When moving to less pins/spec then MCU in dev kit

[estersci2]

Posted on April 10, 2013 at 12:34

Hey,

I would be grateful for some help - first time moving from dev kit to own board. I have a software design based on the stm32 f4 G-Eval kit, developed using the g-eval libraries as the foundation. I am using Keil uVision 4 as my development environemnt. 1. In production, I will not need so many pins on the MCU as per the MCU in the G-eval dev board. Could you please outline what changes have to be made when you use an MCU with a lower number of pins? For example, for I2C1 do you simply select a lower MCU partnumber in the project settings and use the pins marked as I2C1 on the datasheet of that new MCU, or is there more to it? Are code changes to the library required or example?  2. Likewise, when using a chip with a lower amount of flash memory etc than the kit, do you have to make any changes in code, or is it again simply a matter of changing the MCU partnumber in project settings? Thanks

[Tesla DeLorean]

Posted on April 10, 2013 at 13:06

ST has a pin fitting app called Micro Xplorer, other people have used Excel sheets, and others just go through the data sheet/manual and confirm what functions escape on what pins, and if you need to use a different USARTx to free up pins for other things. The lower the pin count the more crowded the peripherals become. You will need to reflect the pin and board changes in the include file related to the board, or however else you are assigning and configuring pins and peripherals. The library doesn't need changes, the board support code does. Watch defines in the C/C++ tab.

In Keil you'd pick the correct device, and confirm in the target window the size of IROM1/IRAM1 configurations. For the most part there shouldn't be any absolute memory references that need changing. As far as I know all F407 parts share the same die, just packaged and marked differently.