6

I have acquired a RM48L952 that is an ARM Cortex R4F from TI and I would like to start learning ARM development with this chip. My question is can I use any JTAG programmer to program it? Something like a Open DIY JTAG Programmer or does it have to be a brand specific one? The same applies for the tool chain, can I use gcc or do I have to use brand specific compiler?

Thank you

  • Check out this article, it may be of help too. http://www.microcontrollercentral.com/author.asp?section_id=2379&doc_id=248729& – fceconel Aug 14 '12 at 13:39

3 Answers3

4

About the JTAG, you don't need a proprietary one, unless the board you have has a proprietary connector. Even then, if you can discover its pinout you could make an adapter connector. The main difference between open/free JTAG projects and a commercial product is that the later will probably have higher speeds, which makes a big difference depending on the complexity of the programs. There are cheap commercial JTAG adapters that are parallel/serial, and somewhat more expensive ones that are USB; given the choice, I'd always prefer USB.

About the toolchain, it's perfectly possible to use GNU tools, (take a look at gnuarm.com) but the downside is that you have to find documentation about how to interface with the device-specific hardware; also you may have to prepare a specific link script for the memory map of your device. The more rare/exotic it is, the less probable it is that you'll find a script already made. And these link scripts can be pretty complicated. If you use a commercial toolchain specifically tailored for your device, you'll find everything already set to make it work on the device.

I'd also recommend to buy a dev board (like this one) for that MCU; It'll be a lot easier to make everything work if you can divide-and-conquer. And having a tested, functional board is the best way to keep software and hardware problems separated.

fceconel
  • 2,659
  • 22
  • 19
  • Acquiring a dev board like that one is what I want to avoid the most because it makes the development cost rise a lot. I guess I'll have to look into those link scripts... – Leonardo Marques Aug 07 '12 at 10:07
1

You can use OpenOCD as a hardware programmer/debugger and the GNU toolchain (e.g. yagarto or GNU ARM)for the software side of things. For support look at Linux Support for the ARM Architecture and Install the GNU ARM toolchain under Linux.

suha
  • 1,774
  • 14
  • 17
0

Get the stm32f4 discovery board and work thought the examples. You will have enough on your plate without trying to design a board as well. The discovery has a USB ST-Link2 programmer built in and the whole thing is about 15 bucks. Coocox is a pretty good IDE for experimenting because its free and unlimited. Unfortunately the documentation sucks and their forum is nearly useless as nobody accept the authors seem to be able to answer questions and as they are working for free the immediacy of response is lacking. That said if you are willing to work at it Coocox is an easyish solution if you don't want to add Linux to your already long list of things to learn. ARM is complex to learn because unless you have a super fluidity with pointers and typedef structures etc you will also have to become an advanced C programmer in the process of learning the hardware. the CMSIS peripheral library is written with advanced C programmers in mind so you won't even know what's going on till you get your c skills up to a decent level. This is of course my perspective on this and may not apply.

Bruce Duncan
  • 191
  • 2
  • 7
  • I'm very comfortable with Linux and C already what bugs me is all the interfacing between code to chip in ARM. – Leonardo Marques Aug 07 '12 at 10:09
  • Yeah I hear you. I have done a lot of work with microchip devices and when I first got into ARM I was surprised at the shear volume of code that is required to do even the simplest tasks. I wish for a core that has all the peripherals and speed of ARM but the ease of use of Microchip. – Bruce Duncan Aug 07 '12 at 17:51