2022-03-09 06:55 AM
We are new to HyperFlash and OctoSPI and it seems that working through the Datasheet and write my own driver from scratch will be very time consuming... Does anybody have a driver or sample code on how to erase / program a HyperFlash (S26KL512S) with Octo-SPI interface? We use the STM32H7A3 processor.
Any help is very much appreciated!
BR GS
Solved! Go to Solution.
2022-03-11 02:39 AM
Thank you for your input, Tesla DeLorean!
Yes I already contacted Infineon, and in fact they did recommend a driver package. However this LLdriver supports I guess 20 different chips and composes out of hundred of functions and reference.... Not easy to implement a straight forward solution and is also very time consuming. Maybe we might better switch back to the Quad-SPI chip and use my ready to use available software solution for that one. Implies a small Hardware change, but is maybe faster ready to use.
BR GS
2022-03-09 01:00 PM
Have you tried squeezing the other end of the tube?
The proponents of the HyperFLASH / RAM standards would seem to be the guys you need to motivate, working demos and plug in boards on STM32 platform(s) to illustrate their viability to much larger customer bases is a way to get their sales up.
ST's likely dumped in some third-party IP to implement the standard in silicon, and unless something gets designed into a DISCO board from Vendor X, probably not much motivation to blindly code a more extensive framework.
I know this "first guy up the mountain" problem is annoying, but I suspect the memory guys are perhaps better equipped, and know the pit-falls and timing nuances of their own devices. And be the ones to test-fixture the logic analyzers up to wrangle the issues.
Adding context so I can find it easier later..
2022-03-09 01:06 PM
S26KL512S doesn't look easy/cheap to fixture, would need a custom PCB interposer to plug into a NUCLEO-H7A3
2022-03-11 02:39 AM
Thank you for your input, Tesla DeLorean!
Yes I already contacted Infineon, and in fact they did recommend a driver package. However this LLdriver supports I guess 20 different chips and composes out of hundred of functions and reference.... Not easy to implement a straight forward solution and is also very time consuming. Maybe we might better switch back to the Quad-SPI chip and use my ready to use available software solution for that one. Implies a small Hardware change, but is maybe faster ready to use.
BR GS