Replies to This Discussion

Permalink Reply by jaron on April 28, 2010 at 12:32am

For me as a software guy the FEZ ("Freakin' Easy!") approach seems to have some advantages over the Arduino:
- Decent IDE
- Supports runtime debugging
- Supports threading

So what are the disadvantages of the .NET Micro Framework, the USBizi chipset and the FEZ embedded system (I know there is some Microsoft technology involved)?

• ▶ Reply

Permalink Reply by Tim Michals on April 28, 2010 at 6:44am

1) It is not real-time
2) The BSP is not open source. It might not matter for this platform, but moving it to another platform is an issue. I brought this up several times with the vendor. They are strong on not releasing the source code to the BSP and supporting libraries.

• ▶ Reply

Permalink Reply by Martin Seven on April 28, 2010 at 7:25am

I wouldn't trust it in a timing-critical application (like stabilizing a quad). With so much abstraction going on, thread scheduling with all kinds of braindead logic and garbage collectors running about, proving conclusively that the system will always have the exact same response with the exact same timing is difficult. It's better to just have a simpler chip, like an AVR, programmed "directly on the hardware" to do that kind of a task.

Might be good for higher-level thinking though, like complex navigation schemes and large volume telemetry. It's a little poor on IO pins though.

And it kinda sits on the boundary between low-level and high-level - it's too advanced and complex to do simple physical computing stuff (like the quad example above), but it's too weak to be used as a high-level processor (e.g. image processing). For the same kind of money you can get an ARM9 routerboard clocked at 500 MHz, put linux into it and use it as a general, high powered computer, perhaps in concert with an arduino.

Plus: it's not opensource. Trusting your multi-thousand dollar aircraft to what's essentially a black box feels wrong. It's impossible to unit test it - all you can do is hope that whoever wrote that software wasn't on meth at the time and that it'll work flawlessly. If you do encounter a bug in the firmware, then all you can do is tell the creator about it and wait. The board doesn't have its JTAG broken out though, so upgrading the firmware might be a challenge too.

• ▶ Reply

Permalink Reply by Gregg "Cabe" Bond on April 29, 2010 at 11:01am

All unfortunately excellent points I hadn't considered :/

• ▶ Reply

Permalink Reply by Troy Reabe on April 28, 2010 at 10:20am

The chip price is also more than the ATmega, you do get more speed but it does cost you in price and power. but if it goes full open source I can see it catching on.

• ▶ Reply

Permalink Reply by Martin Seven on April 28, 2010 at 10:58am

The author makes it very clear that releasing his code for us noobs is not gonna happen. We're being suggested to stick with our "hobby toys" if we don't like it. He really needs to work on his people skills if you ask me :-)

• ▶ Reply

Permalink Reply by Tim Michals on April 28, 2010 at 11:28am

Oh I agree 100%. That is why I'm pushing forward on my microchip PIC32 project and see if I can get microframework working with freeRTOS....

• ▶ Reply