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.

AWR1642: Programming a blank QSPI Flash Part on AWR1642BOOST

Part Number: AWR1642
Other Parts Discussed in Thread: UNIFLASH, ,

I have a new design based on the AWR1642BOOST.  I had blank Quad SPI Flash chips soldered down because the bootloader flow shows that I would be able to program it.  With SOP mode set to 101 (05) I get an error under Uniflash:

9/11/2018, 11:29:34 AM] [ERROR] Cortex_R4_0: Serial port COM5 specified does not exist, is already open, or permission is denied!!
[9/11/2018, 11:29:34 AM] [ERROR] Cortex_R4_0: !! Aborting operation!!
[9/11/2018, 11:29:34 AM] [ERROR] Cortex_R4_0: Not able to connect to serial port. Recheck COM port selected and/or permissions.
When I replace the Quad flash part on an actual AWR1642BOOST eval board, I have exactly the same issue, with the SOP jumpers also set to 05.
Is there an additional step that I am missing?  Does the blank flash part need to be programmed with some initial firmware as opposed to being completely blank?  The block diagram implies that the boot ROM inside the AWR1642 handles that, but it is not 100% clear.
I do see the AWR1642 on my new design send out 8 SPI clocks to the quad flash, so I believe it is functional.
Thanks in advance.
  • Hi John,

    The bootloader is in ROM and should be able to detect the COM ports without flash, and regardless of SOP settings. According to Figure 4 of SWRA551, the bootloader goes directly to UART when in SOP5 mode.

    In the "Settings & Utilities" page of Uniflash, there is a "Format SFLASH" button, which erases the flash. I took one of my 1642 EVMs, put it in SOP 5 mode and erased the flash with this function. I then power cycled the board and restarted Uniflash, and it is still able to detect the EVM's COM ports, as is Device Manager and TeraTerm. My guess (in your case) is that perhaps the COM port numbers have changed, or another program has "stolen" the ports, or perhaps there's an issue with your flash - is it a supported flash? See Figure 6 of document SWRA551.

    If all of the above checks out, perhaps there's a difference between a "blank" and a "formatted" sflash. I would need to check on that.

    -dave
  • Hi Dave,

    Thank you for the reply.  Yes, I am wondering about blank vs. formatted SPI Flash.

    When I try to Format, I still get the COM port error.  I have done this countless times and I was initially suspicious that something had taken over the COM port, as you suggested.

    I wonder if a formatted flash has a boot block with a bootloader program, whereas a blank flash does not?

    During all of this I have been able to duplicate the problem by replacing the quad flash chip on the AWR1642BOOST and certainly the problem seems to follow a completely blank flash part.

    I am not 100% certain that the flash parts are supported, it is said in this forum that they are - however, I have had scope probes on the TX and RX signals between the XDS110 and AWR1642 and see no activity, although now that I understand the "UART BREAK" prototol is simply the AWR1642 pulling the line low for one character slot, I was looking for actual TX/RX activity.

    I never thought of using Tera Term to send traffic over the USB UART and see if it is coming out of the other side of the XDS110 to the AWR1642.

  • John,

    I just spoke with another team member, and the only reasons we think this could happen are a) the sflash is not compatible, or b) it is damaged.  Have you tried taking the working sflash from the 1642 EVM and installing it on your board?  Here is part of Figure 6 from the bootloader doc that shows the supported flash types:

  • Dave,

    You were exactly right.  I thought the cross that I had was compatible, after a thorough examination, so it must be a subtle difference.  I have the AWR1642BOOST programmed using a blank MX25L3233F.

    Thank you so much, this completely resolved my problem.  Next, I will try to burn the flash on our production design.

    I will need to research this to ensure we have compatible parts for a long time.