#US - Utah
1 messages · Page 5 of 1
A direct strike will hopefully just go through the ground connected but 🤣 its gonna melt everything so destroyed anyways
The numbers would be hop limits for routers and clients which would be separated.
So clients would wait until all the router hops are possibly consumed.
- router hop limit 5, clients all wait 5* max router response delay.
If routers reply within 1s window, it would wait 5s before clients take turns.
Could also work the same if client delays defaulted to a much longer time that router delays
So usually all the router hops are finished by the time any clients try
So really the only cost is latency
Congestion too since it results in more traffic when functioning correctly
Sent a test message on the bridge overpass at 106th south. Should be in full view of both Nelson Peak,and POTM. I don’t see any radios.
I’d theorize that they are too close to other equipment, but potm is far from towers?
I can talk to both all day from my house. I was just closer to POTM this time. NPR should also have been fine. No idea.
I feel like one answer is going to be: get at the location of a node and observe packets to see if it’s device specific?
Maybe test ducking below the hill so trace routes must use the router. The peak over and run the same trace route from client?
It's not just you. I could traceroute them directly pretty consistently this week. For some reason I can't today. Both are recent on 'Last Heard' but all traceroutes fail.
There are a painful number of routers up by me according to the app. In the Layton area alone there are 6 router_client and 4 router .. Which seem different than client and client_mute. What are router_clients?
Former config option. Now router doesn’t allow WiFi or bluetooth. It used to
Too many people chose this option, so devs changed it to make it less appealing.
I can see why. Jesus. And yeah most of them seem to be on 2.3 firmware not 2.5
Yup
This is what motivated our switch of MF
Folks think they “need” a router, because they saw it on YouTube and haven’t read up. Happens a lot
Over 25% channel utilization is considered “getting congested” and we were routinely over 50%
Got ya. The problem is becoming clear to me now. This is unfortunate. Its causing me to rarely see NPR since all my trace routes keep bouncing between those routers first
Exhausting my hop limit where it shouldn't
I guess I could ignore a node right?
You can, but it won’t ignore you
If you are seeing other routers up there I assume you aren’t on channel 51? You won’t see NPR at all if you aren’t on its same channel
I'm on 51. Theres a cluster of routers near the mountain side neighborhoods east of Layton Parkway router
I didn’t realize anyone up there was on this discord and switched with us. I’m surprised
What are the chances they’d have randomly chosen what we did?
Who knows. Its possible. Especially since 51 is a high channel and zero defaults to 20
And people when choosing different channels default to the high and low points
They've got to be on here. 51 randomly out of 104?
😂 yeah they might be on here or might not. Old firmware. Lots of 2.3 and 2.2 on the telem data. I mean have you tried just reaching out to peoples nodes and saying hey we got Discord but we are trying to make "explain the effort" and see how people respond?
You'd figure they have a similar goal haha
Yes, we tried. It's like herding cats.
Yeah at some point I guess you can only do your best
Oh well. I'll just ignore their nodes. It really doesnt help me in my situation though. I tried to send something to nelson and watched the packet and it just went bounced in those and failed hahah
Rip
Meshtastic Bay Area group has the same problem (or worse) and have organized switching to MS with a custom channel name. I think it's going to happen anywhere meshtastic is popular enough. You will end up with poorly configured or extremely out of date firmware nodes.
Yeah pretty much I'll try to stay up to date on all my nodes but if I make something custom/private.. I'm going to do an airwave survey and choose the least populated and move it there.. To be polite
Your other issue will also just be trolls. People who sit in. See whats a problem. Poke it for fun and giggles so you have that element you have to deal with too once communities get big enough
So I do get it. Just thought I'd say publicly that I'm seeing too many routers north of centerville and if those people are here. Please change to client 😂
Your killing me 🤣
We had quite the fiasco with MQTT for months
Most of us weren’t sure if they weren’t just trolling at some point
Although if the possible channels are 1-104 and I have a given random distribution of peoples choices. I'd avoid +/-2 of 1, 37, 50, 73, and 100
Unironically when sampling a sufficiently large sample of people.. The most random number 0 to 100 for people is 37 the second is 50.. The third is 73 and the 4th is 100.. With the 5th being 1
What happened with mqtt?
I used to send out a “join our discord channel” message every few days.
There is a public mqtt server that the devs have. Thousands of hosts globally. They use it to test higher loads. Backward compatibility, etc.
If you downlink it we get inundated with useless traffic.
If we don’t find a way to get consistent coms here in the valley I’ll take a step back and just leave my roof node up.
This will become a thing I use when I hike/camp and that’s it.
I’ve been at this for a while. Lots of other folks to carry the torch.
