Mesh range / external antenna?

Does anyone know what kind of range should be expected in reality for mesh devices in a residential house? I know the FAQ says “10s of meters” and “similar to bluetooth” but thats not very specific.

I’m asking because I’ve got 10 xenons scattered about my house, and every one of them except the one acting as a gateway on an ethernet featherwing warns me “The device is disconnecting from the Particle Cloud too frequently.” - some have disconnect counts in the hundreds.

I was thinking about putting an external antenna on the gateway and also one device per floor. Do you think this would help? I found this U.FL 2.4ghz antenna, any reason I shouldn’t order a few to try?

1 Like

@sean13, the range is very dependent on obstacles such as walls and floors. I first arrange Xenon “risers” as my Xenon/Ethernet gateway is in the basement. These risers are in good line-of-sight with the gateway and reliably maintain their connections. I then arrange a Xenon that connects well on the ground floor. This Xenon will be used the the ground floor devices. Finally, I setup a Xenon on the 2nd floor, again making sure I have a reliable connection to the ground floor riser Xenon. You get the idea. I don’t yet have enough Xenons to create a 10-node mesh yet.

I have not seen any antenna recommendation from Particle as of yet. Your choice however seems to fit the bill though so it is worth trying.

One thing I have done so far is to NOT do Particle Cloud operations on the Xenon mesh and instead keep them to Mesh pub/sub. All Cloud operations are done on the gateway instead with results sent to the mesh via Mesh.publish(). For now, I have found the mesh more stable and reliable with this approach.

1 Like

@peekay123 When I receive my stuff (international seems to be taking an age) I would like to setup something similar but connecting to a larger number of Xenons. Just to confirm, you have a Particle Ethernet FeatherWing Adaptor with a Xenon acting as the Cloud/Mesh Gateway. Is connecting to the Particle Cloud via ethernet straightforward and reliable? Do you have any indication about memory usage for each node added yet. The theory seems to be 10-12 nodes is about the limit for a mesh or is that just repeater to end node ratio?

From my initial testing using an Argon and 5 Xenons the Xenons rarely have a connection to the Particle cloud even though the Argon is connected and breathing cyan just fine.

The only thing the Xenons are good for at this point is for the MESH Pub & Sub to each other and to the Argon which can push data to the Particle cloud.

I let this setup run all night and now and 1 Xenon has fallen out of the loop while Xenons 1,3,4,5 are still reaching the Argon and which is publishing every mesh message it receives.

There is plenty more work needed to be done to make this all work as intended.

1 Like

Yes

Yes, very reliably.

No, I haven't been looking at that yet. With the current firmware, ALL nodes are repeaters. Typically, you don't have a single node connecting to 12 other nodes in a mesh. The idea is to have a mesh geometry where you have, typically, any given node see up to four other nodes. I don't believe any given node "knows" the whole mesh. Instead, it knows only its connected neighbours. What is not clear to me is how the routing is done back to the gateway(s).

@RWB, Stand-alone, the Argon works fine. However, I believe the Argon is not stable as a gateway yet. There are several posts to this effect and issues have been raised with the Particle Team. If you have an ethernet shield I suggest trying it out using one of your Xenons as the gateway processor.

So Far Argon is staying connected just fine and publishing like a champ.

The Boron is super stable also as far as staying connected and with the Boron provided a more consistent Particle cloud connection to the Xenons vs the Argon.

I did not order the Ethernet wing so I can’t speak about its performance.

I’m needing to manually reset the Xenons periodically to bring them back to a connected state.

Range seems to be not that bad for but more testing needs to be done.

1 Like

@peekay123, our setups sound similar. My xenon w/ ethernet wing is in the basement, and then other xenons are scattered about the house. However, for the rest of my xenons, I didn’t plan out risers & routing like you did, having the ‘10s of meters’ thought in my mind. All of mine are certainly within 10 meters of at least one other one. This evening I will try laying things out more intelligently, and bringing them online slower. I think I will also source a few of those antennas.

@armor connecting my xenon with particle ethernet featherwing adapter was very easy & it is rock solid. I had a problem at first with ipv6 packets messing with the stability of mesh-connected nodes, but I’ve resolved that. See here for that story.

@RWB & @peekay123 thanks for confirming that I’m not alone. I’ve done very little with the devices so far, except try to get them all talking to the cloud.

1 Like

@sean13, the recommended antenna, it seems, is the same one used by the Argon and sold in the Particle Store I believe.

