Re: Debugger recommendation wanted

Another option is a TI/Luminary development board. It has an USB-JTAG interface, which can be used also as the JTAG dongle for an external system, ARM standard 20 pin ribbon cable connector.

The thing works fine with OpenOCD and the GNU toolset (e.g. Yagarto). I'm using a LM3S6965 dev board, but others should work as well.

A plus with the dev board is that you can test the Cortex-M3 toolset and much of the code with a known-good board before starting with a fresh design.

AFAIK the ST Cortex boards also fit that model, but with JLINK.

They are OK, but beware of Atmel boards, the J-Link is bastardized, it uses USB VID/PID outside of the standard J-Link range, and it also uses different connection points to the standard ones, so it does not work with OpenOCD.

All the low cost (~$10) STM32 discovery boards come with a built-in SWD adapter, so you can use it to debug your own projects. Works very well with openocd.

That was part of the "model" that I was referring to.

I tend to use ST processors almost exclusively these days (they happen to do what I want, they haven't let me down, and I'm familiar with ST's peripherals). I generally get one of those eval boards any time I'm setting out to use a new-to-me processor.

ST seems to really understand users needs in terms of what to bundle inside their ARM chips as well as understanding what package types users prefer. They don't gouge on evaluation boards, they are either free or very low cost. And their Cortex ARM products are very aggressively priced and they have a good distribution network. They show up at major embedded shows like the Embedded Systems Conference and ARM TechCon with a big presence while some companies like NXP aren't there at all.

They're still kind of strange in some ways. Their field support is quite wonderful with the technical seminars and development board handouts. Head office is a bit different. Once upon a time their corporate website was Flash only, couldn't be Googled, etc. Most of that has been shaken out, but when I tried to get their touch-sensing library, there it was -- Flash again. I'll try without the library and see how good their documentation is. It's a shame to hide information like that, though.

Mel.

are you sure ST uses J-Link? They call it ST-Link and the driver software doesn't look like Segger's. But I didn't yet check whether it is J-Link compatible.

I think Segger is currently the best "low cost" hardware choice since they have the best support of new derivatives and are widely supported by debugging software, even iSYSTEM winIDEA (my choice) supports J-Link although they also sell own hardware.

Oliver

AAAAAGH! Massive brain fart. Yes, it's ST-Link.

According to "The Definitive Guide to the ARM Cortex-M0" by Joseph Yiu, from Newnes:

"Supports JTAG connection and serial wire debug connections. The serial wire debug protocol can handle the same debug features as the JTAG, but it only requires two wires and is already supported by a number of debug solutions from various tools vendors."

I wonder how much the J-Link and ST-Link debuggers are just ARM 2-wire debuggers, rather than anything company-specific.

ST-Link. Not JLink. Sorry, I suffered a massive brain fart. I'm smart when I'm allowed to move slowly and deliberately -- really -- but I often screw up on the snap decisions and have to go make up for mistakes later.

No problem, works with OpenOCD.

I was going to correct you but thought there might be some higher-end ST dev board that used it or something.

I was informed that STLink works with other vendors parts (on the openocd list). Not tried it myself.

J-link is already processor-agnostic, it is not tied to a particular manufacturer.

There seems to be a new emerging standard CMSIS-DAP, I think it works a bit like ST-Link but is non-proprietary. That is, a microcontroller on the debug pod acts as a bridge between USB and SWI/JTAG, handling the low level details with lower latency than USB can manage.