I have like 10 radios at this point 🤦♂️
I mean don't do something that upsets you seems pretty sane. Especially if you've been working at something and others dont seem to want to do the same. If it becomes too much of an issue for me I'll literally put my own stuff up.. Different frequency, different channel, different rates and forward local only. I just popped in to see what was public and how I can help. I'll keep my apartment one public though and follow what y'all do
I would like to point out that if one of those routers sees another of those routers repeating a packet then it won't repeat it at all
And if the far away router sees it and repeats first then another doing it won’t cut down on your hops.
I’ve considered making a long FAQ, and a current state, board for newer or interested folks. Just too busy lately.
Repeaters and routers act the same. Repeaters have less configs
That would be awesome man. I'd appreciate it if it was made at some point haha. Considering one of the routers is named rizzmedaddy86f3 I'd say hes a troll 🤣 but could be innocent too
Although that name. Has me skeptical AF 😂
The default channel went from barren wasteland to trollville in like 4 months. It was crazy 😂
Thats what happens when you make something public these days. Its why I honestly had second thoughts from the start on doing anything public with my buddies who are interested. One of them said this is a potential outcome 😂
Oh well at least its funny
And not mission essential
That said if I do make something mission essential. I'm sorry guys. I'll probably borrow the same real estate but won't have the same configs, channels, etc. I'm assuming meshtastic is okay/adjacent frequency friendly with that type of thing?
Hmmm weird. Ive been on 51 all day. Can't send. Can't see NPR but I should be able to see it. I should be inundated in signal. I have a 10 dBi super well matched type N. Vswr max being 1.2 across the spectrum for transmission as well so it should here me. And before those who know to ask. Yes I did calibrate with the sma to type N cal file and standards. :/
Come on t beam you got this 😂
Maybe just am unlucky today? Its been cold.. rain snow mix up north
I do get tons on 20 though
Hmm yeah I guess up north I'll have to stay on 20. 51 works in the city though
Until there is wider adoption you are correct. We “just” started testing channel 51 when you joined.
So I got one of my nodes up last weekend. The telemetry data is stuck in Celsius. Any idea how to un-stuck the telemetry? I've unselected convert to F and saved and redone it, but it just always stays in C
I was out at a park near our house in Millcreek. I was able to consistently ping poster's nodes and mill_mesh nodes through NPR which was great to see.
There isn't a display, I'm viewing the temp remotely. I tried it anyways and it's s still C
I'm so excited for a well organized Chan 51 haha. I noticed someone in fed heights has a router. Its really strategically placed. For that weird corner. Is he participating because it would really make the hops to my buddy in the avenues shorter from up north
I'm 90% sure that is Pete Ashdown, as I've seen his mobile node at the same address. He also placed the Xmission router. He commented on here about it a few months ago.
#1197577977781821541 message
Yeah I've done that a bunch of times
Do you have the CLI installed? You could try that way.
meshtastic --set display.units IMPERIAL
oh, ok. I don't have anything with a temp sensor, just saw this in the display configuration and assumed it might work.
The "ingeniously" placed Xmission and Federal Hgts routers are a strong reason why I left ch. 20
Honestly I'd love to know the hardware. Other than what's broadcast. Those two are the reason I can talk to SLC. For some reason despite even slapping a directional on and pointing it directly at nelson 😂. I cannot get nelson to be heard on my tbeam echo. If I want that I'll need to go purely directional sadly and probably an amplifier then filter.. Then into the box. Idk why I can't hear nelson on 51 with direct LOS. Like I can see nelson with my insane monocular 😂. But nope. Can't be heard. Can see you and xmiss router though
So if anyone knows the hardware. That would be great because its clearly strong.
I just moved to a new place with direct LoS to Nelson (19 miles away) I’m going to be testing station g2 with a 12dbi alpha yagi and an amplifier . 😂
Is Nelson currently on LF 51?
Does anyone know how much attenuation glass windows usually add? 😅
It is still on LF 51. I last saw it 42 min ago and can't seem to traceroute it rn.
I wonder if you have a direct connection with my work node in NSL? It's about 4 miles North of Marmalade, but depending on where you're at, probably no LOS. It's named DSR2.
Its very cold up there right now and snowy
I just got a traceroute through it, so it's definitely up and running.
!eb4c89df --> !ac7f43c3 (?dB) --> !0e81d07d (?dB) --> !db59904a (?dB)
Route traced back to us:
!db59904a --> !33676da0 (6.0dB) --> Unknown (?dB) --> !ac7f43c3 (1.25dB) --> !eb4c89df (6.25dB)
I (MM / 6da0) currently have a 0.3 SNR, 95 RSSI connection to NPR, but it is often unresponsive when I traceroute after a long interval. Does it have a power option set that puts it into a sleep mode?
I’ve wondered the same. @hallow dagger will know
First trace to npr failed, second succeeded
Same, responded super fast just now.
I think I might be getting packets. I just got a trace from cottonwood with my antenna facing NPR
Yeah, roof sees it with its typical 0-1 SNR and 95-105 rssi
Anyone seeing d968?
