This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AM3356: UART boot fails

Part Number: AM3356
Other Parts Discussed in Thread: TPS650250, TPS65216

on a new customer board, UART boot fail, host send Xmodem package upon receive CCC. captured waveform as below, when host send the first package, target AM335x response with NAK 

Compare with BBB board wave form, AM335x response ACK after the whole package:

Below is customer's SOM board PCB, route out UART0 TX/RX to the post card board edge directly, there is no serial or pull up/down resistor. use USB to UART TTL cable to connect on the solder point(RED) to capture waveform. 

connect with RS232 from base board, same result:

Use logic analyzer to capture the RX signal, the seems add one more bit at each byte.

if use USB to TTL cable, occasionally can boot successfully. the successful waveform captured with logic analyzer as below:

Question:

Tried with several PC as host, same fail result. Any clue to debug this issue?

  • Please check your terminal settings. They must be 115200bps, 8 bits, no parity, 1 stop bit.

  • Biser,

    we confirmed it. same setting as working with BBB.

  • Biser

    I think my question is not resolved. I need help.

  • Tony,

    Tony Tang said:
    Use logic analyzer to capture the RX signal, the seems add one more bit at each byte.

    I had a closer look at your logic analyzer screenshot. A couple of observations:

    • You should set the analyzer to trigger on the falling edge (not the rising edge), as UART idle state in the logic-level domain is HIGH. Once you fix that settings all your framing errors etc that are flagged by the analyzer software should go away
    • If I "measure" the signal in the scopeshot using an image manipulation program I see that 1 bit is about 15 pixels wide and that 4 bits are 60 pixel wide, which means an entire byte should be 10 bits x 15 pixels/bit=150 pixels wide. Which is exactly the case for the waveforms you show. So there is no extra bits here.

    Tony Tang said:
    Below is customer's SOM board PCB, route out UART0 TX/RX to the post card board edge directly, there is no serial or pull up/down resistor. use USB to UART TTL cable to connect on the solder point(RED) to capture waveform. 

    Can you try a different SOM module? Maybe there is something wrong with this particular module?

    Tony Tang said:
    Bif use USB to TTL cable, occasionally can boot successfully. the successful waveform captured with logic analyzer as below:

    It sounds like there might be some intermittent hardware issue at play here, potentially related to clock or power startup, or other reasons. Can you see if you can characterize the failure behavior more, such as:

    • What happens when you start the process from a power cycle vs. a device RESET
    • What happens if after power cycle you wait let's say 10-20 seconds before attempting UART boot
    • When you do the experiments with the USB to TTL cable, is the circuitry involving the D11 level shifter disconnected? If so, please disconnect it.
    • I see the D11 level shifter circuitry has a 1K pull-down resistor to GND on the RS232 connector side. Is this even valid? RS232 signal levels are either a negative voltage or a positive voltage, typically in the range of +/-6...12V. Tying an input like that to GND could potentially result in some intermittent behavior of the RS232 level shifter, resulting in "garbage" getting received by the AM335x, which could be a reason for the NAK

    The idea is to find a way to reliably re-produce working and non-working scenarios so we can isolate what may be causing it.

    Regards, Andreas

  • Thanks Andreas,

    This new version board uses TPS65216 instead of  TPS650250 of previous working board. 

    5pcs prototype boards have the same issue. 

    Andreas Dannenberg said:
    What happens when you start the process from a power cycle vs. a device RESET

    I did a manual reset after power on, Same result.

    Andreas Dannenberg said:
    What happens if after power cycle you wait let's say 10-20 seconds before attempting UART boot

    That is the most way we were doing, same result.

    Andreas Dannenberg said:
    When you do the experiments with the USB to TTL cable, is the circuitry involving the D11 level shifter disconnected? If so, please disconnect it.

    Now we only use the SOM board to test, no level shifter.

    I got two pcs boards on hand, one can download to num #2 package, one fail on first package.

  • Andreas,

    I found out the reason, customer made a mistake on mounting resistor on sysboot[15:14], it is reversed. 1 connect to GND end.