Moderator Comment by John Church on April 5, 2011 at 1:05pm

Visionaries... Where would we be without them? Realizing potential and latching on to build bigger and better things is absolutely the way to go, and this community is masterful at that for sure! There are plenty who scoff at Arduino, but I bet I have more time on my home-brew gadgets than most in-the-box home engineers!  :)   (not to downplay those guys, but you know who you are, and you can attest...  ;)   )

Comment by I.S. on April 5, 2011 at 2:03pm

Happy to see the trend going sky high on that chart, and agree on all points you mentioned and glad you noticed the potential so we can enjoy APM today.

However, I would like to see Arduino also offering more powerful platforms (as ARM) as mainstream,

not just as niche solution as Maple and similar.

Comment by bcr on April 5, 2011 at 6:56pm

Popularity alone doesn't make something good. Once you're past the learning curve, the ARM MCUs are far superior to AVR8.

3D Robotics Comment by Chris Anderson on April 5, 2011 at 7:14pm

bcr: yes, that was exactly what we heard from engineers three years ago. We ignored them and are glad we did. Autopilot teams that decided to go with "superior hardware"--ARM, Propeller, Gumstix, PIC32 and the like--had to deal with complex toolchains and relatively small pools of qualified developers, which raised the barrier to entry. Many succeeded anyway due to a few brilliant people at the core (Paparazzi and the UAV DevBoard, for example), but still tend to have smaller dev teams. Others have stalled out completely.

Note that you can use Arduino without having to use their simple IDE. The APM dev team mostly uses Eclipse and other pro-quality dev tools, and most of the code is C++, not C. But it's all still Arduino-compatible, so even newcomers can work with it in the basic Arduino framework.

That said, Arduino is now extending to ARM and we are, too, thanks to the work of Roberto Navoni and others. But because we started with an easy-to-use computing platform that was more capable that it looked at first glance, we feel we were able to get critical mass easier and sooner, which will help us navigate the more complex migration to 32 bits and RTOSs with a lot of momentum achieved already.

Comment by bcr on April 5, 2011 at 7:24pm

Fair enough and well said. I do think the move to ARM will have many advantages.

Comment by Krzysztof Bosak on April 5, 2011 at 8:11pm

Exactly this is why I was amazed abotu some other autopilot claiming 140MIPS? then extra 140 MIPS for another processor just fo rthe IMU? this is something like 280.

AVR 16MHz is around 14-15.5 MIPS I estimate. So you can have as many MIPS as you want. You can have also MS Word opened on quad core 4GHz machine - the programmers become careless so easily, no matter how much you have, any processing power can be wasted easily.

Space shuttle is maybe 8x motorola 68000. But with less redundance, you could fly space shuttle on 2-3 atmegas.

1 atmega should be enough for anything less than cruise missile, if you got Engineers.

Most autopilots are RABBIT3000 which is only overmegahertzed Zilog80 with extra peripherals.

Developer Comment by John Arne Birkeland on April 6, 2011 at 12:21am

The 8bit AVR is good enough and easy to use. But I would like to see a Arduino priced IMU with true EKF filtering, and that sadly is to much to handle for the 16mhz AVR.

Comment by I.S. on April 6, 2011 at 2:14am

Recommended this nice article on Sparkfun Apollo landing with 1 Mips

Any nice tutorial for working with Arduino using Eclipse or AVRStudio?

For the sake of curiosity how many MIPS are left nowadays on the APM?

Moderator Comment by Roberto Navoni on April 6, 2011 at 2:24am

In starting stage of developing Arduino could be a good choose , but after some development became a limit.

infact the code is around up 100 kbyte now and standar Arduino UNO have only 32 kbyte of flash .. only

So now with Multipilot32 that use STM32 is possible to continue the development of AutoPilot Project on ARM M3 at 72 Mhz with 512 Kbyte of Flash and 64 kbyte ram.

Sure Arduino have a good and huge community and good Brand , the arduino core team are doing a great job :)

What was possible to do on AVR platform are doing.Is not only a problem of speed but also on the bus interface for example. So if you put more device on i2c for example  it freeze ... i doing a lot of test on this limit but i cannot solve it . So in my new design i use a micro with 2 i2c bus one for ESC and other for sensor.

So really the DIYDrones developer don't use only Arduino framework to solve the problem of performance on imu , but develop a lot of good library that using directly the native code of AVR processor. Don't forget AVR Freaks community that was usefull for this years on development on AVR micro processor. http://www.avrfreaks.net

Best

Roberto

 

Comment by I.S. on April 6, 2011 at 3:05am

Robero, I don't think the point is on flash memory with large & chip memory chips and Atmega in 256K flavour (the coming APM board), but the processing capabilities, not only mips, but also DMAs ando others.

And don't want to leave without mention all your good work in your ARM based APM. Keep it going.

 

On the other hand, no one thought about migrating to Xmegas? It would be the closest to the current Atmegas (AVR core) and doubles de mips (by doubling the MHz up to 32) and has DMA, and so on, so more than twice the speed I think.

Comment

• ‹ Previous
• 1
• 2
• Next ›
• Page