Luckfox Forums

We've developed a custom application, and it worked great until today. Today we've started to experience issues with unstable boots. Sometimes (50% occurrence) the boot will just not happen. Interestingly enough it only happens when the RX pin of the serial console is not connected to the board.
There is no error messages that would explain why this happens and we can't even replicate it when the serial console is connected. Is there any related documentation about why the RX pin of the serial console could cause such issues?

Hello, this seems to be a hardware issue. The RX Pin may not be the root cause of the problem. There are no relevant documents that mention a similar situation.

We are seeing exactly the same behavior on RV1106 (Luckfox Core 1106).

On our hardware, UART2 RX (debug console) is not hardware-defined — there is no pull-up or pull-down resistor on that pin, so it is left floating when nothing is connected.

With this setup:

Boot is unstable (~50% failure rate) when UART2 RX is not connected.

Boot becomes 100% reliable as soon as a USB–TTL adapter is attached (even with RX only).

This makes us suspect that UART2 RX might be sampled very early during ROM boot (possibly as a boot/debug/JTAG strap), and that leaving it floating leads to undefined boot behavior.

Question to the Luckfox / Rockchip community:
Is UART2 RX expected or required to be hardware-defined (pull-up or pull-down) on RV1106?
Is this mentioned anywhere in the official Rockchip or Luckfox documentation or reference schematics?

One more detail I forgot to mention:

When only the UART2 TX pin is connected (RX left floating) and the boot fails, the serial output shows RKUART before stopping.
This only happens when RX is not hardware-defined; as soon as RX is connected or pulled to a known level, the boot becomes reliable again.

We also tried adding an external pull-up resistor on UART2 RX:

• With a 10 kΩ pull-up to 3.3 V, the issue was still present, and the RX pin voltage measured around 1.4 V.

• After changing the pull-up to 1 kΩ, the RX voltage increased to about 2.04 V, and all boots became reliable.

Based on all of the above, this looks like a hardware-level issue during early ROM boot.

However, we would still like to ask:
Is there any software or firmware configuration (U-Boot, loader, maskrom settings, etc.) that could mitigate or avoid this behavior, or is defining UART2 RX at the hardware level the only supported solution?

Any guidance would be appreciated.

The default FIQTTY of RV1106 uses UART2 RX which has an internal pull-up. It starts functioning from the BootRom stage and can be used for mirroring burn-in through the serial port. Our past designs usually involved external connections, or connecting an OR resistor in series for protection, and we have not encountered similar situations.
What I have learned is that the only two stages where the input through the serial port can affect the boot are: one is when entering the burn-in mode, using special UART commands through the FIQTTY interface to burn in the image, which does not trigger during normal boot; the other is entering the uboot stage by pressing ctrl+c to enter uboot command control, which can be reconnected to the RX pin to confirm in case of boot failure.
Since we have not encountered similar situations at present, we do not have better suggestions. Personally, I tend to believe that the fundamental cause of the boot failure is due to other factors, and connecting the RX pin just solves this problem. For example, the 3.3V voltage of Core1106 is unstable due to other loads, and when the RX pin is connected, the stability of Core1106's 3.3V is ensured by the pull-up resistor of the TX pin of the serial module.