STMicroelectronics has joined the ARM mbed project

@ Rajesh - so even more exiting :slight_smile:

@ Gus - What would be required to use GHI OSHW boards with mBed?

The possibility of running native code on GHIā€™s powerful boards (with possible support for Gadgeteer modules) is VERY exciting!

I am not sure what mbedā€™s license is to use their services. But I did think about this before.

Interesting. It looks like the top 1/4 of that board snaps off.

@ jasdev - I agree. An industry standard framework like mBed combined with a standard module system like Gadgeteer sounds very exciting.

Time to go hit the mBed booksā€¦ Any other suggestions for best training/hardware/debugger configs to quickly get up to speed?

In the first place, MBED is Software Development Kit.

I think with a high degree of abstraction, MBED simplifies the development of simple and medium complexity devices.
Depending on the situation, but gives you the possibility to reuse the same code base on MCUs from cheap CORTEX M0 up to advanced CORTEX M4.

As an example, for a simple IR transmitter, it makes no sense to use expensive Cortex M4 MCU, but for that task is more then enough Cortex M0 for $0.85.
Another example, some device for processing and exchange some readings or data with main controller.
In most cases is more than enough Cortex M3 MCUā€™s in price range 2-5$.
But that reduce development time, reduce PCB layers, PCB size, time from prototype to final product and of course final product costs.

So I think the question with additional features as an online editor and etc. is not important, when you have a full featured IDE.
And you can add at any time any library with just 2-3 clicks. What is important, is core library updates.
But I think itā€™s possible in Eclipse to sync source code with github.

Also, in general, development with MBED SDK is not harder than with NetMF.

What I wish, I really would like to see MBED as addition to NetMF platform.
I think DaisyLink still has potential, and some DaisyLink layer or library for the MBED will simplify development of DL devices. Another option is to use RPC.

Also i wish to see a MBED based small board as Medusa , that can be with gadgeteer sockets or just headers, but can exchange data with NetMF devices by wired/wireless connectivity.

@ slawek - I see the other end of the spectrum as more exciting. A powerful board like Hydra running native code with a large choice of Gadgeteer modules would be awesome!

I donā€™t know about the economics of it, but it would open up a whole new market for GHI OSHW boards and Gadgeteer modules.

Daisylink on mbed would be quite cool, as even mbed online IDE is so much better than Keilā€™s.

I do not see mbed support for LPC1113, so which of mbedā€™s is the cheapest now?..

Yes Gus. I like NETMF and would like to see it go to the next level. However Mbed is real easy to use including networking and USB.

I have used this with the online compiler only, with no debugger but for something like building a simple app to fetch weather info from weather.com, parse and set a RGB color it was easy and I did not need the debugger.

Mbed + Mbed libraries + GHI mainboard + Gadgeteer sockets + GHI modules would be a great option.

I am not taking sides here but life is all about options and speed.

No sure if that final pricing, but pricing seems really reasonable.
Also i think STM have done a great job to support MBED on full line of MCUs.

http://www2.mouser.com/Search/Refine.aspx?Keyword=NUCLEO

If someone is interested to try MBED on stm32F0Discovery or any board based on STM32F1xxx MCU, I have uploaded two KEIL projects with basic templates and RPC functionality.

RPC can be tested with any terminal application and .Net framework library which i think can be easily ported to the NETMF.

https://mbed.org/cookbook/DotNET

2 Likes

EM::Blocks basic mbed project for STM32F0xxx MCUs.
Works fine with STM32F0Discovery board.

Because all GCC toolchains using NewLib library, the binary file output ~15K larger comparable to the Keil.
But, itā€™s free, really fast compilation, and still a lot space left even on devices with 64K flash.

I think that project is good starting point to implement DaisyLink on mbed devices.

EM:Blocks is such a sadness compared to Visual Studio :frowning: Iā€™m surprised ST doesnā€™t provide anything decentā€¦

BTW, Iā€™ve bought mbed lpc4088. Nice and powerful piece of hardware, say I must. mbed abstraction layers are also kind of cool, but the whole developing environvent is still a huge drawback for meā€¦

The purpose of that project was to be able compile STM32 mbed target with GNU toolchain.
I think that project can be easy migrated to Eclipse+GCC toolchain.

We are working on project with LPC1769 LPCXpresso board using Eclipse based LPCxpresso IDE. I can say itā€™s good as Visual Studio. Also having ā€œlow levelā€ features that Visual Studio donā€™t have.

Well then obviously you donā€™t work with VS long enough :slight_smile:

Anyway, back on topicā€¦ Iā€™ve bought 2.7" epaper display from Embedded artists yesterday, and decided to give it a go with mbed and LPC4088. I failed miserably :slight_smile: After two hours of trying and looking at 2 pdfā€™s and one Excel sheet, I gave up :slight_smile: Usual C stuff: once I figure out which pins are connected where (after half an hour of browsing), compiler starts vomitting that it cannot find something. It is those moments when I realize how good Gadgeteer hardware (and software) interconnection platform is.

Obviously, epaper display is not very mbed-able at the moment.

Weā€™ve got another small project here to read photodiode value. Maybe Iā€™ll give mbed another chance. This time with LPC1114.

STMicroelectronics Reveals Affordable, Extensible Platform for Efficient Prototyping with STM32 Microcontrollers

The STM32 Nucleo-F030R8, STM32 Nucleo-F103RB, STM32 Nucleo-F401RE and STM32 Nucleo-L152RE are available immediately, priced from $10.32 per unit.

1 Like

@ $10.12, with Arduino headers, onboard SWD, online and offline compiler support, that is good value.

@ slawek - Those new STM32 chips are dead on arrival for .NET MF. The largest is 512MB of flash and I cannot go back to such a limited amount of flash. My minimum for practical purposes is 768MB. If ST really cared about .NET MF they would have made one with more flash. Personally, the CerbuinoBee meets all my needs, but a super cheap .NET MF platform will always be good for promoting platform adoption.

I think maybe youā€™re off by a just a few powers of twoā€¦ :wink:

And few years ahead of our timeā€¦

512MB Of on-chip memory, you can run windows on chip not just NETMFā€¦ :), so I think you meant to say 512kb :slight_smile:

@ ianlee74 - Holy cow. Youā€™re right! I meant KB not MB. Ha, sometimes I canā€™t believe how little memory we are working with here.

My point still stands that these do not have enough flash.