The websites meshmap.net and meshtastic.liamcottle.net occasionally display inaccurate neighbour connections.
One reason for this issue is that meshtastic nodes running firmware versions older than v2.3.2, when connected to MQTT, mistakenly identify remote MQTT nodes as neighbouring devices. There might be other causes.
Yes - that would seem to be the case. Itās sitting at 4.16 volts.
Good to know its working. Where did you get your 18650 battery from?
Bought a Ā£4 solar controller with variable output voltage. Will connect it to a 30w solar panel for offgrid powering the Kincorth node.
Still thinking what battery to use, cheaper the better. Maybe reuse from something, though about batteries from Ecigs
Got a extra 30w solar panel that might sell or swap for something usefull
I got a pretty decent mains charger from amazon and that came with two 18650 batteriesā¦AND a āfreeā LED torch! lol
I visited my ā9eb8ā location again today to add position information to the node. I donāt live at that location but I do visit regularly.
I noticed that there were around 15 names in the node list. They ranged from Cults (4 miles southwest) to Newmacher Golf Course (6.6 miles north). Most are in Aberdeen City, and a couple appear to be companies.
I havenāt had the chance to test the range yet, so I will leave the node running there for the time being. Eventually, I plan to try it at another window sill pointing north. The line of sight will be towards Brimmond Hill with Bennachie on the horizon, and I expect it will still cover much the city centre.
Its wild how quickly the amount of users in Abz expanded. Few weeks ago when I first tried there was 1 other user. Now 6 - 7 regulary and upto 9 at times. Didnāt see the cults user from city center yet. Hope more will join
Will be interesting to see what happens when you try other window.
Is your setup backed up by battery/solar?
The Abz Mobile Repeater is. And gathering bits to make the Kincorth Base solar powered. Need to find a battery either 12v or a 18650 on a budget
When I drove about Aberdeen a month ago, I only saw 1 node as well. I expect most of the nodes existed before but everyone is switching on, seeing nothing, and switching off. I am seeing nodes with names like ābedroomā so maybe they are just being used for personal projects with raspberry piās or arduinos.
With our nodes now running 24/7, then the mesh routing tables are increasing and being distributed further. We have contributed to the visibility. The global maps show 1 in Peterculter, and I saw another temporarily pop up in Inverurie. There was a cluster of 5 nodes between Peterhead and Fraserburgh which have also disappeared off of the map. I have seen evidence of over 20 nodes in Aberdeenshire so there is scope for improvement.
My node doesnāt have a battery or solar panel. I am not yet comfortable about adding a Lithium battery backup as there is a fire risk. These things can supply a lot of current.
Scotvapes on George Street sell 18650ās but they are missing the button at the +ve end (flat top). I think you can order online from uk.rs-online.com and collect from their trade counter in Altens. I have never tried this but they have lots of LoRa bits and pieces. Argos used to sell them too.
During the aurora event starting around the 20th May, my Aberdeen node list increased dramatically with long range contacts into Norway and the North of England at the same time. Another aurora event is predicted later this week from 5th June onwards.
For me, my longest 0 hop connection was to node_tjeldsto in Norway (344 miles - Fair), followed by GarfieldUK in England (255 Miles - Good). A couple of Norwegians did message me directly but unfortunately I wasnāt around to reply.
There might have been longer ranges but most nodes I see do not have their location set.
The trick seems to be aurora, and a low angle take off into the horizon which is ideally over sea.
Sorry for the late reply. It was a node set up on Kincorth. Tried the fixed location and everytime saved/sending the entered l9cation it jumped to the north sea! Node been moved to central location now.
Currently short range as indoors. Looking for a waterproof case with space for battery and solar charger
I picked you up sometime towards the end of last week, Thursday or Friday I think. Then you disappeared again and I havenāt picked you up since. Danestone area here.
