Maybe you should sell the idea to SmartThings, then? I would totally buy half a dozen on these boards right now!
Thanks for the vote of confidence
I am guessing that SmartThings had to deal with the same problem in the past. They are using a CEL MeshConnect module, as far as I know. Like the ZigBit module mentioned above, those are a bit more expensive. They do, however, come with FCC certification, a MAC address, and a pre-tuned antenna, saving costs at that end.
Planning on giving the ZigBit another try, since it might just end up being much cheaper for a small production run like this.
@florianz are you still pursuing this? I’m working on a couple of projects, an IR trip wire for my driveway among other things, and I could use a board like yours. I’ve design and made my own Arduino-compatible boards in the past (e.g.: https://github.com/sudarkoff/thermostat/blob/master/eagle/board.png), but they were all through-hole and used a non-HA ZigBee radio. In other words, I like your design much better! Would you be willing to share the schematics/PCB layout? I promise to use it for personal projects only and not share with anyone else.
Yeah, I am still working on this @sudarkoff. I am going through a major redesign to incorporate the ZigBit module, because of the FCC and MAC address issues. It’s taking me a while.
I am happy to share what I have. I will send some stuff your way, as soon as I get a chance.
Thanks! And I’d be happy to pitch in on the software end. I’ve been writing in C/C++ for over 20 years, it’s very much my forte.
Driveway IR Trip Wire, now there is something I’ve wanted for some time.
I’m experimenting with chaining events for more intelligent state detection. E.g.: interior garage door opens, then motion in the garage detected, then exterior garage door opens and closes, then the driveway trip wire is tripped, then George and his car or bike leave the location and the time is between 7AM and 9AM - George is driving to work. Same in reverse between 5PM and 7PM - George has arrived from work. No application at this point, just trying to see whether it’s even practical given the inherent latencies in the system.
florianz, what’s the difference in your board and something like this?
I’m brand new to all this, so the simpler the explanation, the better.
The board you linked to needs to be populated with the actual ZigBee radio module (CC2530 from Texas Instruments), and it doesn’t come with a software stack. Texas Instruments may provide software for you to jump start your development to make this board talk to a ZigBee Home Automation network, but you may need to purchase additional hardware to do the programming.
Once you put on the radio and develop the firmware, you’ll end up with a board that is very similar to what I am working on.
Thanks, Florianz. Trying to learn this stuff and the only way to do it is to ask what may be a bunch of dumb questions. So, that board I linked above, plus one of these, and some software, provide the components necessary to develop the backbone of smart devices?
Once you’ve developed and perfected a device, would all these parts be required for each such device you wanted to build? Or might there be cheaper groupings of components required, depending on the device you’re developing?
If you want to build a specific device - say a temperature sensor - you can start development with the components you listed. Once it becomes times to mass produce, you can cull all the unnecessary stuff (LEDs, headers, power plug, 5V DC converter, etc.) and create a very specific circuit board, for that very specific temperature sensor, using only the necessary components. Does that make sense?
Yes, makes good sense.
So while I’m learning, if you have a chance, please take a look at this device.
I have a high-end Parasound preamp that offers RS-232 bidirectional control. Would something like this enable me to develop a SmartApp capable of sending/receiving RS-232 commands to/from the preamp?
This really is just an edification exercise. Now that I have Indigo working like a champ, and since it interfaces nicely with Global Cache devices, I’ll probably just grab a Global Cache WiFi-to-RS232 bridge and handle it that way.
Maybe… The product page you linked to doesn’t yield any information about interfacing with the device on the Zigbee side of things. There is no mention of the profile it uses or where to get more detailed information about any how to control it (attributes, commands, etc.) You may be able to interface with it with the SmartThings hub as the coordinator, but I doubt things will just work out-of-the-box.
Even though I have not given any updates in a while, I am still working on this project.
I am really bummed about the FCC certification costs, and the Atmel ZigBit module turns out to be simply too expensive to build this development board within the constraints that I have set myself. I am now evaluating other Zigbee transceiver module options, outside of what Atmel has to offer.
Hi florianz Certainly understandable … Do you happen to be planning on posting the firmware you had developed for the RFR2 any place, or is that project now considered dead, and time to move on?
Just out of curiosity what costs were you looking to obtain for a 15.4 certified module, and at what annual volumes ?
Looking forward to hear about what you finally select as a viable solution.
Absolutely. I have it in a private BitBucket repository, which I am happy to give you access to: Just PM me your BitBucket handle. It’s not so much a “firmware” as it is the BitCloud framework with some bugs fixed and one of the examples modified to work with SmartThings. Nonetheless, it should definitely help you hit the ground running if you want to play around with the Atmega.
Well, frankly I have no manufacturing experience. Everything I know about it is from reading other project stories. Also, I think there isn’t really a high demand for a product like this. The target group is sort of niche, because beginner makers would probably want to stick to the ThingShield, and advanced makers can go straight to designing their own PCBs using whatever components they want. So this is for the “in-between skill level” makers, or for prototyping of a production device.
Having said that, I may want to build 100 - 200 as a first batch - so we are not really talking numbers where quantities make a significant dink in costs.
I am currently exploring the option of using a non-programmable Digi Xbee surface mount module as a network processor, and an external MCU (like an Atmega328P or low-power STM32) for application code. This has the added benefit of making Arduino IDE integration much easier, because the network stack of the firmware doesn’t have to be compiled and flashed together with the application code.
What are your thoughts?
Have you looked at available ZigBee modules, like this one for example:
Making your own hardware can be fun and educational, no doubt about it, but considering time and cost, I don’t think you can beat $26 shipped.
My first prototypes are around that cost (not including assembly, granted) and have a richer feature set, as well as a smaller form factor that doesn’t require the annoying rubber ducky antenna.
The module you linked to doesn’t come with software of any kind, so chances are you still need an expensive compiler toolchain (Texas Instruments requires IAR) and potentially a programmer / debugger. There is an open-source compiler for 8051 based MCUs (like the CC2430) called SDCC, but I am not sure if it’s mature enough to compile Z-Stack, which is TI’s own ZigBee library. Up until recently it wasn’t able to produce good enough binaries, but SDCC is under active development, so chances are the situation is better now.
Also, the module you linked to doesn’t come with FCC certification or MAC address allocation, so I am left with the same problems as with my own prototype.
The head-start I would get from using a module such as the one you linked to is marginal.
The only off-the-shelve module that comes close to my requirement is Pinoccio. It just needs different firmware, but it also comes with a whole bunch of unnecessary (for this application) features, which significantly drive up the price.
Yeah, the more I think about it, the more I wonder why bother with Zigbee if you can buy a WiFi module for $15 these days and connect directly to ST cloud.
Apparently, there’s OpenWRT port for this puppy.
The problem with wifi is that it’s not a great choice for low-powered sensor networks. There aren’t any wifi chips that come anywhere close to the low power consumption of a Zigbee radio. Even the new CC3100/CC3200 is still ways off. Wifi is also not a mesh network.