next topic »

[Reply #60 on:]

Just discovered one more creature:
http://microcontrollershop.com/product_info.php?products_id=4380

Would be nice to hear review from people with hands-on experience, is there is something they didn't mention on the web?

Well i got the board, and it has potential with all those extra specs such as on board sd , Li-on battery charger, much lower power consumption that just about any board, how ever it is suppose to be a "Maple clone/compatible" which for the most part it is, however
the silk screening on the board uses an Arduino format, A0-5 and D0-13, but the Maple uses D 0 -14 and then 15 - 21) and that throws everything off, because when you use the Maples IDE and lets say you want to set pin14 on your boards high but pin 14 on the board and what the IDE sees as pin 14 are not the same due to the labeling discrepancy , so that is annoying as hell. Aside from that these people took pin 21 and used it for the CAN driver , with no female header for the pin , and sucks being that pin 21 is the data line for one of the I2C ports.




Now the Maple issues, it seems, TO ME, these people try so hard to make the board arduino compatible that, while it does help them sell more, it cripples the chip which has at least 20 more GPIOs then what they actually  break out with the board to keep it arduino compatible.
I2c is suppose to be pretty straight forward and as i was trying to interface it with an external eeprom  it was not working and i know my code and wiring was right so i went to their forum and on of the "Key Masters" which i think are the developers came at me with "the wire library is... okay, but not super reliable."
and that little statement goes with just about everything on their board.


if all you want to do is blink leds then that works but  every nice little "feature" offered comes with "is... okay, but not super reliable."
I dont see why they are so stuck on making it arduino compatible and porting all these libraries instead of working with the chip they have  and taking it for a nice good run .
I am sure that stm32 is more then capable of i2c what can be so difficult about opening sending and ending a transmission?  they really are crippling the poor Arm

so then i figure i dump this clone and go for a real maple..but.. have you seen their Errata page? stuff like pins being accidently wired to reset making them unusable , pins forgotten to be broken out, having to manually cut out traces etc...

[Reply #61 on:]

The Mapple is strange...
They are doing everything from scratch because the CMSIS librarys are to slow LOL.

[Reply #62 on:]

But with the CMSIS, it's easier to port across different vendor.

[Reply #63 on:]

Isn't CMSIS the equivalent of AVR #defines for port values except it's done as structures of pointers?

I can't see why it would be slow or are there functions that map logical to physical pints like the Arduino?

______
Rob

[Reply #64 on:]

The Mapple is strange...
They are doing everything from scratch because the CMSIS librarys are to slow LOL.

Can you provide a citation for this? Maple is trying to copy Arduino, so they are probably just trying to dumb down the code.

[Reply #65 on:]

While we're talking about 32-bit boards, have a look at this one

http://www.kickstarter.com/projects/synapsesg/axon-mcu-the-future-of-net-hardware

______
Rob

[Reply #66 on:]

Hummm, I don't like it, .NET is not real-time and really slow. So a 400Mhz processor will not get you far, unless you only use assembly-made stuff (can't recall the name). I tested it and couldn't get anything done, as most of the stuff I create requires real-time processing. :-\ The thing is slow as heck and the Microsoft software is a total bloat... but heck, that's me... ;-)

Wk

[Reply #67 on:]

I agree, I'm currently trying to figure out how to do about 5 things in < 1uS, I don't think anything based on .NET will have a chance

I think it (and others we've been talking about) are a sign of the times though, essentially the only issues are memory size and processor speed, and we all know what they do every year or so.

I'm currently designing a board that needs as part of the app a 24-bit comparator and a state machine, it has to be fast but I can live without hardware speeds, so I've just added a 50MHz 32-bit processor to do the job. At $2 why not?

______
Rob

[Reply #68 on:]

Hi,
This is Terry Massey one of the developers behind the aXon-MCU. I first want to thank Rob for bring up the project here. Secondly I would like to offer some information about the aXon-MCU. While we are designing it to run the .net Micro Framework, It is utalizing an Atmel AT91SAM9G45 Which is a ARM processor core. It is also capiable of running such realtime Operating systems such as FreeRTOS http://www.freertos.org/ or even raw code compiled for the processor which would be realtime. While we are focused on the .net Micro Framework we are happy to offer support to any open source group wanting to work with the board. We also belive that it should be capible or possible to even use the Arduino code now that they have ported to an ARM core for the new Duo Board. Every project has a place and while some require realtime others do not. Our goal with the .net Micro Framework is to bring one if not the largest populations of developers over to the embeded world. While the arduino code base and IDE are very powerful and has it's place which we develop shields for it and the power of the aXon-MCU is only meant to open doors for developers who would like a smooth transistion to low level hardware/software development.
Thanks,
Terry Massey
President
TheShieldStore.com

[Reply #69 on:]

Hi Terry

At the price point you have, I'm interested in Ethernet features with a RMII or MII, USB OTG, and native SD bus. It looks like the chip itself have all of these features, will these features be supported in the software?

Using the MII or RMII might be challenging to some hobbyists, do you plan on making Ethernet PHY breakout boards?

How realistic would it be to make sort of micro-ASP.NET server project?

Are you aware that there's something else called the "Axon" on the market? http://www.societyofrobots.com/axon_select.shtml

[Reply #70 on:]

Frank,
As far as the Ethernet features are concerned we will have software support as well as a shield of sorts; I kinda hate calling it that since that is more an arduino term but a stacakable board that is designed to provide not only 10/100 ethernet support but also uSD storage onboard. We cant really release or finalize that untill we are confident that the aXon board is released. I was unaware of that board untill earlier today when I was notifed about it. I am really shocked that my google searches didnt find that name months ago when we are comeing up with the name for the board. I unfortunatly am unable to change the title of the Kickstarter project once launched. And We will likely have to do some thinking about that issue as we near launch of the project. As far as an HTTP server there is some great code on the netduino forums I have tested with the .net micro framework  on a netduino board. ASP.net support might be possible I have offten thought of creating a slimed down MVC Managed API for the .net Micro Framework. The nice thing about it is the apache license it is released under. There is so much potential for the .net micro framework.

On another note with the real time I just thought about. The delays in the .net mIcro framework are there in the slower processors and I am sure to some degree it will exist in ours. However native code drivers with a managed wrapper can resolve many of thoes issues.

I also dont want hijack this forum with a bunch of questions about my board. I really just wanted to respond to the concerns originally addressed. I am happy to answer all the questions anyone might have but I think the Kickstarter project page or even private messages would be a better forum.

[Reply #71 on:]

They are doing everything from scratch because the CMSIS librarys are to slow LOL.

Can you provide a citation for this? Maple is trying to copy Arduino, so they are probably just trying to dumb down the code.

Starts here: http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1274652722/173#173

My impression was that it wasn't so much that CMSIS was awful, or that the arduino libraries were awful, but that there was significant disconnect that putting the Arduino API on top of CMSIS was a problem...