I was in Aberdeen on Friday, driving around with a Taixali node in the back of my car. Although I donāt live in Aberdeen, I set up a Foresterhill node, which I thought might be beneficial to anyone using Meshtastic in the city. The location is elevated and reaches over 40 km north of the city. Many years ago, I used to operate a 10W VHF packet radio from that site and could reliably connect to the Netherlands, Belgium, and Germany during summer Sporadic-E events. I havenāt seen anything from these countries so far on Meshtastic, but during Aurora events, my Foresterhill node comes to life with connections from Norway and the North East of England.
My trip into Aberdeen was to update the Foresterhill firmware to a newer version. It didnāt go well, and I ended up taking the node back home to fix it. Iām not sure if the issues were due to bugs in the firmware or the phone app, but the settings were all over the place and refused to be corrected. I believe Iāve resolved the issues now, so Iāll try to return to Aberdeen this week to get the Foresterhill node back up and running. Youāre on my node list, so Iāll try to message you if I see you online.
That explains a lot! Your Foresterhill node was a godsend, it must be on the top floor or something. Hope you get it fixed, it could turn out to be a critical node in the Aberdeen(shire) mesh.
That is my Foresterhill node back in service. It is nestled between skyscrapers which creates a few dead zones but as new 24/7 nodes are appearing, coverage is improving.
Thanks for this. My poglad node got repurposed as Linksfield to fill in that area, and my personal node is now MM0OQR-L8.
Weāve got loads of Newcastle area nodes showing up today. I think theyāre being picked up by the guys near Peterhead due to tropo along the coast, and then being relayed to Aberdeen via Foresterhill.
There was a significant aurora event yesterday, and another one occurred last week. I have an aurora monitoring app on my phone, and the Newcastle nodes appear and drop out in perfect sync with the aurora events (day or night). Foresterhill has direct connections with the Newcastle nodes with no hops, but a building obstructs my angle from reaching further inland. With binoculars, I could probably see the Peterhead area nodes.
I have been considering whether I need swap out the Foresterhill antenna to increase the gain slightly but still keep it omnidirectional. However, I cannot increase the gain too much as at narrows the vertical spread which means that at my height above ground, I would be transmitting directly over the nodes in Aberdeen and they wouldnāt pick me up.
I use an aurora app as well, but it said I had a 0.3% chance of seeing it yesterday lol. There was definitely tropo though, all along the coast from Fraserburgh to below Newcastle. On Monday it stretches between Aberdeen and Manchester, according to the forecast, so who knows what might happen.
Regarding your possible antenna upgrade, itās all very interesting. What Iām finding though is that with just handful of nodes in the area, the performance degrades exponentially with each hop. A location request or trace route to Foresterhill comes back in under a second, as expected, but try it with a node received directly by Foresterhill e.g. 125YAGI and the time increases to 25+ seconds. Another hop away and it can take well over a minute to return, and text messages can take multiple tries over several minutes before they go through - if they ever do.
And yet my node is basically talking directly to Foresterhill, with good positive SNR, and Foresterhill has proven great connectivity throughout Aberdeenshire.
Compared to many places, we have a small mesh here so far. Is the algorithm already letting us down so badly?
I have seen similar issues and have been thinking about what could be going wrong.
The link between Danestone and Foresterhill is strong and fast, likely because we are both radiating towards each other in line of sight, using vertically polarized antennas.
However, the link between Foresterhill and the 12.5dBi Yagi is a different matter. While this connection may also be line of sight, a Yagi is highly directional, and since online maps show it only neighbouring Linksfield, I suspect it is not aimed towards Foresterhill or Danestone. Consequently, we may be on the less effective side of the beam. Additionally, most users tend to mount their Yagi antennas horizontally polarised. This misalignment likely results in some packet loss between Foresterhill and the Yagi, leading to slower/failed communication. Adding a third node with a similar issue into the path then there will be delays on the way out, and all the delays repeated again on the way back. It could look exponential.
A Yagi is good for a long-distance dedicated node-to-node link, but I might consider adding an omnidirectional REPEATER node alongside it. This would ensure that all packets transmitted via the directional Yagi are also repeated in all directions, benefiting everyone with the Yagiās gain, and removing the Yagiās deaf side. However, we might all have to increase our hop count. Iāve noticed nodes appearing at Foresterhill with hop counts reaching seven.
