STM32F446, QuadSPI Bank1 CS signal

[JSaez]

Hi folks,

currently I am working on an communication with a sensor via QuadSpi using the STM32F446 controller. The QuadSpi Bank1 was chosen, but unfortunately the hardware designer did not connect the Chip Select signal (CS) to the pin that corresponds to the QuadSpi Bank1 (according to the pinout is either PB6 or PG6) but to the pin PC7. Could this be a reason for the QSPI communication not to work? Indeed, it is not working (in the oscilloscope I only see the QSPI CLk signal but the data inputs remain low). The same code works in another board where the pins were correctly chosen according to the pinout. But I thought that this issue would not be meaningful (not to connect the CS signal to the right pin).

Thanks a lot for your support

Jorge

[Clive1 (HNL)]

Sorry, what's the question here? HW guy connects wrong pin SW guy thinks some critical connectivity not important, yet still doesn't work?

Options:

Drive the CS pin that's actually connected as a GPIO, and modulated around the access code.

Drive the entire interface in a bit-bang fashion to prove it works otherwise.

[waclawek.jan]

Unlike in some other circuits like SRAM, in sFLASH (and many other SPI-connected devices) the chipselect is not merely a gate of the strobe signals (WE/OE) so that it's enough to keep it low all the time if there's no other connected device on the bus, but here it's also a vital part of the state machines governing the function, so it must be asserted/deasserted, framing each command.

JW

[JSaez]

Hi,

thanks for the reply. I just thought the QSPI CS signal could be represented by any GPIO (as for instance the SPI CS) and not by a specific one.

I will try the options you recommended.

Best Regards

Jorge

[JSaez · Accepted solution]

Hello,

Just for your information. The QSPI interface works with the selected GPIO as CS (It would with any arbitrary GPIO, I think). The error was in another place.

Thanks for your answers

Jorge