QSPI fast boot issues

[johnfielden]

We are using a Spansion S25FL127S QSPI part.  We are having difficulty on some units booting in quad mode using a 100MHz clock.  Are there any reported issues using this part with the Vybrid VF6?  It is quad SPI, but single data rate only.  We seem to be able to reliably boot in single data bit SPI mode at 100MHz, but quad seems to fail right after we command the part to go into QSPI mode.

Thanks,

John

[johnfielden]

This appears to have been an unrelated power sequencing issue.  The problem has gone away.  We are now booting reliably at with 100MHz QSPI, single data rate.  Takes under 200ms to fully boot our OS. 

[johnfielden]

We are attempting to boot from the QSPI.  There is no code for us to run prior.  When a unit fails, we see a rising clock edge from the Vybrid on the QSPI clock.  Just one edge, then a level.   No clocks makes it difficult to pull data from the QSPI device.

When it works, we get lots of clocks, but again, that is a working case.  100MHz QSPI using execute in place. 

[pato3]

That's very interesting. Pushing to the limits, way to go !!

Concretely what results do you get when you compare code execution speed between OCRAM and your XIP flash now ?

[alejandrolozan1]

Hi, one thing you could do to narrow down the issue:

Try to execute some code in RAM and force the PC to execute in QuadSPI memory or write and read to it. That way you will know if it is a HW problem.

Also make sure that the LUT is configure correctly.

Each flash vendor has specific read/write/erase commands supported, and while many commands will overlap, not all do. The LUT also determines how many bits wide a read will be, and what type, based on the sequence put into the LUT.

Best Regards,

Alejandro