You've made some interesting choices here. Lots of people here are questioning using a gateway, but I think this is a great choice for what you are doing, and required to get decent battery life. However, I think using ZigBee will end up pricing you our of most markets in the long run. Especially if you elected to go with the very pricing Xbee modules from Digi. These are developer friendly, but way to expensive. If you are serious about building this as a product, I would investigate alternative RF solutions, there a lot of good ones out there.

Check out the 6lowpan standard, which is specifically designed for this kind of communication. Dust Networks (http://www.linear.com/products/smartmesh_ip) has a good implementation. The other alternative would be to go with some of the proprietary RF standards, I know this sounds prohibitive at first but will end up saving you money on your BoM in the long run. Check out Atmel's LightWeight Mesh (http://www.atmel.com/tools/lightweight_mesh.aspx) which is based on 802.15.4, the same as Zigbee and BT, so if you want you can even run ZigBee on the same hardware. Linx also makes a great proprietary RF module (https://www.linxtechnologies.com/en/home) which is extremely cheap even at low quantities.

If you are using Xbee (which it seems like you are) then you are likely not using a SoC, so designing in a different radio may still be relatively easy.

PM me if you have any questions, I've been working on this IoT stuff for a while now. Congrats on the crowd funding campaign and best of luck.

As an outsider to the space, your comment makes it sound like ZigBee is never a good option. Is this so? Is there a 'target market' for ZigBee?

Industrial.

You'll probably find that most price-sensitive consumer devices that don't need to interface to a standard protocol like WiFi will use very basic RF communication with a homebrew protocol. It makes the engineering cost much higher than using more complex devices, but spread over hundreds of thousands or millions of units, that cost vanishes.

OTOH, industrial markets are a lot less price sensitive. Also, quantities sold will be much less, so the device's unit price has to be higher to make building it worthwhile.

Exactly, it is till good if ease of use and time to market is the major goal and price isnt a big question, which is certainly true in many markets.

However, I think we will see it fall out of favor in consumer markets unless thye drastically reduce the license fees.

Congrats, this is an interesting project and props for doing it all on your own! I thought our kind has gone all but extinct these days :)

That said, here's a problem from just watching the video - all the focus on the design of the device assumes that it should be on a display. That's the domain of the home automation market for passionate amateurs - "Look, look, I automated my door!" A proper installation is always discrete and does not clash with existing interior design of the space. It would be something that blends in easily and, ideally, paintable over. Definitely not a computer-mouse like puck that just screams "look at me!". Your whole point of this being "technology that is not worth hiding" is way off for practical applications. I realize that you probably trying to pull Nest here, but this is an utilitarian non-interactive device, it has no business being visible in my space. I would strongly suggest adding an "OEM" model that is just a flat white box.

That sounds cool! I do recommend you highlight a few practical use cases, i.e. what benefits does (or will) it provide? Try to lead with why rather than with what. It took me a long time to realise they're all the way at the bottom.

I also recommend not repeating the words "notifying object" so often - right now my eyes might go from: "Notifon, the notifying object." (1st headline) to "Transform any object into a notifying object!" (2nd headline) to "Notifon / Transform any object into a notifying object" (video) to "Preorder now" to "Notifon is a wireless Internet-connected device that sends you instant notifications based on real life events." and I've learned little more than the 1st headline told me.

Thanks for sharing this and good luck with your crowdfunding campaign!

Thanks for those good recommendations, I think you're so right with the "notifying object" words! I think the Indiegogo page is better, I'll merge it with the main website.

This is cool marketing but it is way expensive than I expected. I am not sure pricing is DIY-friendly but you would certainly find people to buy that.

I use a technology called ZigBee. It's a wireless protocol made for Internet of Things devices. This part represent about 60% of the production cost. This chip is expensive but gives years of battery life. Curious about what price did you expected?

I expected this to work through either wifi or a very cheap USB dongle. The price point for a base station with one device should be 49.99 and each additional device should retail for 14.99. That's my honest opinion.

This is really interesting to me too. I've been looking at a bunch of technologies to make "jelly bean" sensor nodes (i.e. cheap enough and simple enough that I don't have to think twice about popping one into a project).

Those price points are hard to reach for retail pricing, but are pretty close to the "in bulk" pricing that I've got things down to (once you factor in labour, NRE for solid firmware, etc). If you want a cheap and unreliable setup, you could probably get the node price down to around $5 or $8, but I don't think I'd want to rely on it.

Thanks for the reply. This kind of price can be reach using Bluetooth Low Energy. Those chips are really cheap but as you might know, the range is too low.