For me, LF has zero improvements over MF. It’s a different story for you @leaden crow , right?
If it's an improvement, it's very minor. It's sixes to me whether we switch back to MF or not. I was just going to leave that up to someone who had a strong opinion.
Same. Just curious
The biggest dilemma is that it's some work for gdane to switch them back, but other people have even more remote nodes that would need being switched from MF > LF51 if we stay.
We should wait for gdane to respond, see if he has any preference. If we stay on LF 51, I will send some messages on MF. I know we left at least a few people behind.
I think that gdane's routers are on v2.4. Does 2.5 allow remote administration using packets? Any reason why Meshtastic doesn't use watchdog timers to recover from crashes?
You can use the admin channel on 2.4. 2.5 just adds more secure admin communication
So does gdane need to physically visit the nodes to change them, or just the time and effort hassle doing remote changes?
Edit: not physically visit the node, but had difficult time getting direct connection when switching from MF #1197577977781821541 message
speaking of power options. Im kind of wondering what if any best recommendations people have for power saving and sleep on a tbeam s3 core and a rak 4361. I havent experimented enough but if someone found power save settings (maybe wake on packet?) and gps settings and neighbor info settings for those devices that save a decent amount of power.. id love to know. It would be a bit easier than me testing constantly. The raks way outlast my tbeams but I kind of expected that from a nordic semi m0 core
When I trace to DSR1_RAK, it's straightforward there MM->NPR->DSR1 but the return is DSR1->Meshtastic ffff->Meshtastic ffff->bp_roof->MM. It is also slow and unreliable. I hate to harp on Meshtastic, but why not use a router coming back? Who or what is Meshtastic ffff?
I see the exact same thing tracing to you. Outbound is one hop from NPR. Return is through a node named Unknown (?dB)
#1197577977781821541 message
So clearly the closest route is 1 hop through NPR from either direction.
I'm guessing that Meshtastic ffff is the same as Unknown (what I get from the CLI)?
Is this a node using 51 for a private channel that handles any packet it sees on 51? Is that why it doesn't show up on our (LF) node list? If they changed Rebroadcast Mode to KNOWN_ONLY, then they wouldn't be intercepting our traffic?
DSR1_RAK? I don't know why it wouldn't show up in your LF node list. No private channels, only the public/default one with Key AQ== on channel 51
slcmesh shows my nodes and messages. I think it's a you problem. 😆
Oh, I misunderstood, you're asking about the unknown node
We should probably ask the Devs. @vital hemlock might know which one would be best to direct the question
It's also possible it's a repeater? If somebody configured one of those, my understanding is they act as a router but don't reveal themselves in node lists
I turned on the debug logging. I'll try some traceroutes and see if I can figure out anything from the logs
I just think it's a problem if we a trying to set up a reliable network based on well-placed routers, but unknown forces are muddying the waters. We should think about all switching to KNOWN_ONLY to at least limit that variable. It could reduce traffic on the routers.
Well if someone wanted to throw wrenches in our plans. Making a bunch of repeaters is probably the best way. 😆
Interesting idea, I'm going to leave my debug log on for a while. I actually had it on, but one of my upgrades overwrote the config.yaml with default. I was looking at the old tracefile and it was 84MB for probably like 1 week on Long Fast 20. I think if we had a ton of private traffic, or malicious repeaters, we could probably figure it out with the tracelog. Side note, I need to switch my roof node to an RPi.
Repeaters don’t announce themselves. But if you’re hopping through a node you simply don’t know the name of sometimes I’ve seen it show as a repeater. Let me find an example.
I doubt most users will deploy repeaters. They want to have more control most of the time. Also, if someone has competing routers you can add them our routers ignore list. However, I’ll grant it’s easier to just say “local only” and not have the default channel.
Where is the router ignore list?
Is it for Chan 51 or Chan 20?
It’s completely separate from channel or modulation
https://meshtastic.org/docs/configuration/radio/lora/#ignore-incoming-array
Understanding the LoRa configuration settings on your Meshtastic device including region, modem, hop limit, and more.
But I don’t know the cli command to pass it. I’ll look in a bit
NPR doesn't have any sort of sleep mode enabled, at least not that I know of. There's nothing funny in hardwre, just direct battery and solar to RAK4631. Are there any "sleep" modes in firmware that maybe I'm not aware of?
It's not a big deal to change. I could probably do it from home if I did NPR first and then POTM. I personally don't see any reason to stay on LF-51 if it hasn't provided any benefit. NPR obviously has other issue(s) that weren't resolved by the change. At least MF had an added speed benefit. I was really hoping that @slate urchin would be able to test on LF-51 first since he seemed to have the only consistent/reproducible issue routing over MF vs LF.
I have the issue as well, when I'm at work today I'll do tests with 51.
