Using another product, I’ve been working on a device that monitors and posts to the cloud data from renewable energy systems. I’ve recognized that in building this product I’m reinventing things that are already created here for the Spark Core. I did at one time consider the Spark Core, but there is one thing keeping me from doing this, and that is the device must use a wired network connection, not Wi-fi.
Any talk yet about a wired network Core? Same cool everything, just connections over a cable. Perhaps this is possible to hack together with a Wiznet library and current Core, but the deep integration may not be there without a lot of firmware hacking.
I might take a look into this, but wondering if anyone else has an interest or taken a look.
not sure i'd call the rpi finicky. sure its a microprocessor not a microcontroller, but at least you can do useful stuff like SSL, SSH, webserver etc. the arduino ethernet boards are even more limited that the spark in what they can do.
i still dream of a board with onboard wifi, arm chip, gigs of ram and onboard storage, runs linux for months from 2xAA batteries, is a inch square and can run arduino sketches. don't even mention that yun crap.
No matter which board or hardware you suggest, the point still remains that capabilities of the Spark Core and the Spark Cloud need to be reinvented.
I’m using devices from Tibbo, I love them. They are easy. They work. I passed on BBB and RPi as they are not industrial enough for my deployment. Both even if any of these boards or the dozens of other ones, I’m developing cloud software and board firmware for remotely flashing programs, getting device status, security of data transmission, etc.
So, just judging the interest of a wired network core, as not all applications need wifi, and if this is something the community wanted to start hacking together.
I appreciate this is off topic but, really, nothing could be simpler that maintaining the Pi’s o/s. Firstly, there is nothing to maintain, it just runs and runs and runs, like Debian always seems to. Updating it is trivial (if you want to) using the standard Debian apt-get tool, as is installing all kinds of software from 1000’s of packages. The Spark is nice, very nice, but is more difficult to program and to debug, and doesn’t do pre-emptive multtasking, nor does it have 500MB RAM and a couple of GB SD “disk”. All the Pi is missing is A2D, and the small footprint, and it uses (only!) twice as much power as the Spark. I like the Spark, but the Pi is brilliant. One of mine is my mirrored USB 2TB filestore, music server, podcast aggregator, my local name- and DHCP- server. And it has been up 22 days! Oh, and it talks to two Sparks using UDP, getting the temperature from one and telling the other to turn the heating on and off.
I agree with you. also looking at adding BLE to the core during my personal time since everything is open sourced and I believe that the enterprise would love to have an ethernet port for future board
you can get a2d in a single chip, so its really not a problem. i was pretty disappointed with the power requirements of the spark, i can't get my spark to use less than 35mA even in deep sleep mode, with arduino's you're talking under a micro-amp.
A Spark Core with Ethernet, and maybe PoE would be great for some projects.
One of the issues with the Raspberry Pi is that the SD can get corrupted with ease. I spent a considerable amount of time on a product and then found it was not stable if you turned it on and off too many times. A device like an Ethernet Spark would fix that.
Agreed, although hacking RPi to mount the SD read-only with the writable bits in tmpfs is quite doable. The web doesn’t seem knee deep in good examples, but I have a skunk-works project that relies on that.
[Off topic] What I do is partition the Pi’s SD into / and /home - for the same reason. It is months since I have experienced SD card corruption. Or maybe I do but it is fixed but never brought to my attention by fsck sorting it all out for me at boot.
ethernet seems like overkill for a spark though, without seriously beefing up the chip and memory you might as well just use a long usb cable for getting your data serially.
Trying to pull this back to the topic of a wired network Spark Core…
There are many applications where a wired cable is preferred over a WIFI or USB connection. For instance, I’ve got some industrial situations where a carrier board can be made to connect to custom sensors and RS-232 devices. This carrier board is likely to be mounted on DIN rail in a metal or stainless steel cabinet. I’d need a hardwired connection for this and other networked devices as WIFI just isn’t available or not going to be penetrating of the enclosure or environment.
I’ve not looked into details just yet, but perhaps a hardwired Core could use one of the Wiznet IC instead of the TI WIFI module. I’ll have to study the current firmware to see the division of responsibilities now between the WIFI module and the STM microcontroller and review where the network stack is split. The Core would have all but the magnetics and jacks, which is added on the application specific carrier board. POE would certainly be possible too.
The advantage of the Spark Core is the cloud services, remotely programming and flashing firmware and such. In my current deployment of non-Spark hardware I find myself building cloud software that has functionality existing in Spark already.
Any other gurus want to band together to see if we can frankenstein a hardwired spark core?
I’m not sure if this helps but the “core-communication” and the “core-firmware” are separated right from the started for easy porting in future boards and usage by someone else on other platform
Also, i’m not a totally experienced electronics person which i may appear to be, but i’m always happy to provide my help in any way
i've got to have another go at it, as my spark was behaving so oddly when i tried deep sleep that i was wary of trying it for long, had to factory reset it and reflash firmware twice before it would connect to the cloud or serial afterwards, red and multicolour leds flashing, seemed to also fry a bmp180 i had connected. but it had no peripherals when i was measuring the current and it was definitely 35mA which i thought was just regular sleep not deep sleep.