Jump to content
43oh

Free STM32F4 discovery kit in US/Canada


Recommended Posts

  • Replies 69
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Popular Posts

http://hackaday.com/2011/10/17/how-to-d ... k+a+Day%29   This article may be of interest to some.

Here's a start. I'm using Ubuntu 11.10 (just released) and it appears that the Universe repository already has an ARM package available for GCC 4.6.   sudo apt-get install gcc-arm-linux-gnueabi

STM32F4 discovery kit giveaway  

Posted Images

Thanks.

It would be separate, not related to 43oh. I would hate to loose people to ST though. But at the same time, I would hate to miss on new cores.

 

ST's hardware is fine, but their tools are not really accessible to a whole hacker community like AVR, or MSP430. Just getting by the debugger alone is a pain, since it's not cross-platform friendly. Through mspgcc and mspdebug that support is really brought to the masses of MSP430 users. I understand that TI has at least contributed to these projects in some limited form or fashion. And the code-limited versions of CCS are good for many users as well. The same goes for AVR. Processing / Wiring / Arduino make the hardware accessible to hobbyists due to the open tools. TI is trying to replicate that with Launchpad, to some degree of success, through highly-accessible hardware, a free IDE, and community resources.

 

To my knowledge, ST has nothing of the sort.

 

There are so many ARM Cortex-based products out there, you couldn't target them all with just a site devoted to ST. At that rate, why not just have a site devoted to all MCUs. AVR? PIC? ARM? Coldfire? SuperH or any number of Renesas' hundreds of cores (I joke). The list goes on.

 

I'm no expert on ARM, but I'm not sure that one manufacturer's product is all that radically different than the next. They all just tailor the microprocessor for a different application. Peripherals are the main difference. M4 is an M3 with DSP functions, among other things. Technically, we already have that in Concerto. I almost think that TI's new Cortex M4 product is just a "me too" product, where they really want to sell Concerto instead. But maybe they are just intended for different applications. Concerto seems to be solely aimed for Industrial automation and power applications. Whereas Stellaris M4F seems to be more general-purpose.

Link to post
Share on other sites

ST's hardware is fine, but their tools are not really accessible to a whole hacker community like AVR, or MSP430. Just getting by the debugger alone is a pain, since it's not cross-platform friendly.

 

If I understand correctly, all of the commercial IDEs for the STM32s use the GNU toolchain at the back end. What is at issue is the ST-LINK for programming/debugging that is proprietary. It does seem as if protocol has been reverse engineered, and an implementation is available for Linux at:

 

https://github.com/texane/stlink

 

Alas, I do not believe this will work under OS/X, my preferred development platform.

 

Through mspgcc and mspdebug that support is really brought to the masses of MSP430 users. I understand that TI has at least contributed to these projects in some limited form or fashion. And the code-limited versions of CCS are good for many users as well. The same goes for AVR. Processing / Wiring / Arduino make the hardware accessible to hobbyists due to the open tools. TI is trying to replicate that with Launchpad, to some degree of success, through highly-accessible hardware, a free IDE, and community resources.

 

Unfortunately, if TI is trying to replicate the success of Arduino with the LaunchPad, they are completely clueless as to how and why the Arduino has succeeded.

 

Don't get me wrong, the LaunchPad is a great product and would be a good deal at $20 + shipping. TI is practically giving them away at $4.30. In fact, second day shipping on a single unit is probably more than $4.30, so TI is taking a loss from the shipping alone!

 

I have a sneaking suspicion that there are a bunch of people at TI who think that because they are giving away such a great tool that it will translate to Arduino-like success, but they couldn't be more wrong. That's a part of the reason why I am concerned for the lifetime of the $4.30 LaunchPad, and why I'll continue to stockpile them while I still can.

 

And if anyone from TI reads this, contact me offlist -- I'd be glad to give you a free clue on how the open-source hacker community *really* works, and how to properly leverage it.

 

There are so many ARM Cortex-based products out there,

 

Which is a good thing, because the competition and volume production has made them amazingly affordable.

