During the HALT test at IFIC Valencia it was found that the combination of the backplane20sfp_v2.0 + SCB v3.4 failed booting at temperatures below +5 degrees.

This issue is due to a combination of circumstances:*

• Backplane20sfp has a CP2015 UART [1] which powers-up in "suspend" (see [1], paragraph 5) when no USB cable is attached. IO is configured in GPIO mode which measn that the TXD output is configured as input and the ARM DBG_RXD line is floating.
• The ARM apparently can't handle the floating RxD (see also [2], paragraph 10.5). Either a defined '1' (non-asserted RS232) or a '0' will satisfy the ARM.

Solution:*
Put a pull-up (4K7) at the CP2105 TXD outputs to avoid floating lines during "suspend" and signal "non-asserted RS232".

Other observations:*

• The SCB V3.4 should have taken care of a defined DBG_RXD input, i.e. in a future version there should be a pull-up on the SCB board.
• The SCB v3.4 works fine in combination with the miniBackplane_SFP_V3_3 [3] since on the miniBackplane CP2102 devices are used at 3V3 and their outputs are made 2V5 (ARM VDDIOP0) compliant using a resistor divider, hence effectively having a pull-down when the CP2102 powers-up in "suspend" mode. Therefor the ARM DBG_RXD is defined and the SCB never fails boot when plugged onto a miniBackplane.
• Apparently temperature influences the leakage current on the ARM and/or CP2105 thus influencing the floating voltage.

[1] https://www.silabs.com/documents/public/data-sheets/CP2105.pdf
[2] http://ww1.microchip.com/downloads/en/DeviceDoc/Atmel-6438-32-bit-ARM926-Embedded-Microprocessor-SAM9G45_Datasheet.pdf
[3] https://www.ohwr.org/project/wr-switch-hw/blob/master/circuit_board/wrs_v3/Docs_PDF/miniBackplane_SFP_V3_3.pdf