Posted on May 17, 2011 at 13:46I want to avoid buying a commercial development PCB, as stated above. They are expensive (for example, €30 for stm32-H103 from olimex - which is nothing more than a breakout board with header pins and a usb + jtag connector), and I do not need high quality just for testing (dual layers, silkscreen, soldermask, etc). Second, I have all components and tools to make a PCB here, so it's really not worth the cost. Last, commercial development boards are made with just that in mind - development. They are big, ugly, and aren't design to be expanded or used in future designs (easily) - plus I'd still need to design a PCB at some point. Not that I'm saying they aren't useful for some, I just find them too expensive for what they offer, plus, I prefer breadboarding and hooking up wires for experimenting.
Anyway, what I want to know is what I need to do to make a STM chip programmable over USB/DFU.
a) As far as I understood from the datasheets, the bootloader is written at fabrication, so I assume I can skip the need to build a JTAG programmer and put a JTAG connector onboard. Is this correct? Can new chips be programmed directly over USB without the need of burning a bootloader first?
b) Does the need for BOOT0/BOOT1 signals still exists over USB/DFU programming? How is this implemented? As I understood it, if the chip starts with USB disconnected, it goes into application mode, if it resets with USB connected, it starts in programming mode.
c) What special precautions do I need to take when compiling code to be loaded over USB/DFU?
d) Can code loaded over USB/DFU cover the whole flash range?
Thanks.