As for range, it is much like WiFi with walls, pipes, floors and whatnot affecting the signal. So it is very much YMMV. The “riser” approach allows me to pick the best signal path to go between floors. If you have a continuous stairway from basement to top floor, for example, that makes a perfect place to rise. Locating a Xenon on each floor near that stairway then gives you connectivity on each floor.

Don’t be surprised if you see someone come up with a Xenon wall wart that can act like a floor-to-floor mesh “connector”!

@RWB, I found the same thing with the Argon. It is stable but the nodes not so much. After an OTA to the Argon or a node everything goes to heck and the Argon seems to drop the mesh and the nodes lose connectivity. I have to bring the Argon back up and then the nodes one at a time to recover. However, this is not always the case so it is not consistent.

With the Boron and ethernet everything is quite stable as you have found also.

During then Particle cloud outage today in had an Argon and P1 up and running.

After then the issue was resolved on Particles end the P1 reconnected and kept publishing just fine.

The Argon never reconnected and remained in a flashing cyan state until I manually Hit The reset button. Then it did start publishing as intended from all five mesh publish events.

Anybody else have an Argon running and experience the same thing?

Looking forward to the day the Argon is as stable as the Photons.

1 Like

All my mesh devices seem to be up (looking at Console remotely) including my Argon though it is running Tinker and has no connected nodes.

1 Like

Yesterday, outdoors, without antennas, using Marco Polo Mesh Test I was able to get 90% reliability at 150 feet - clear line of sight.

At less than 150 feet it was “seemingly” perfect (I didn’t test to 5 nines or anything). It would be nice to try again with antennas but I don’t have any right now.

4 Likes

I got my Mesh devices just a few days ago, and finally threw together a quick test between an Argon and a Xenon last night. The Argon has an antenna for WiFi, but neither device had an antenna for Bluetooth. My test sketch uses Mesh.publish() from the Xenon (data from the light sensor, every 5 seconds), with a corresponding Mesh.subscribe() on the Argon, which then turns around and does a Particle.publish() to the Cloud. I could then monitor the events on my phone with the Particle app.

I put the Argon in my kitchen, which is about 5m from the front outer all of our house. Then I took the Xenon outside to see how far I could get before losing signal. It was probably about another 5m from my front door before I started losing events. I was a little disappointed until I remembered that this was without antennae for Mesh. With that in mind, 10m (maybe a tad more than that?) though walls probably isn’t too terrible.

I might try scavenging the antenna from my other Argon kit and trying again, to see how much difference that makes, and maybe moving the antenna on the first one from the WiFI u.fl to Mesh. I’ll also try moving the Argon to other locations, because my eventual goal is to get mail arrival notifications from our mailbox (roughly 30m from the front door). That might require a Xenon acting as a relay located in the corner of my garage closest to the mailbox.

1 Like

I wouldn't do that. You can use any WiFi antenna for the mesh (the suggested one is the Argon WiFi antenna). Mesh does NOT use Bluetooth. It uses 6LowPan on 802.15.4 with Thread. If you have an external antenna for a Photon (pcb, duck), it will most likely work with mesh.

1 Like

Does the mesh antenna go on the antenna out labeled “BT”? Why BT?

1 Like

Am I correct that any external mesh antenna has to be selected in software, and that the firmware doesn’t yet support doing this?

These are the antennas I ended up purchasing. Hard to beat that price (that’s for 2 of them)

1 Like

@rickkas7 can confirm external antennas working or not. Pretty sure he said that does not work now.

Thats what I thought. I’m fine with that, just wanted to know if I should leave 'em on the shelf or start playing with them, and wanted to make sure others were aware as well.

Well, when I said "antenna for Bluetooth", I meant "for Mesh". I realize it's not technically Bluetooth, but aren't the Mesh protocols related, and using the BT chipset on the Nordic? In any case, I was referring to the 2.4GHz antenna that came with the Argon.

Why should I not move that antenna to the "BT" u.fl connector? Shouldn't that give me better Mesh connectivity?

Oh, and while we're talking about those suckers, does anybody have any tips/tricks for getting that tiny u.fl connector to, well, connect? I spent 15 minutes trying to get that Argon WiFi antenna connected last night, and I spent another 10 minutes earlier trying to put one on my Xenon, before something came up and I had to put it down. :face_with_symbols_over_mouth:

From the docs
"Two U.FL connectors for external antennas (one for Thread/Bluetooth, another for Wi-Fi)"

Probably because "BT" is shorter than anything you could come up with for "Thread/Mesh", space is limited, and BT is more recognizable.

Just my guesses though. Also, if they named it mesh, the question would become "Why not BT?" :wink: