external flash init timeout on S32G2-VNP-EVB

[Vicky1235]

I used S32G2-VNP-EVB board to debug flash, I found that when I debugged flash, it always initialized timeout. 

The RTD version I used is S32G RTD AUTOSAR 4.4 Version 4.0.1, and I used the example is Fls_Example_S32G274A_M7. When I debugged, the interface of Fls_IPW_InitMemories was returned STATUS_QSPI_IP_TIMEOUT.

When I used Qspi_Example_S32G274A_M7, the Qspi_Init was returned STATUS_QSPI_IP_SUCCESS, but I still can't erase/read/write flash.

[Daniel-Aguirre]

Hi,

We have replicated the behavior you are seeing under our EVB3 + S32G274 platform. We retightened the socket and we could proceed no problem, as seen on the image below:

Do you have the black plastic frame between the metal frame and the chip? This should enable the metal frame to (at an extent) uniformly apply pressure to the whole chip. Since the QSPI pins are located at the edge of the chip, we saw that even extremely tightened the example could not run. Once we added the black plastic frame and even apply less pressure (but still well tightened), the example could continue executing no problem.

Please, let us know.

[Vicky1235]

I also initialized Qspi successfully:

But I can't read data from flash, Qspi_Ip_Read returned STATUS_QSPI_IP_TIMEOUT:

I didn't understand the reason of read data timeout. Could you help me to check out the situation?

[Daniel-Aguirre]

Hi,

The "STATUS_QSPI_IP_TIMEOUT" seems related to the S32G not able to communicate with the available QSPI.

Since the EVB platform uses a socket to be able to swap chips, the pads are not soldered and if the user does not apply the correct pressure to the chip some pads could be not touching the pads. Since the QSPI pins are located at the edge of the chip, the pressure needs to be constant over all the surface of the chip.

We have experienced the behavior you are seeing, but by putting the black frame above the chip and applying the correct pressure to the socket we were able to run the example successfully.

If you have any RDB2/3 available, you should not encounter this situation, since the chip is soldered directly to the board, hence all the pads should be able to communicate with the QSPI without the variable of the applied pressure to the surface.

Please, let us know.

[Vicky1235]

Hi,

I want to get RDB2 board for S32G234, could you provide a link for purchasing the RDB2 board for S32G234

[Daniel-Aguirre · Accepted solution]

Hi,

RDB2 platform is not provided with S32G234M chip, it is only shipped with a S32G274A.

If the EVB is still ending in a NOK regarding QSPI, the problem should be either the pressure of the heatsink or the pressure not being applied equally on the chip.

As said before, we had the same behavior, but by putting the plastic washer and applying the correct pressure, we can run the example no problem.

Please, let us know.

[Vicky1235]

Thanks for your replying, I used RDB2/3, the Qspi is OK, and I can normally erase/read/write external flash. But I use EVB board, Qspi is not OK. Could you give some advice to resolve this problem. I have measured Qspi voltage, the voltage is normal.