First, up thank you to the community for your help last week, I was able to get my project up and running in large part thanks to your help :).
Unfortunately, I am now facing a new issue. So what I tried to build was an electronic scoreboard linked to 3 wireless clickers I built. The clickers and scoreboard were powered by Xenons and there was an additional Argon acting as a node for code upload.
Initially, the system worked fine and we disabled the particle cloud to operate on mesh only, so as to have better stability. We did connection and distance tests and the clicker functioned well beyond the maximum needed distance with no additional nodes in between the clicker and scoreboard to increase the range.
However, on the day itself, the system was plagued by connection issues. By the use of programming the onboard RGB Led, we know that the clickers were publishing mesh messages but the scoreboard was not receiving it. The only difference between our test setup and the actual event was the presence of a large audience, about 150-200 people.
We were working well within the range of our maximum distance, the clickers and scoreboard indicated they had not lost mesh connection indicated, but messages were simply not reaching the scoreboard.
Could the presence of audience personal smart devices have affected the mesh communications? I had thought the mesh operated on different technology and frequencies.
And is there any way to protect against this in the future? We hope to use particle mesh as the backbone for a fully automated scoring system if we can fix this issue.
Thanks again all for reading this and for any light you could shed on why this issue is occurring.
2.4GHz interference will always be something you need to take into consideration - WiFi, Bluetooth and Mesh operate within that band concurrently.
Additional nodes/repeaters and some acknowledgement mechanism may need to be implemented.
Mesh operates ontop of UDP and hence delivery of packets is not guaranteed by the protocol.
There is a link somewhere on the community forum to a video of Steve Jobs trying to demo some WiFi device to a large audience of press where he has to ask them to switch off their WiFi hotspots, etc. You get the idea.
At exhibitions when we demo products with WiFi we gave up a long time ago and now hardwire (via ethernet) each one. If you do a scan of WiFi networks typically there will be +50 visible. I tried a simple button press over Mesh at a show in January - I probably managed 3-5m effective range even with external antennas.
Some people use alternate channels with WiFi (rather than the defaults) but this selection isn’t possible with the Mesh radio, unless someone else knows better?
Thanks for clarifying guys, that makes things a lot clearer. If possible may I seek your advice then on how I could proceed forward.
My end goal is to build an automated scoring system for a sword fighting martial art. When we duel we wear gloves and helmets.
The goal is to have sensors in the gloves and helmets that can automatically detect the hit and transmit it to a scoreboard for the public to know what is going on.
The communication method has to be wireless as we move around quite a lot in a circular ring, about 9-10 meters in Diameter. As the intention of this is for the public’s benefit, the system must be able to operate reliably in high crowd circumstances like the one I experienced.
I will try what you both mentioned by including acknowledgements and more mesh nodes, but could I seek your advice on if you feel particle mesh is a feasible communication choice for my application?
Or is it just not possible from a technical standpoint to have mesh reliability in those circumstances?
One thing I’d reconsider is the use of mere (dumb) “clickers” and rather have each of them hold the current score and retransmit its current count on each event plus regularly in intervals till it gets an acknowledgment of the central node which matches its own score.
Each of the “clickers” should also subscribe to the events published by the others and monitor their scores and retransmit theirs too.
The central score board should keep track of any discrepancies reported by the “clickers” and use the maximum score of each individual contestant regardless of which device reported the score (assuming the score can only ever increase but never decrease).
Depending on the error rate the central station could inform the “clickers” to increase/decrease the retransmission rate to find a good balance between data integrity and radio silence/power preservation.
With this scheme I’d think Particle mesh should be capable of doing what you want with minimal extra circuitry.
Of course using a less crowded radio band would rid you of the 2.4GHz headache but would add some other complexity to the setup (HW & SW).
Use BLE to communicate between 2 Paired Gen3 Particle devices - 1 on the sword fighter with wired sensors to the device and 1 ring side and this could then communicate using a store and forward buffer to the score board or be hard wired!
Consider using another device and mesh technology on the sword fighter, ring-side and on the scoreboard - Bluetooth Mesh is low latency and more robust in noisy environments. It is new and there are not many devices around to develop off. Particle have said they are not doing Bluetooth Mesh.
Particle Mesh is based upon Open Thread 1.0 (I may be wrong about this version) - the Mesh technology is aimed at Home Automation and came out of Nest Labs. It is not ready for a commercial environment with lots of noise and is fairly low throughput. Just this summer Open Thread 1.2 specification came out - this is aimed enabling Open Thread use in office and commercial/corporate environments. Implementation is some way off. My experience of trying to setup and run a Particle Mesh (even with the latest and best Device OS 1.4.2) in an office environment has so far fallen short of being usable - range, stability of nodes and reliability of the publishes - which a year after Gen3 was actually delivered is a tad disappointing.
I want to say thank you for the post and for the discussion.
It's very informative for Gen3 Mesh users.
I have a thought that may or may not help.
Since we believe the Gen3 Mesh is seeing the other 2.4GHz devices as noise, maybe add a Xenon at the ring (to forward the sensor Xenon's data) and use directional antenna between the Argon and the ring Xenon.
I assume the "ring" Xenon would be very close to the athletes and Xenon sensors.
I also assume the weak link is getting this data back to the scoreboard with the external 2.4GHz noise.
Think of this (1) stationary link as "blinders" for the Argon, in addition to increased RSSI from the (2) directional antennas.
The "ring" Xenon could also become the Brains of the operation (since it's the closest to the action), per @ScruffR's suggestions, with the Scoreboard acting as a peripheral with the directional antenna link.