I left it on 51 overnight and it's definitely seeing more nodes now than it did on medium fast
My first Trace route attempt was a success. I've pretty much never been able to do that on medium fast
Thanks. I'm reading it now. Just wondering where the list is
On Android is more obvious than on IOS. I'm not seeing it in IOS
I'm also getting more persistent trace routes on LF:51. This was an interesting path this morning from Millcreek.
Android traceroute only supplies the intercepting node's name. The ignore list seems to require a decimal number identifier. How would I find that for the node "Meshtastic ffff" since it is not in our node database?
You can also do the traceroutes from the CLI. I hear the python cli has the most robust logic.
We just approved an iOS PR to fix traceroute bugs for how it's displayed - to be released soon. (very possible with current logic for numbers to be totally wrong on iphone)
I was unable to get traceroute responses from NPR with this setup. (RX was SNR -18.25, RSSI -118)
I will attempt the same with an open window and remove the glass attenuator (window).
I could only get the CLI to return a forward trace. Unfortunately, the intercepting mystery node happens in the return path. These problems will be fixed by switching to KNOWN_ONLY or possibly LOCAL_ONLY, especially for our routers. Each local node may wish to add particular nodes near them (up to 3) to its ignore list. For instance, dragonslayer may wish to add Meshtastic ffff (whatever its decimal id is) because this node is preventing his outgoing messages from using our routers. Doing this (in addition to the KNOWN_ONLY setting) slightly reduces traffic congestion by preventing Meshtastic ffff from shouting in the first place vs. ignoring the shout because it is unknown.
I will try to do this, but I don't know how to find the decimal id of that unknown node. I just got another trace from you where the return path was through Unknown, but I don't seem to have anything in the debug log for the return message.
Route traced towards destination:
!eb4c89df --> !ac7f43c3 (?dB) --> !0e81d07d (?dB) --> !33676da0 (-3.5dB)
Route traced back to us:
!33676da0 --> Unknown (?dB) --> !ac7f43c3 (1.0dB) --> !eb4c89df (6.0dB)
I'm confused. It's either not being logged, or it's in the encoded payload?
{"bytes":"85742D5C996A2AB44A289EC2861A955A176F0A761C62C9F93E125BF7AD8915951616B050638DBE102B9273D9246D8531474060867F41D6C0895D37D0C3539C8DAF343513494DB4438254","channel":8,"from":862416288,"hop_start":4,"hops_away":2,"id":652868292,"rssi":-57,"size":74,"snr":6,"time_ms":3174366,"timestamp":1730745012,"to":3947661791,"want_ack":false}
{"channel":0,"from":862416288,"hop_start":4,"hops_away":2,"id":652868292,"payload":{"route":["DSC2_RPI","DSR1_RAK","Nelson Peak Router","Mill_Mesh"]},"rssi":-57,"sender":"!eb4c89df","snr":6,"timestamp":1730745012,"to":3947661791,"type":"traceroute"}
Thinking about this. You ignore them, but they dont ignore you. Just because you route a packet to someone else doesnt mean they wont also pass it. You just ignored their reply to you.
I assume that a few of these wont really be a problem except for in a few smaller usecases.
Also, I thnk the ffff just represents a node for whom you dont yet know the ID of. But we can ask the devs.
I'm confused by that though, I must be wrong. The node ID should be "in" the packet you recieved. Though I've seen this a number of times before. We do need to ask.
The ignore flag actually adds the IDs to the message header. This is why it is limited to a MAX of 3. If a node sees its own ID in the ignore header, it will not respond.
Oh interesting, it didnt used to work this way. I've had large ignore lists in the past because we had so many folks just "making noise" on LF
This is a better iteration.
Can we just make the router Mesh private for now and only send out keys to people as they confirm the location for their router(s)?
That way we incrementally add more and more nodes and if someone breaks it we tell them to disable it.
ffff indicates that it's a "repeater". We should discorage poeple from using this mode since it prevents us from seeing what's happening with the mesh.
I actually just saw some details in a PR that was handling how to display repeaters and it's determined when nodeId == 2^32 - 1 AKA 4294967295
Great find, Jacob. With an empty ignore list, the trace was MM->NPR->DSR1 ... DSR1->ffff-MM (Ironically, the repeater is working well for this route today). I then added 4294967295 to my ignore list and got the same trace. Maybe the return trace doesn't add my ignores... so it's not a definitive test. dragonslayer will need to try it.
Note that 4294967295 in the return route may also mean it's a node using old firmware (<2.5), because only then the route back was introduced. Also, it might be a node that can't decrypt the traceroute.
I haven't finished unpacking, but I'm already testing a NPR specific Yagi. 😄
Does anybody know the parameter to set the ignore list from the command line? I can't seem to find it in the docs.
TBH, the shooting tripod makes for a super sturdy antenna holder. Just needs a 90degree adaper for the yagi.
