I know that others have asked about the possibility of using Electron+Photon to make a cellular gateway before. Generally the response from the community has been either “it’s coming” or “why not just buy a hotspot”.
I want to create a hotspot that will have a low duty cycle to keep power consumption to a minimum. Ideally the hotspot would sync Time to the particle Cloud each time it wakes up. It would stay alive to a minute or two and then go back to sleep. The Cluster of Photon sensors would do the same (I actually already have them doing this). For example, every 15 minutes the entire cluster and the hotspot would wake up (everything is solar powered in this scenario), get connected, read data, and then go back to sleep.
I’ve thought about doing this with off-the-shelf hotspots from the carriers, but none that I have yet found can turn themselves on when power is applied (they require a physical switch to be depressed, and the thought of putting a servo into the mix feels way too Rube Goldburgish).
It seems that the ‘perfect’ solution would be an Electron+Photon, but I am open to alternatives!
Imagine if you will…A solar powered IoT sensor cluster monitoring conditions far from the grid all built using Electron and Photon. One application could be to deploy clusters around remote oil and gas wells. Another could be to detect emissions from water treatment facilities. Help me! This has to be possible!
@Darmitage, the Particle platform doesn’t support publishing from one device to another, only to the cloud. It also doesn’t support a cloud publishing “gateway”.
A Photon solution would still require a WiFi AP to connect to. A “master” Photon could collect data from the other Photons via UDP or TCP (unencrypted) and pass it to an Electron. This seems very messy. All this to say that it may be more feasible to use Electrons all around.
I have messed around with the Bluz gateway quite a bit with an Electron as the WAN access. A Bluz Gateway would be fine BUT the range on the BLE modules is very limited. I’ve also investigated the possibility of using a Raspberry Pi. Any observations on the Pi approach? Is it possible to set up a Photon as the ‘master’ and use it as an AP? …I didn’t think it was…
My initial reaction was no…I like to more elegant WAN/LAN sensor cluster model. BUT after spending more time thinking this through, I think you are right. The added cost per node is small and largely offset by the cost of a hotspot as long as the cluster of sensor nodes is relatively small per site.
I estimate 6 nodes per site is where it starts to not make sense to use Electrons on every sensor node.
AND I think it is reasonable to assume that Particle.IO is hard at work on a variant of Electron that will use LTE-CAT 1, and in the not distant future LTE-CAT M1. As these modules arrive it will be increasing difficult to justify the WAN/LAN approach for a sensor cluster.
I'm not sure that would be more elegant, I would say it would also be more complicated to code the communication.
You can keep the solar panel cost down by just using 10w glass solar panels vs the non-glass solar panels that have a higher dollar per watt cost.
With the Photon's you would need to add a battery and battery management board to the mix which is going to cost you also, especially if you want to be able to report the fuel gauge status like you can do with the Electrons.
I've played with RFM95 2 way radio chips with 1-mile range but they need a microcontroller built in and you still need a battery and solar panels to charge it and coding the communication between the radio nodes and the Electron becomes more complicated than using a simple Particle Publish and Subscribe method.
After playing around with many different options for wireless IoT solutions recently I have really come to appreciate what the Electron can do and how easy it can make getting sensor data to the internet with very high reliability.
Recently I have found Losant.com to be killer IOT Backend service to receive, database, and then show that data via highly customizable dashboards.