There might also be an issue caused by ROUTER and ROUTER_CLIENT configurations. I have read that that if node1 is out of range of node2 and uses an intermediate ROUTER node to establish a link, that route is considered static. If node1 then moves to a position right next to node2, it will still attempt to communicate through the ROUTER node instead of directly with node2. It might even be the case that node1 can no longer reach the ROUTER node.
This becomes even more problematic when mobile nodes are configured as a ROUTER_CLIENT and are driven around the mesh creating a rather chaotic situation. Also, if nodes have directional antennas, and they are on rotators, they shouldnāt be a ROUTER as they are not static and will appear/disappear to other nodes. I understand that the ROUTER_CLIENT configuration is being phased out in the latest firmware, which I anticipate will lead to improvements in every mesh network.
My antenna upgrade idea presents a conundrum. LoRa operates on a line-of-sight basis, and I know I am achieving this as I am being picked up from 40km away due to my height. However, with increased height comes a disadvantage. My setup also picks up signals from everyone else, not just LoRa. The base noise level at Foresterhill is high across all frequencies - HF, VHF, UHF. I suspect that most LoRa boards do not have robust front-end filtering, and many are tuned for 915 MHz. Given that 4G mobile frequencies (800MHz and 900MHz bands) are adjacent to 868 MHz, I wonder if my node is being desensitised by the numerous mobile masts and phones within range. The antenna at Foresterhill is marketed as 5dBi, but since it is based on a dipole design, I expect it to be closer to 2.5dBi. I considered adding a 868 MHz bandpass filter, but the insertion loss would exceed 2dBi, effectively negating the antenna gain. Unless I come up with a good solution, I may choose to leave everything as it is.
I visited my Foresterhill node today. In the three weeks since I rebuilt the Foresterhill node, the node list has grown to 105 entries. We have a larger mesh than the online maps suggest, as I can see local nodes that do not appear on any of the maps.
I also took the opportunity to test my criticisms regarding ROUTER configurations from my previous reply. I placed two of my nodes on the same table at the Foresterhill site. One node was Foresterhill, which has always been static, while the other node was Taixali, which is typically mobile and usually sitting in the back of my car. Both nodes are identical and configured as CLIENT.
When I sent a message from Taixali to Foresterhill, just centimetres apart, the message was successfully delivered. However, when I traced the route, it showed Taixali > 12.5Yagi > Foresterhill. I expect that this occurs because the 12.5Yagi node is configured as a ROUTER and that overrides the direct connection. As it stands, Taixali cannot reach Foresterhill unless the 12.5Yagi is reachable by both nodes.
I also traced the route from Foresterhill to Danestone, which showed a successful direct path. However, every attempt to trace the route from Taixali to Danestone completely failed. I suspect this is also due to the ROUTER configurations again. As Taixali has moved about, it could have previously encountered a ROUTER node that could establish a path to Danestone. Now, when Taixali tries to connect to Danestone, it relies on the availability of that intermediate ROUTER. Unfortunately, that ROUTER may now be switched off, out of range, or its directional antenna may be pointing in a different direction.
So, I have two identical nodes sitting next to one another on the same table, each experiencing different problems reaching different nearby nodes that are all in perfect line of sight.
In my view, the ROUTER algorithm does let us down badly when the mesh includes mobile nodes or rotating directional antennas. In all meshās, every node should be configured as CLIENT by default, and ROUTER should only be considered in very rare specific situations.
Valuable insights, there. Thanks for testing. It seems ridiculous that the mesh has got itself into that condition and the firmware canāt detect that it needs to use some other route then the unavailable one.
Iām interested that Foresterhill is seeing so many nodes, yet isnāt retransmitting their beacons across the mesh. Is the hop limit being reached or something like that? How can anyone hope to contact them if only Foresterhill knows theyāre there?
This is all one big educational exercise, which I suppose is a good thing.