Link to post
Share on other sites

Unfortunately, if TI is trying to replicate the success of Arduino with the LaunchPad, they are completely clueless as to how and why the Arduino has succeeded.

 

Don't get me wrong, the LaunchPad is a great product and would be a good deal at $20 + shipping. TI is practically giving them away at $4.30. In fact, second day shipping on a single unit is probably more than $4.30, so TI is taking a loss from the shipping alone!

 

I have a sneaking suspicion that there are a bunch of people at TI who think that because they are giving away such a great tool that it will translate to Arduino-like success, but they couldn't be more wrong. That's a part of the reason why I am concerned for the lifetime of the $4.30 LaunchPad, and why I'll continue to stockpile them while I still can.

 

And if anyone from TI reads this, contact me offlist -- I'd be glad to give you a free clue on how the open-source hacker community *really* works, and how to properly leverage it.

 

I'm not quite sure that there is more that they could do to leverage it. Open the FET source code? Write an IDE that's essentially a poor man's text editor that hooks into a compiler? Someone could just as easily adapt Processing for any target.

 

Recognize that Atmel is not responsible for the success of the Arduino. It just happens to be an MCU that's on the board. But you can see that there are a hundred clones out there that use the form factor but implement numerous architectures. For all intents and purposes, one could pretty easily adapt the form factor for an MSP430.

 

TI is trying to attract the hobbyist to their chips. Those could equate to real-world sales later on. Frankly, I think that they are successful and that many people in the MSP430 division are interested in the open-source hacker community (which is why they are targeting it). And since their sample program is also pretty liberal, it's a lot more attractive to some people than buying an AVR or an Arduino. Anyone that cares that much about "open-source" / free as in speech & beer is already going to be able to figure out how to get GCC working on their machine. The rest is just about accessible hardware (price and ability to make it "work"). The mere fact that they have so many parts available in DIP can only be to keep the hardware accessible for prototyping. Does anyone really use DIP for anything else anymore? The socket alone adds extra cost in fabricating a board when dealing with large quantities. Pretty soon it will be tough to find an Atmega328 Arduino with DIP, since Atmel's priority is SMD.

 

It would be stupid for them to give up on Launchpad. Considering that it practically uses all TI parts (and few of them at that), even taking a minor loss at $4.30 or breaking even has the potential to get them future sales in hardware, IDE sales, or more expensive development kits. I now own around a dozen TI development kits thanks to Launchpad. And I'm pretty sure that I am not the only one. TI also sends samples of a vast majority of their parts for free, express shipping included. I don't think that they or any other manufacture are cancelling that program any time soon. The same logic applies to Launchpad. In that case, you are getting a sample and you're essentially paying for Shipping. Now buy 4 Launchpads. Pack them all in 1 box for $17.20. Get the idea? Most people buy multiple Launchpads in one-shot. At $4.30 a pop, it's an absurd deal compared to almost anything else available.

 

If you've seen my posts here, it's pretty apparent that I am a big fan of open-source, and I've been an avid Linux user for almost 15 years. I personally find Launchpad and the MSP430 to be just as accessible as an Arduino for my purposes... Moreso, in fact. The corporate aspect of the Launchpad vs. Arduino's grassroots approach doesn't really impact me. I need a tool for a certain application, and MSP430 is more flexible and less expensive in almost all respects.

 

I think that a lot of people at TI *do* understand how the open-source hacker community works, because they *are* members of the open-source hacker community. Take a look at the people who designed it, and wrote the software examples. They are the members of the engineering and marketing teams. Just read source code to the documentation and Processing examples that ship with the Launchpad and FRAM boards.

 

Back to the ST Discovery though... Even though we have solutions like the ST-Link software, it's still annoying to have to jump through hoops to make it work by reverse engineering the emulator. And it likely doesn't work with newer ST-Link V2 boards yet (I haven't confirmed). Sure, there is JTAG. And the emulator could also be reprogrammed with Versaloon or similar. Maybe even OpenOCD.

Link to post
Share on other sites

You are missing the big picture in exactly the same way I believe TI is missing the picture.