BLE doesn't sound too bad? 100 meters according to WikiPedia: http://en.wikipedia.org/wiki/Bluetooth_low_energy#Technical_...

This can of course be far from the effective distance, but since the data being communicated is so small it should be possible to keep sending the data until you can acknowledge that it has been received.

Also, you could look into creating a mesh network between all Notifons, so you could extend their range.

http://stackoverflow.com/questions/14136218/bluetooth-mesh-n...

You can't build a mesh network with sensors on idle mode.

I agree this is way too expensive.

Looks great! Out of curiosity, why go with the hub/node approach using ZigBee for sending such a small amount of data? There are WiFi modules like TI's CC3000 (http://www.ti.com/product/cc3000) available for very low cost that would make separate hub hardware unnecessary. This is the product used in Spark Core (https://www.spark.io/), for example. I imagine there are a few different reasons why you might spring for a hub-based solution despite the higher cost, but just wondering what your thinking is.

We use a wifi module in a battery operated environment (we kill power to the module when it isn't in use).

It uses about 225mA @ 3.2V.

The connection/handshaking take about 5-7 seconds to talk to a local server, plus the time for whatever data you are sending.

When talking to an internet based server, the connection time can extend out to 20+ seconds.

So you can see wifi isn't the greatest choice for low power/battery operated environments.

(Now on something you don't mind recharging a few times a week, it is perfect.)

Exactly, for Notifon the transmit time is 40ms. That's why it can run on a small and sexy coin battery (CR2032).

I much prefer standard mignon types. They are an ubiquity.

I know, it's all about the battery life (about years). The ZigBee chip require only 40mA in transmit and less than 1uA on idle mode... You can't do the same with a WiFi chip, ask Spark... Also, with Hub/Node there is no "association" need like entering the WPA key. It's all automatic

Interesting. Seems similar—complete with slightly creepy base station—to Sen.se: https://sen.se/store/cookie/#mother

(Though it appears Karotz is a separate thing: http://store.karotz.com/en_US/)

There's also Electric Imp, but it's a bit more of a platform/DIY: https://electricimp.com/

My first thought was that it looked very similar to Ninja Blocks: http://ninjablocks.com/products/ninja-blocks-kit

Eww, the Sen.se is kinda out there. Thanks for the links though, I keep thinking there is a winner here somewhere.

How do you know the garage door is "open"? If you receive an identical notification whenever the sensor is triggered, the door could be open or closed.

With something like a garage door, knowing its state is very important, perhaps more important than knowing when the sensor has been triggered. When someone leaves the garage door open (kids), this is what we need to be notified about. Is the Notifon up to this task?

Also, range might be a real issue when the nodes are placed outside or in the letterbox.

One more thing, and this is just a personal preference in response to the idea that I "won't want to hide the device". I don't understand why people think white plastic is the best choice of color. So many cheap kitchen appliances are white plastic. If you have floorboards and warm living areas with warm tones everywhere, shiny white plastic objects are an eyesore. IMHO.

Of course the Node can trigger two states. So you know when it's open or closed. You can set a different message for each state.

Sorry, but just saying "it can do it" without explaining how, is not convincing!

The magnetic switch in the video - I have used before, and they do not distinguish between "open" and "closed".

When someone opens the garage door, a notification is sent: "garage door open".

When someone CLOSES the garage door, a notification is sent: "garage door open".

Unless you can provide further explanation, the device only notifies of "activity", not specific activity such as open or closed. In which case, you're being a tad misleading.

I think you should try again to play with your magnetic switch. Here is some help: http://en.wikipedia.org/wiki/Reed_switch

Ok, thanks for the useful link (I think). When I played with wireless door sensors and Arduino, detecting door state reliably wasn't happening. Maybe it's the wired vs wireless factor that gives yours the edge. Maybe that's the trade off to know open/closed state with certainty, is to use wired sensors.

Been working on overcoming some similar issues for another IoTs project.

The wireless chipsets for IoT are really limiting IoT because there are so many tradeoffs. Zigbee seems to have most of the technical challenges solved yet have such a high cost. Raw rf/radios require custom (error prone ) software stacks but are super cheap, wifi uses TONS of power and BLE is fairly limited in range even if it in theory can get better range... I think as soon as someone solves the chipset for the IoT, its going to be as simple as slapping I2C sensors/actuators and 3d printing mechanical parts for fast innovation.

Wow, Amazed someone used ZigBee instead of Bluetooth Low Energy. There's just so many more BLE chips out there and Android and iOS phones are even compatible nowadays.