Do nodes display uptime in any telemetry data?
What if the node is just getting absolutely BLASTED by wind and it's got flaky wiring or antenna connections. 😛
Yes, I see it in my tracelogs
{"channel":0,"from":2894021571,"hop_start":4,"hops_away":0,"id":2022885209,"payload":{"air_util_tx":0.578555583953857,"battery_level":49,"channel_utilization":1.48000001907349,"uptime_seconds":72056,"voltage":3.71700000762939},"rssi":-55,"sender":"!eb4c89df","snr":6.25,"timestamp":1730751048,"to":4294967295,"type":"telemetry"}
Well, I'm not getting any ACKs from NPR from either traceroutes or DMs.
I'm quite tempted to attempt to run some long term diagnostics to see if there's any trends over time for success vs failure.
Gonna remove the amplifier and a 15' cable. Going direct to antenna.
omfg, that did it!
I got my first ever traceroute. Just took direct LOS and a 12dbi Yagi. 😂
It wont ACK dms but it's giving me really clear signal.
Okay, moral of the story is probably don't trust a $2 amplifier from ali-express.
It's not responding to my traceroutes or messages right now. Give me that yagi 😄
Hey, got one through. Just took a few attempts
Sometimes you just have to prime the airwaves I guess 🤷♂️
rssi -98, but SNR is -15.75
I noticed that the summit was covered in clouds until about 1hr ago.
oh hell, yess. I can traceroute POTM now. 😄
│ 5 │ Meshtastic d968 │ !a2ebd968 │ d968 │ UNSET │ UNSET │ N/A │ N/A │ N/A │ 101% │ 4.99% │ 0.03% │ 6.75 dB │ 2 │ 0 │ 2024-11-04 15:56:25 │ 7 mins ago │
No Serial Meshtastic device detected, attempting TCP connection on localhost.
Connected to radio
Sending traceroute request to !a2ebd968 on channelIndex:0 (this could take a while)
Route traced towards destination:
!eb4c89df --> !ac7f43c3 (?dB) --> !0e81d07d (?dB) --> !a2ebd968 (3.5dB)
Route traced back to us:
!a2ebd968 --> Unknown (?dB) --> !ac7f43c3 (-0.25dB) --> !eb4c89df (6.75dB)
^ this is extremely exciting to see
Yeah, it looks like whatever's going on the mesh is actually functioning.
I'm going to start testing the signal from my non-yagi nodes.
There is something going on with this. After several trace failures, everything starts working. I was wondering about FreeBirds comment on power saving yesterday. This post is super old, maybe not relevant at all anymore, but I'm tempted to put a router on some irrelevant channel and see if I can determine anything.
https://meshtastic.discourse.group/t/router-and-client-what-is-the-difference/7060/5
Let me know if you want any help writing scripts to test it.
Should be pretty simple right?
Park em near eachother and just spam messages to one of the nodes and track failures over time. If there's windows of downtime, something's up and we need to loop in the firmware devs.
exactly what I'm thinking. probably don't need any help. we'll see if I get anything useful from it in a few days.
@shut gulch do you know by chance? I recall you speaking to power issues in the past
You might also try spacing the messages out. Start with 10 seconds and keep incrementing up until they start having failures. If you reach 30 mins then work your way back down?
I wonder if more messages on the mesh keep nodes awake. It shouldnt be like that, but it could be
Another thing of interest. I've seen the SNR pogo'ing today. Most of the time I'm -2 to +2 SNR to NPR and POTM, but I've seen greater than -15 randomly today. That's a lot of noise, for us.
Does anyone know if there's any easy to use utilities for tracking node metrics over time?
I'd be curios if this is weather dependent. I did see NPR getting absolutely blasted earlier today.
Now it's sitting in the sunshine.
Anyone know how accessible NPR actually is in the winter? (is it gonna be as brutal as I think it would be to upgrade the firmware so that it gets snrTowards?)
side note: the over-the-wire snrBack needs to be divided by 4 to get the real value
I think the note is pinned to this channel but I host a small grafana dashboard. It's filled via data that I collect through the serial port on 2 nodes, my roof and also poster's node. It will have any metrics that devices send across the wire...that one of our two nodes sees.
Until we set the routers to known_only, private channels could be streaming Netflix through them.
I don't think traffic is the issue currently. NPR is reporting 1.5% utilization.
Life has been crazy for the last month or so. I haven't had time to follow the conversation here.
Are we running med/fast or long/fast nowadays? And which channel?
Yup, has been busy with testing a few different options. There's a number of routers currently being tested on LF on channel 51.
We had previously been trying to switch to MF/45, but felt we might have lost too much signal penetration. (not sure how many stayed on it)
We are currently seeing that MF vs LF might have been less critical for the routers than we had anticipated.
My guess is that we might land back on MF/45 if we ever figure out what's going on with the routers.
