I have a question for the Infineon employees.
It looks like there is a bug in the ASC bootloader of the XMC1403/4
through which it is not possible to communicate with the IC.
The smaller devices from the series are probably not affected.
When delivered, the XMC is configured to ASC-BL via BMI.
If after a power-up the sequence 0x00 0x12 is sent to the XMC (P0.14),
he usually answers with BSL_ID (0x5D).
Apparently, the ASC bootloader in the XMC140x has an autobaud function,
with the bug that the first found baud rate is retained forever, even if it is completely wrong.
As a result, the smallest voltage dip (e.g., turning on a JLINK) on pin P0.14 is considered communication and baud rate detection.
The effect looks like this:
- Voltage on and P0.14 stuck on 3.3V. Then sequence 0x00 0x12.
Reaction: successful answer.
- Voltage on and sometime after that a short or long break (glitch) in the voltage at pin P0.14.
Reaction: A response will NEVER be sent again.
Is there any way to communicate with the XMC,
if once after switching on, a (switch-on) pulse on the line P0.14
and the XMC has set the baud rate (wrong) to this glitch?
This currently prevents the JLink from being able to change the BMI on XMC140x devices!!!!
Cyclic transmission of the sequence (even over hours) is not a solution!
It looks like there is a bug in the ASC bootloader of the XMC1403/4
through which it is not possible to communicate with the IC.
The smaller devices from the series are probably not affected.
When delivered, the XMC is configured to ASC-BL via BMI.
If after a power-up the sequence 0x00 0x12 is sent to the XMC (P0.14),
he usually answers with BSL_ID (0x5D).
Apparently, the ASC bootloader in the XMC140x has an autobaud function,
with the bug that the first found baud rate is retained forever, even if it is completely wrong.
As a result, the smallest voltage dip (e.g., turning on a JLINK) on pin P0.14 is considered communication and baud rate detection.
The effect looks like this:
- Voltage on and P0.14 stuck on 3.3V. Then sequence 0x00 0x12.
Reaction: successful answer.
- Voltage on and sometime after that a short or long break (glitch) in the voltage at pin P0.14.
Reaction: A response will NEVER be sent again.
Is there any way to communicate with the XMC,
if once after switching on, a (switch-on) pulse on the line P0.14
and the XMC has set the baud rate (wrong) to this glitch?
This currently prevents the JLink from being able to change the BMI on XMC140x devices!!!!
Cyclic transmission of the sequence (even over hours) is not a solution!