I'm new here, and overall pretty new to the embedded tech space. Learning fast as they say. I've been searching all over the place for how to generate the best user experiences in dev kits. Price, OOB Experience, Looks, Tool Chain, Extensibility, etcetc.
As an undergraduate student I played extensively with Altera and Xilinx FGPA teaching boards. Without wishing to be done for slander, these were... awful on almost all the counts above. The only saving grace was the fact that one didn't actually have to get out a soldering iron in order to prove the toolchain didn't work, as they had thoughtfully included a plugboard for your components.
There seems to be little consensus on what people actually want from a dev kit. I guess it depends on how long you want it to last before you have to just build a board and "do it properly".
That said, there are lots of experiments in how to make this stuff work well. The price/looks/functionality of the STM32 primers seems to be the best combination. Mine arrives tomorrow, so will able to test out the toolchain, OOB experience etc. Extensibility I'm a bit worried about, as I didn't buy the "Pro" version so I'm not sure about how much stuff I'll be able to add to it, we'll see. I've previously used the XMOS XDK (https://www.xmos.com/products), which was great (particularly the toolchain) but expensive. Recent kits are (much) cheaper though I'm yet to try them out.
Anyway, my question is:
$1
Perhaps if we all shout loud enough, someone will listen and integrate this stuff into the next version! ;-)
-Screen. It's too darn small. It has a not good pixel number, and it'd be pretty nice if we had something bigger (even if the pixels'd be bigger). Also the area be nicer at 320x240... -Touchscreen. It sucks. The size doesn't permit much finger manipulation, and the stylus is still pretty inaccurate. -Memory. Nowadays memory is very cheap. It'd be extremely nice to have a 256MB of actual RAM, and it wouldn't raise the price significantly. Now with all that RAM we could actually like... run Linux on this >.< -LED's. We have 4 LEDs. This seems like a good number until you realize that you can only control two of them. The charged LED is nearly useless. Would be much nicer if we had at least 4 free LED to work with, or at least three with the charged LED gone. -CircleOS.
The good:
-Audio -SD card and library in general. -output port. Except for the fact that boards are too expensive. -Cortex M3 is very nice in regard to porting.
About the negative things: - Next generation has a 320X240 screen (but it will be more expensive...). - Touch screen: accuracy has been inproved on CircleOS 3.8. But, for sure, it will be easier too use with the larger screen.. - Memory extension: that is for later.. - LEDs: more will be present on the next Primer, but this is not a priority. LCD makes LEDs much less usefull... - CircleOS: it is not a preemptive RTOS. Just a low layer that encapsulates the access to the hardware resources. It could be extend... but we need help from the community.
The next version will be more flexible and it could receive a LAN adapter, but it is not planned for the first sub-versions. We think that a wired LAN does not make sense for such a (mobile) product. Wifi would make more sense...
About the negative things: - Next generation has a 320X240 screen (but it will be more expensive...). - Touch screen: accuracy has been inproved on CircleOS 3.8. But, for sure, it will be easier too use with the larger screen.. - Memory extension: that is for later.. - LEDs: more will be present on the next Primer, but this is not a priority. LCD makes LEDs much less usefull... - CircleOS: it is not a preemptive RTOS. Just a low layer that encapsulates the access to the hardware resources. It could be extend... but we need help from the community.
The most important thing for primer3 to have is memory and screen. With these two it would be incredibly more useful.
As for LCDs, those are amazing for debugging. For example, we can quickly put them to blink in an already done applications without fooling with the output to the screen.
Additionally, they help to troubleshoot when the screen is not available. For example, I was playing around with FreeRTOS and at one point I broke the boot. So I made LEDs light on: we have two LEDs, so I quickly put in four points in the program (01, 10, 11, bb).
What help do you need from the community regarding preemptivity? It is really not that hard to implement.
I'd be also willing to see what can be done regarding primer3 and linux
Last edited by ShadowPhoenix (2009-08-04 20:32:09)
For the next release, SWV will be available for debugging (it can be used as a dedicated fast printf). Anyway, ok, we will add more LEDs when possible (that does not cost so much...). At the moment, no immediate solution for extra memory... But we have to think about this.
For the next release, SWV will be available for debugging (it can be used as a dedicated fast printf). Anyway, ok, we will add more LEDs when possible (that does not cost so much...). At the moment, no immediate solution for extra memory... But we have to think about this.
As for memory, I just looked at the micron page and they have 64mb for $5 even if you buy a single chip(x32), and 256 for $8 (x16).
However, I realize the amount of effort it takes to actually put them on board and have them fully working. The wiring is easy (I believe the connections are very standardized), debugging is hard...
About the negative things: - Next generation has a 320X240 screen (but it will be more expensive...). - Touch screen: accuracy has been inproved on CircleOS 3.8. But, for sure, it will be easier too use with the larger screen.. - CircleOS: it is not a preemptive RTOS. Just a low layer that encapsulates the access to the hardware resources. It could be extend... but we need help from the community.
Great news! When will CircleOS 3.8 be available? Do you think about adding a full audio codec with line IO to Primer 3? I really hope so...
BTW: I think Circle OS is o.k. It is very close to the hardware, which is sometimes exactly what I want. There are a number of RTOS around there, and for bigger projects I use some kind of Linux. Those boards usually have more RAM. If you need more RAM and/or power, look at FriendlyARM or BeagleBoard.
CircleOS 3.8 is ready and it will be published very soon. We postponed the release because it uses the new ST-library, and a new RKit-ARM needs to be released too (at least for those who compile CircleOS). I agree with you. There is a need for different kind of boards/RTOS. The Primer concept should be adapted to different configurations. Even the small Primer1 makes sense, but there is no need (nor any possibility) for Linux on Primer1.