The key is to get notifications when you're away. BLE don't have a good range (about 10 meters). ZigBee is more like Wifi (about 50/100 meters).

That you managed to do everything is impressive, especially given the costs involved in production tooling for plastics. I assume you had to build a large quantity of PCBA's as well, to get your per unit cost down to something you could actually make a profit on, which means that was another large check you had to write for components and assembly.

Can you give a ball park figure for how much you invested, bringing this product from concept to market?

Thank you. In fact, I haven't yet launch the mass production. I'm running a crowdfunding campaign on Indiegogo to have at least 400 preorders. As I have done a lot myself, I didn't had to invest so much. The only part that was done outside is the design of the plastic enclosure. I found a good designer in china, it cost me only 700$ for both Node and Hub with 3D printed prototype. In total, I think I spent about 1500$. I had to buy an oscilloscope, some componants and others tools...

  The only part that was done outside is the design of
  the plastic enclosure. I found a good designer in china

Would you mind sharing how you located this designer/plastic shop? I've had mixed results with mecheng partners outside of my personal network, so I'm always curious about how others do it..

I found him on elance.com

Would you mind sharing the contact details of the designer & 3d printer with me (email in profile) ? Thanks!

This is pretty cool. Can you describe how you went about designing the slick outer packaging? Did you use an external vendor to design & produce? I am doing a DIY project with my kids and wondering how to get the outer packaging for the arduino board inside. Any pointers appreciated...

The best way is to found a freelancer in China. I used elance.com

This is brilliant!

I'm right now looking into building something similar but for industrial/agricultural applications in unattended machine monitoring. I'm planning to prototype using a similiar Digi device: the 900 series with 1 mile range.

I would never have thought of a consumer version. Fantastic idea.

This looks really cool. Just a side note, you might want to update your website to use screenshots from iOS 7. The current screenshots look a little dated because they are from iOS 5 or 6.

The web page gives me pretty much absolutely no idea what this product does, except for the abstract "notify me of things". How does it do it, how do you interface to it, etc?

In fact there are more details on the Indiegogo page. Can you take a look and tell me if I have to add some details? (http://www.indiegogo.com/projects/notifon-the-notifying-obje...)

The key thing is what you wrote there:

"Sensors

Notifon is universal. You can use any sensor that has two states, such as a switch, a push button, a magnetic door switch, a relay... Notifon allows you to predefine a set of actions for each state."

Just put that (or a simplified form of it - a graphical illustration is helpful) somewhere quite visible on the actual product page?

Looks very nice. When I went to the page, I wondered, "notifying about what?"

Your video did a great job at presenting the product and explaining what it's good for.

Good luck with your crowdfunding.

Thank you, I wanted the video short and demonstrative

You've got precious little runway left on that campaign, which launched in early March. What other promotion have you been doing to get people to buy in to Notifon?

That's the point, I think I launched the campaign too soon. I didn't had a good press coverage on the launch day. One week later, I got answers from Mashable and Techcrunch reporters, they seemed very interested by Notifon, until I sent them the campaign link (with only 3% of the financial goal reached).

Good luck with the rest of it. I like the idea and it's definitely clear you have the tech down.

Good first pass. But the page you link to needs copy-editing. There are a few sentences that don't have agreeing verb tenses. Not much, but it is there.

1) All screenshots are of iPhones, is there Android compatibility? 2) PushingBox only allows login with a Google account.

Both of these mean I could never be your customer.

My main phone is an iPhone but of course Notifon is compatible with Android and Windows Phone. For now PushingBox require a Google account, I plan to do a lot of development to make PushingBox even better. So don't worry, you can be a customer ;)

What do you need to do for FCC compliance? Is it all taken care of because you're using ZigBee?

interesting, it looks like you spent a lot of time on it, congrats on launching. Did you also consider using zwave over zigbee? I've been looking at installing a zwave hub and a few light switches, motion sensors and thermostat. Care to compare/contrast?

Thanks, Z-Wave is also a good technology but I hate the "association" part, when you need to be at least 3 feets from the Hub. I don't think Z-Wave will survive in the next 5 years. Look at Philips Hue, they also chose ZigBee.

I really need to understand and learn how it works on such less power. Kudos.

Simple, when you open your door, the Node wakes up during 40ms, transmits at 40mA to the Hub and goes back to sleep consuming less than 1uA. You can calculate the battery life yourself: http://oregonembedded.com/batterycalc.htm

This is really awesome. Great work.