 

[ warning: the following is a poorly worded response that barely touches on all of the issues involved. I hesitated before because I do not have the time to properly express all of my feelings on the issue. But for what it is worth, read on... ]

 

The Arduino is popular because Atmel chose to support gcc as a first-tier development platform without any code size or optimization limitations. This attracted the hobby market away from PICs (which already had "free as in beer" tools) and led to the development of a substantial code library. This in turn attracted more users, creating a feedback loop. It got to the point where there was so much code that it was difficult to keep track of it all, or use it together. Then someone got the bright idea to develop a system to make all of these libraries play nice with each other, and even better, do it in such a way that that it was as easy to use as the BASIC stamp. Even better, it was cheaper and faster than the BASIC stamp in commercial form. It was *much* cheaper when you rolled your own.

 

But it all started with a free (as in "beer" *and* "speech") unrestricted toolchain. That is what is key. *Everything* about the Arduino's success follows from this.

 

If TI wants to have any hope of even making a dent here, they are going to have to get their head out of their ass and realize that they are a HARDWARE company, and not a software tools company. I am not saying that TI should ditch CCS/CL430, as this product line is absolutely necessary to support their commercial users. Here is the thing that so many people don't realize about commercial tools: Commercial/Industrial users are not paying for the tools,they are paying for the SUPPORT they receive with them.

 

It is well worth noting here that one of the reasons why MicroChip is so much more successful then Atmel in the commercial realm is because their support. One of the things I hear over and over again from professional engineers is that MicroChip's support -- both for their hardware and development tools -- is truly top notch. I have also heard that Atmel's AVR division has yet to turn a profit. Apparently, they just can't attract the volume developers -- the volume developers who demand first-class support for their tools and hardware.

 

I say this to point out that the availability of a no-cost toolchain has absolutely no bearing whatsoever on the sale of a toolchain with commercial support. If TI *really* wants to get into the hobby market, they *must* embrace gcc and an open toolchain. TI can easily embrace gcc without providing commercial support, and without effecting sales of their commercial tools.

 

At the present time, gcc for the msp430 rests *entirely* on the hands of a single person: Peter Bigot. Were it not for his efforts, none of the chips TI has introduced in the past two years would have any support in gcc. It is unclear whether Peter has a client funding this development or he is doing it on his own, but in either case the long-term future of gcc for the msp430 is tenuous at best. As a result, most hobbyists are just not willing to dedicate the time required to learn the MSP430, and are instead looking at other options like -- you guessed it -- the STM32.

 

The only item that is unclear is just how TI should embrace gcc, and of that I am unsure myself. One short term measure is absolutely clear: TI can hire/contract someone to generate proper documentation for mspgcc. The current mspgcc documentation is dated 2003, and very little documentation was generated as a result of the now-dead mspgcc4 project. Peter Bigot has all but stated that he will not be generating any documentation on his changes to mspgcc, but he does seem exceedingly open to answering questions.

 

If TI makes a commitment even by finding some way to get the current mspgcc properly documented, it will go a *long* way towards opening doors in the hobbyist community.

Link to post
Share on other sites
It is well worth noting here that one of the reasons why MicroChip is so much more successful then Atmel in the commercial realm is because their support. One of the things I hear over and over again from professional engineers is that MicroChip's support -- both for their hardware and development tools -- is truly top notch.

 

I have dealt with Microchip for ~20 years as a hobbyist, technician, engineer, and consultant. In all cases the support I got was very good. They don't ignore the little people as Atmel seems to do. The Microchip forums are open to everyone from noobs to pros and support there is generally very good.

 

The latest dev tools from Microchip run on Linux, Mac and Windows. For corporate use there is no need for Linux and certainly not Mac - but they did it anyway - they value all their customers.

 

The Microchip compilers for the 16 and 32 bit parts are GCC. Microchip did the port and offers paid support if you want it. Source code is of course available if you really want to do your own build.

 

So there is no need to nag Microchip to support Linux and Mac - they have that. No need to demand a FOSS compiler - they did that too.