You know. When I wonder if there's something to this about SNR with short turbo when we have high amounts of baseline traffic/noise.
^ Perhaps running SHORT_TURBO gets some advantages if the issue is airtime congestion that makes up for the noise resilience.
I might try running this test again, but downtuning my radio output power to be the minimum value possible.
Corrupted protobuf?
oh, I think he's correct. It might be that the code was incorrectly assuming repeater when it was actually just an unkown node. (repeaters never broadcast their name so they are always unknown)
I don't think you'll see anything from up North on either MF or LF (channel 51), but you could try. @maiden wraith has had some bad luck with his Francis Peak Router, so that's probably down until Spring. I think it would be the key bridging all the nodes up North to the new Nelson Peak Router.
Hate to be the bearer of bad news, but that's a compressor, not a generator.
haha, I assumed
I'm not judging.
You're a better man than me. I heavily use the edit button. I think I edited my edit yesterday.
So, I've never really looked at Meshtastic data over the air before now.
a 200 byte message sent on LF is causing 4s of airtime.
testing the same on SHORT_TURBO now.
I should have brought this up more prominently in the past. This is why MF seems so much snappier.
LS is…slow
So, I'm not even close to seeing RF from the RX side of far away nodes with my SDR.
Nice to see a few more nodes poping up today on the map.
I bet several of them are my test cluster getting meshed through my yagi. 😄
I agree. His node made a huge difference. Thank you
Awesome, Thanks!
TBH, the router NRP was at 10% utilization last night at 3am. (yay moving stress insomnia) This node count + our node telemetry usage might already be near the node count capacity.
Do we want to try any other experiments with the routers?
I'm very curious if doing SHORT_SLOW, SHORT_FAST, or SHORT_TURBO could result in improvements from Meshtastic related congestion. (and how much real life reduction in range we see)
actually, bad packets seems like something to target testing next. (since it could be caused by non-meshtastic issues too)
This could probably be done separately from any router changes.
I'm open to trying some of the SHORT_* modes on POTM but I'm reluctant on NPR for the time being when there's the possibility of losing connection from the valley entirely due to reduced range. I don't think that would be the case but need to prove it out on POTM first.
Makes sense
Another thing to note @white field , the nodes will "chatter" less as the mesh traffic increases. The firmware is designed that way out of the box now.
To summarize experience on LF-51, seems like there was marked improvement for a couple of nodes, moderate improvement for a couple more, and the rest didn't really improve from MF. I personally didn't experience any improvement over MF. I had a lot of issues with MF initially after NPR came online but that improved after a couple days. Not sure if that was something I changed on my client node/app or if it was the apparent transient issues people are seeing with NPR, but when it was working it was nice and fast.
Anything I missed?
You summarized what I've heard so far perfectly. It's interesting to me that once more radios were using NPR it seemed to "wake up", whereas before it was intermittent at best. It's very responsive now. Strange.
Yeah I really don't get that. What I wouldn't give to have an rpi up on Nelson to observe the MT logs.
I'm still highly suspect of interference from the tower next to NPR. Doesn't seem like it would be intermittent though, but who knows how those antennas are used.
Nice score! Could come in handy for a multi-radio deployment. Would need some serious battery/solar though, and a mobile data connection.
I ran a baseline overnight. Ran a trace every 5 minutes. Only 1 of 180 failed (Aborted due to timeout), and only 1 other took longer than 3 seconds to complete. I now have it incrementing the sleep by 5 seconds every loop. Maybe if I'm seeing failures at 30+ minutes., I'll try a few quick repeat traces right after the failure.
Do client nodes look at the most saturated node they are meshed to?
Otherwise the ones in the valley won’t know to be quiet when NPR is overloaded
I’m very uneasy about reliance on tuning for this despite what others claim.
I’d rather see baseline utilization near zero for higher on demand reliability
If you look at the logs there are always messages about "mesh busy, not sending" or the like
That's them rate limiting
That feels like something our mesh should not hit under nominal operating conditions.
I'd rather everyone just configure so that there's tons of breathing room.
I want to be able to reliably RX/TX packets no matter the type.
If I have a node that's 4 hops-away and the mesh is congested, the packets have to be very-very lucky to get passed along.
I've setup my remote router nodes in a courteous way IE, "very long time windows between analytics/position/node_info". Due to this, they are extremely unlikely to be heard by other nodes in a congested mesh.
If I were to configure these with a "loud" config, nodes that transmit very frequently will get heard, but they will prevent "quiet" being heard due to congestion.
@main stag , fix your nodes!
I need a higher roof
I really need to figure out a way to mount my a large dipole on my roof. Tho, first I need make friends with the neighbors and then figure out if there's roof access. I'm a renter, HOA be damned! 😂
I didn't see @vital hemlock test at 10:13 but I did log Mill_mesh's response to it.
