choosing right ARM-based Chip

[rezaf]

Hi,
Please help me for choosing the right Atmel's ARM-based chip (ARM9 or ARM11) for my project.
the main important features and peripherals that the chip must have is :
ability to run OS, like Linux or Android
cost-effective
QFP package (Not BGA) (this is the main problem for me to choose one)
with camera interface
audio interface for connecting MIC and Speaker
TFT LCD interface
Ethernet 10/100
also with drivers needed for all this features to install in OS.
also it will help me to start if the chip has evaluation kit.
Thanks,
Best Regards.

 

[bsprajc]

If your not set on Atmel and you don't need to manufacture your own PCB, check out the beaglebone black. It has everything (+ HDMI )... except for speaker and mic input.. you can build an adapter for that, or, plug in an external USB sound "card" into it's usb port. It boots linux and can run android and is $45

 

[rezaf]

Thanks, I don't have set on Atmel and can choose from every company but I want to manufacture my own PCB and can't use BGA chips.

 

[bsprajc]

Thanks, I don't have set on Atmel and can choose from every company but I want to manufacture my own PCB and can't use BGA chips.

Freescale has ARM5 in a QFP, maybe they have an ARM9 in that package

 

[rezaf]

Freescale has ARM5 in a QFP, maybe they have an ARM9 in that package

Yes they have. I found : Freescale (i.MX233 series), Cirrus Logic (EP9302), Allwinner Technology (A13 cortex-A8) .
now I go to research about this chips.

Very Thanks "bsprajc" ,
Best Regards.

 

[MrEd]

If this message is moved to a separate thread, please call it: Moving up to ARM, chosing between TI Tiva and ST32-F4.

I am taking the step up, I have decided, but I have the same problem as you Rezaf. But I have questions from another angle.
At the moment I am choosing between TI TIVA (the Launchpad Eval board) and the ST32-F4 (STM F4 Discovery Eval board). They seem to be pretty similar. The ST32-F107 chip has Ethernet, the 123 Tiva doesn't though. i have both Evaluation boards and the whole thing is not about how much an evaluation board cost even though they were similar in cost.

One of my quandries are regarding all the function on each pin. I read somewhere that this can be mapped to move functions to wished pin but it is very tedious. ST has a plug-in for this for the Eclipse IDE. But I will not use Eclipse, so how would I do this no, "by hand", that's probably a weeks job considering the number of pins and functions on each pin. I don't know yet how TI solved this on the Tiva series.

Another issue is, how well is everything documented? I have started downloading PDF files for both the Tiva 123 and the ST-F407 and the PDF files are kind of 2000 pages on average to get an idea, some are smaller some are larger. To make the right choice I should at least skim through all the documentation, especially the Errata sheets, then i need to know about libraries. Some kind of operating system, like RTOS, which I read about a million times but really don't know anything about. I think a lot of viewers are in the same kind of dillemma. We want to take the step from PIC, AVR, whatever to the Next level. It is all the same concept, right? It is just a lot more of everything and faster. but who has the best support, or in other words, who laid it out in such a way that one doesn't dig oneself into really big black hole, wishing i stepped down to a multiprocessor system of Atmel 8-pin tiny processors...Better the devil you know...But I am ready to face the bull and prepared to spend half a year on getting something with a TFT display, MMC/SD card, customer firmware update somehow in an easy way, general IO like ordinary pins, uarts, SPI, 12C. etc.

Compilers is the next issue. Keil. IAR. MikroElektronica looks tempting with TFT designer and even Basic for ARM. Anyone actually made a commercial product by using their tools. They seem to focus on education so it might be more of getting peripherals going, make a project complete with no multitasking and show the teacher they managed to get it together, but the students are not creating commercial products. Any feedback on the MikroE compilers would be very much appreciated. Considering the ST32-F107 have 1Mb Flash, the compiler will compile a lot of code. I have read the forums at MikroE and they seem to have a few bugs in their compilers, anyone with any experience of this?

Even with many years of C programming experience of commercial products I came to a total standstill after installing TI's TivaWare enormous installation. Actually Ghosted back my C drive after that to study it better before I do it again. IAR and Keil MDK are other options, as well as MikroE as i already mentioned. Anyone got any experience? Just for easier visualisation, let's say i want to make a cool talking, half oscilloscope, voltmeter with a TFT display. And let's say i wanted it to be able to log and show the results via USB on a PC or store it on a USB-stick. I am not designing that but it sums it up nicely what I want to do.

Could you share a bit of your experience, having completed a commercial product? Seeing now what you would rather have done differently? Where did you get stuck? And did you ever get so stuck that you were close to give it all up? Or maybe, was it all just too much, and you gave it all up? Any experience is welcome. I read a lot of how to do and get started, but read nothing about real commercial products, or kind of how to think, or what kind of calming pills to take when the code is approaching 1Mb. I mean 1Mb is a lot of code. A lot of possible madness too possibly caused by a tiny miss print in the documentation or a bug in the chip. I mean, i know the endless nights from crazy stuff in 8-bit processors, finding out strange things with the compiler, and other times really strange things in the chips and finding a work around. The ARM's are so much 'bigger', so I want to make a good decision of which one to begin with. And i think the decsion doesn't lie in the chip but in the development tools and the documentation.

Right now i am looking for some kind of wizard, like the Codevision Wizard, which is really great setting up all the chips peripherals and one is up and programming within minutes. Doing all the chips init for the first time on a new chip is really to ask for trouble, and pin mapping on top of that.

I might sound negative, but I know how easy things are when all works as it should, and how extremely difficult things becomes when something is not working as it should. That is the time we never get paid for! Now i have a feeling i won't be paid for a long time, but if i get to master it. I will charge!!!!

MrEd