 

Everything I have done with MSP430 so far is just a re-implementation of what I have already done with PIC. In most cases a 8 bit 18F series part can perform as well as the MSP430. The 16 bit parts are typically 2 to3 times as fast as MSP430 - in some specific cases much faster. The performance is excellent for the specific class of microcontroller.

 

Availability of PIC parts is generally good, and the sample program is reasonable (you may have to pay a small fee sometimes). Many of the low pin count parts are available in DIP package.

 

There are many compelling reasons to choose PICs.

Link to post
Share on other sites

I say this to point out that the availability of a no-cost toolchain has absolutely no bearing whatsoever on the sale of a toolchain with commercial support. If TI *really* wants to get into the hobby market, they *must* embrace gcc and an open toolchain. TI can easily embrace gcc without providing commercial support, and without effecting sales of their commercial tools.

 

I'm not missing the point at all. And, I still don't think that TI is either.

 

viewtopic.php?f=8&t=1561&p=10486#p10486

 

Read Peter's comment on the issue. But note that Peter does mention a few issues that TI still needs to work out, as we've discussed several times in the forums. Still, the RS-232 issue is not an MSP430 problem particularly but is an issue with Launchpad (and similar) boards with the emulation hardware.

 

I still stress that people should use the right tool for the job. And I think that the growing number of people on these forums seem to indicate that MSP430 is the right tool in many applications since it's accessible with regard to both hardware and software.

Link to post
Share on other sites

Read Peter's comment on the issue. But note that Peter does mention a few issues that TI still needs to work out, as we've discussed several times in the forums.

 

Yes, I have read Peter's comments. My interpretation of them is that TI clearly doesn't understand just what a huge favor he is doing for them.

 

I still stress that people should use the right tool for the job. And I think that the growing number of people on these forums seem to indicate that MSP430 is the right tool in many applications since it's accessible with regard to both hardware and software.

 

I believe there are growing number of people on this forum is because it has achieved "critical mass." If there were really a lot of interest, 430h would be seeing more competition. And really, aside from a few low-volume mailing lists and TI's forums (which are rather a joke) there simply isn't anything out there in terms of msp430 community.

 

But yes, I am all about choosing the right tool for the job. Unfortunately, the MSP430 really only has one raison d'etre: low power applications. And unfortunately this comes at a price premium.

 

However, the MSP430 is also an *excellent* choice for education. The architecture is exceptionally clean. 27 instructions and a fully-orthogonal instruction set make the instruction set a breeze to learn. Lots of registers and 16 bit operands keep code tight. Von-Neumann architecture and memory mapped I/O are much more similar to desktop PCs that most other MCUs that use the Harvard architecture. The peripherals are well thought out. The low-power model is highly conducive to event-driven programming. And finally, the data sheets are well written.

 

But unfortunately, there is almost no good introductory material on the MSP430. TI's example code is a joke. There is only one good book on the MSP430 -- the Davies book. The learning curve for someone who has never used a microcontroller before is incredibly steep. I can't count the number of people I have seen drop into #43oh in a state of utter confusion and then just disappear.

 

Where the Arduino is full of win is that it is accessible to newbies. You can get up and running with Arduino in almost no time at all. There is plenty I dislike about Arduino, which is why I don't use them, but there is no denying this is one area where Arduino truly shines.

 

So this is yet another area where TI has failed. They have the chip most suited to education, and they have reduced the cost of the development environment to not much more than the cost of a Happy Meal, but they simply don't have the tools and materials to get people hooked before they give up and go away frustrated.

Link to post
Share on other sites

wow!

lots of good stuff here on the STmicro board. ;)

 

I ordered mine on the 7th and got it on the 11th, never got a shipping confirmation.

 

I would love to use open source software with this thing, but talk about overwhelming!

 

If a noob to micro controllers and C programming can put in his 2 cents here's mine :

 

I used to search the interwebs looking at embedded electronics thinking how I made an error in my choice of

professions, I mean how cool are these little gizmos that you can program to do stuff, I mean all kinds of stuff.

So finally after I got over being intimidated by them and trying to figure out where to jump in I went for it.