My node is two blocks from teh top of the hill I live on, I'm on the south side. If I put a node on top of the townhomes at the top it'd see so much it's dumb.
I'd leave it as a client, replacing my roof node
Hmm, I'm gonna test out a diff antenna now. I have 100% links with my yagi 12dbi, I wonder if 5dbi dipole will link with the same hardware.
Is Short_Turbo still on 914MHz? 51 defaults to 927MHz on S_T...
Guess you’ll need to move. 🤣
I can trace NPR successfully without using a Yagi. wooh.
Pardon me while I do some spam mesh traffic to test the difference between hardware. (sorry mesh, gonna use DMs so acks are more reliable and it finishes with more reliable data)
Oka test running. DMing NPR every 30-60s.
Gonna collect a bunch of SNR data and then swap antennas to get actual performace characteristics.
Well, the verdict is definitely. Get an alpha-915 Yagi if you have signal issues and want to talk to mountain routers.
SNR went from avg -8 to avg +5. Huuuuge difference.
Will post the plots when I finish collecting data.
I'm also only testing with a Station G2. I'm going to run the same test later with a RAK4631.
Interestingly, with the Yagi, I still get ouliers where the SNR drops back to -13.
link me!
Yeah, I've seen this a bunch of times. It'd be interesting to leave an SDR plot running and figure out what the "blat" is that raises the noise floor so high intermittently. Bizarre.
Whatever that shit is it's broadband per what the SDR can capture. It totally saturates from 900-925MHz.
Histogram of SNR measurements a Station G2 after sending messages to NPR.
12dbi (24") Yagi (blue) vs 18" 5dbi dipole (red).
Those damn blue outliers for the yagi when it dropped to -15 and -14.
I'm curious what caused those.
1000w pager system!? jEEEZUS
Wtf
That would definitely explain the SDR plot exploding.
And warms nearby burritos
watch there be a tower for one on NPR.
I foudn it
I'm right next to this fucking massive system.
300-1000 W
Well, that explains that.
Maybe we should use a lower part of the spectrum. As far from 929MHz as we can be.
There's another on farnsworth.
it's only rated to 100w
Soooo, MF channel 1?
now that I know about this FCC database I really want to figure out how to lookup what else in use on the regional towers.
oh this is an interesting find. Trains are using 28W in these frequencies, and they have 20kHz bandwidth.
906 is dead in the middle away from train frequencies.
Tho someones running some 300w towers at 907.475. (25kHz bandwidth)
I feel like 905 MHz might be a nice zone to try from what I've been seeing here if we want to avoid high power interference.
LF 20 is on to something. 🤣
I guess 12 & 13 are closest to 905 MHz
The following configs look open if we want to try lower bands.
- short turbo - slot 6, 7
- mf/lf - slots 11 - 20
905.6125 is the center of the empty licensed spectrum.
I'm gonna crank my SDR up and check out this range.
I wish there was an ez way to measure, "THIS SPECTRUM BAD, THAT SPECTRUM GOOD"
It's interesting to see it here, the spectra is showing some data operating around 906.4 with a bandwidth of 500kHz.
I bet I have a short turbo node somewhere around my house. 😂
929.7 looks veryyyy busy.
getting paged:
oh, this might just be harmonics in the SDR bc my gain is all the way up.
I just did so. My SDR that surveys a wide band uses a discone mounted offset to a balcony.
They didnt give one sh*t
I am now unconvinced I learned anything about the spectrum.
It was fun to learn about ffc licensing and sdr++
You should try and survey the entire band for quiet spots first
Then take a look in the quiet spot what might be going on there
And remember the span determines the filter but thats post amplification so you'll notice more noise at high gain not just due to harmonics but also the nature of amplification. It amplifies noise too. Your filter for your span just (ideally) filters out anything coming in from outside the band but I'd the noise is already there you'll see it filter or not
Use only the amplification you need to see the signals 😆
TBH, I'm not certain anymore if I'm even able to measure with enough sensitivity to tell what's happening in the RX signal levels.
I have to max out all the amplifier settings and it still only gets 102db of gain. (at those levels I don't know if we are seeing actuall signal noise or just static from the hardware I'm using)
I'm a little tempted to try calling/emailing a company that is licensed to operate that pager system to learn more about what hardware they are using.
The thing that was overwhelming my FFT plots was actually a meshtastic device when I had my gain cranked so high.
I think you are right. If you look at filters like this:
https://acasom.com/products/915mhz-cavity-filter-for-helium-network-amplifier-filter-sma-type-high-out-band-rejection
They are not going to do anything for filtering noise within 902-928MHz (what we are assuming is a problem). It seems like the people using filters in discord search are using them in conjunction with amplifiers.
I tried a few times, my experience was the same. Gdane result looked actually useful though: #1197577977781821541 message
Hey, I see our mystery radio today!
