’m trying to use a RAK19713 (mPCIe-format RAK4630/RAK4631 core module) in the reTerminal DM’s 4G/LTE-designated mPCIe slot. It never appears on USB at all, with symptoms I’ve fully isolated to that slot rather than the module or host:
What works:
- The RAK19713’s bootloader was flashed successfully via SWD (RAKDAP1,
pyocd) — confirmed good. - Application firmware (Arduino/PlatformIO,
rak4630board target) builds and flashes successfully via UF2 drag-and-drop while the module sits in the LTE mPCIe slot — mass-storage enumeration (sd 0:0:0:0: [sda] ... Adafruit nRF UF2) works every time. - The same module, plugged into an external USB-A port on the same reTerminal, enumerates instantly and correctly as
239a:8029 Adafruit WisCore RAK4631 Board, with a working CDC-ACM serial interface and firmware running correctly.
What doesn’t work:
- With the module seated in the LTE mPCIe slot,
lsusbshows nothing at239a:*, anddmesgshows zero USB/xHCI enumeration attempts for that slot at all — not a failed enumeration, just complete silence. This is consistent across two full power cycles and a physical reseat. - Only USB (CDC-ACM/mass-storage) is affected — SWD-based flashing to the same module while in that slot works fine, since it doesn’t depend on the slot’s USB routing.
What I suspect, from your own reTerminal DM manual’s block diagram (section 2.1.4): the legend shows a “Wi-Fi Toggle Switch” sitting between USB HUB2, the onboard WiFi module, and the Mini PCIe USB path, controlled via a GPIO expansion IC (PCA9535, I2C 0x21). This suggests the LTE slot’s USB lines may share a hub port with onboard WiFi through a switch that isn’t currently routed to the mPCIe slot, rather than the slot being an independent always-on USB path.
Question: is there a documented way (dtoverlay, script, or direct I2C write to the PCA9535) to route USB to the LTE mPCIe slot? /boot/firmware/config.txt has no relevant dtoverlay lines by default, and I haven’t found this documented in the Hardware Guide or Getting Started wiki pages.
For context, I found a related (unresolved) report on a different reTerminal model with a similar “4G module disappears after installing reTerminal drivers” symptom: [link to that forum thread]. Wondering if this is the same underlying mux/driver interaction.
Happy to provide lsusb/dmesg output, OS version, or anything else that’d help.
Thanks,
Troy (drafted by Claude)