My first micro was the dreaded arduino I got the duemilanove a couple of years ago. I thought wow this is cool and

the community was also pretty cool, but the part that wasn't cool for me was (and this is my opinion only) it seemed like

only a handful of people on the forums really knew what was going on with the actual chip. To my own demise I like to

know how things work on the bottom level. Then I went ahead and bought a couple of ATtinys and a programmer.

(the adafruit one). I downloaded WinAvr and to be honest was overwhelmed by it. Command line interface what? I didnt

know where to begin and I need a gui so then I found out about eclipse with the avr plug in and after along night I had some leds blinking on my attiny. I thought this is cool but still not so inspired that I wanted to stick with it. Shortly there after I got wind of the launchpad and I dont know why but I have always liked TI so I got a couple of them. Right from the get go I liked the CCS eclipse based software, unlike the arduino it has a real working environment, a debugger ect ect. Then I found 43oh, I think I was the 50th person to join the forum, and it just took off from there. I have learned a ton of stuff because of the launchpad and this forum that I just couldn't get from arduino or avr. Maybe I just didnt dig deep enough with the arduino or the avr community, but I dont really want to go back now I want to go forward. As far as TI's stab at the hobby market well it worked for me and the samples, I haven't had to buy anything, except for the initial purchase of the launch pad itself, oh I bought a couple of the touch shields too but thats it.

 

Maybe Im of a different breed then most beginners but I like knowing why the chip is doing what its doing, or why a certain piece of code does what it does, not just digitalwrite and now the led is blinking.

 

I dont say too much on the forum but while I was reading this I felt compelled to give my noob view.

 

If anyone has any insight on getting the stm32f4 discovery board going with open source or any free software

it would be appreciated.

Link to post
Share on other sites

If anyone has any insight on getting the stm32f4 discovery board going with open source or any free software

it would be appreciated.

 

I know that I am. And it sounds like rockets4kids is too.

 

I use Linux, but as rocets4kits mentioned there are alternatives like stm32flash:

http://stm32.spacevs.com/index.php?opti ... Itemid=103

 

http://code.google.com/p/stm32flash/

 

I am not sure that it supports STLINK V2 though. I'm debating just reprogramming mine with Versaloon or similar. It's just a matter of getting the time to do it though.

Link to post
Share on other sites

Here's a start. I'm using Ubuntu 11.10 (just released) and it appears that the Universe repository already has an ARM package available for GCC 4.6.

 

sudo apt-get install gcc-arm-linux-gnueabi

 

I had looked at GCC 4.6 a week or two ago and it looked like the core was supported. The cortex-m4 appears to be a valid target with -mcpu. I'll post more details as I progress (compiling a project and actually getting it to load).

Link to post
Share on other sites

Question please, guys?

 

It seems the Attolic TrueStudio IDE is based on Eclipse just like the CCS IDE. If I install TrueStudio on my Win-7 laptop will there be any conflicts (registry or elsewhere) between the Eclipse which is part of CCS and the Eclipse which is part of TrueStudio?

 

TIA. Regards, Mike

Link to post
Share on other sites

I seem to be having some luck with Texane's stlink. It seems to support st-link v2 as well. https://github.com/texane/stlink

 

There are a few requirements to build it (Ubuntu 11.10).

 

sudo apt-get install libsgutils2-dev
sudo apt-get install libusb-1.0-0-dev

 

The St-link software can be installed as follows:

 

git clone https://github.com/texane/stlink stlink.git

 

The README and the tutorial.pdf detail how to build it. it's pretty simple if the above prerequisites are met. You will also need to copy the udev rules to /etc/udev/rules.d/ and restart the udev service, as detailed in the README.

 

Afterward, a device node will be available at /dev/stm32l_stlink2 when the board is connected.

 

You can connect to the device with:

 

./st-util port /dev/stm32l_stlink2

 

GDB can connect with:

 

target remote :4242

 

At this point I've made some modifications to the LED example and it compiles just fine, but haven't yet loaded it. GDB seems to connect but gets a write failure. I suspect that it's address-related and I need to work on it a bit more.

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...