It shows up in a trace as "bp! -> all -> DSC2"
All? Seems weird
I may brainstorm some more things to test, but overall I think this was a dead end. The router I setup has basically responded to every trace reliably. It is now at 45 min between traces. I will probably let it run for a few days just because.
Interesting. I have not looked much at what the client or mapping software does. Is there anything that will do rough positioning based on RSSI if you have enough nodes to do even rudimentary triangulation? Not saying we can find this, just curious. I need to message Arrow on MF. I don't know if he is on Discord, but his node in Holladay that was on MF could have been a hop like this.
Does anybody on here still have nodes on MF? I'm just guessing we are probably not going to switch back this Winter? I might send some messages on MF tonight to Arrow and anyone else I see on MF with LF 51 instructions. Thoughts?
A noise filter only helps if you have well characterized noise if you dont. theres not much else you can do except amplify. Filter with a sharper roll off (higher order odd element filter).. And match better ... Obv high gain helps... You can do some post processing but at some point you gotta deal with the real world and it has noise 😆
So I was not able to get NPR at all today at the university, downtown, in the aves, in draper, and I wasnt even able to send. I have not changed my config off long fast 51.. Im not sure why its occurring. Will check debug logs tonight
I am really surprised you didn't see it at even one of those locations. My work node in NSL is an old T-Beam SX1276 with a $2 Ziisor antenna. It is inside a building, and can still reliably see NPR. Not trying to rub salt in the wound. It seems like something strange is going on in your case.
I'm unsure as well. Its a very well matched 8dB gain antenna too. I checked on my vna
And sa
Software maybe? Reset. Reflash? ..2.5.6
If you are going to be at 801 Labs some evening. I could probably stop by and bring different antennas, radios. See if it's more location specific or hardware specific problems?
Couldn't hurt. Is it an NRF device? If it is I would do the factory erase. I am using the 2.5.9 stable and it seems fine.
😂 shiiit I think you just told me the issue
Back to 2.5.9 it is
Thanks for the tip!
@real cedar you're using a t-echo?
NP. I can't guarantee it will solve that, but I was having a ton of RAK reset problems. It seems like some combination of the factory erase and newer firmwares have fixed it.
https://meshtastic.org/docs/getting-started/flashing-firmware/nrf52/nrf52-erase/
I have a supreme and a nrf rak
Both exhibit the same problem on the same firmware
That position is incorrect. But yes that is one of my radios
I'm further south at the North part of the lake
Hmm, I’m not able to hit BCPd. Getting consistent 2-3 hop to POTM. (Hop limit is 5)
It probably does not have LOS right now. It's a portable node not a rooftop node. I don't have rooftop node yet
Ahhh okay
Anyone on MF seen traffic from Fools Creek Peak?
I’m thinking about trying to see if my summit nodes are up. I might be able to get them with a link from POTM or NPR. FCP also had direct los
I think that's the bug they were discussing in #firmware yesterday. I sounds like it only affects direct messages, not messages on the public channel.
Is traceroute a type of direct message?
No traceroute uses the default channel
I believe that direct messaging does too. It just starts with a public key exchange in 2.5+ and a less secure encryption in earlier versions.
Okay, serious question. Do we have coverage in the cottonwoods yet? 😛
I feel like the lift-techs and ski patrol could possibly be veryyyy chill about us placing solar client nodes on their shacks.
Cool thing about these is that if they only really see one router across the valley they could be very good about always repeating, but with the only RX node being the router and clients in the canyons. (not increasing raw valley congestion)
okay, updated assessment on how we might get "legit" coverage in the cottonwoods presuming enough cases of beer are donated to ski patrol + lifties
- 9990 - summit behind lift links to NPR, covers mid-mountain and above at solitude
- Milly - covers bowls of Brighton (can see 9990)
- Alta - hard to get access & coverage (lifts are not near good overlooks that also have LOS to 9990)
- Snowbird - Hidden Peak tram - links to NPR + can cover front or back bowl - can link coverage over to Alta (can see 9990)
I get a vague sense that if our goal is getting signal in/out of the canyon the best place placements would be nodes that can ONLY see NPR and act as a "client mode - backhaul"
Hidden Peak might be worse than a 9990 link bc hidden peak would see other clients in the valley repeating messages. (maybe it's fine to run as a router if it can only see one regional router)
I still plan on placing a node on Desolation Peak (above 9990). I didn't manage it before the Crest shuttles shut down for the year unfortunately. I was close, my hardware was basically done. I just had a bad BMS couldn't get a replacement fast enough.
shit, I have two spares and tbh could have given you a solar node.
That area is actually pretty safe in terms of avalanche terrain. It remains ski tour accessible in winter.
Yes, I imagine it would be easy for a telemarker
I fell over sideways the one time my friend took me to it.
I'm gonna go the above board route and start asking my friends who work at the mountain resorts.
oh you know what. We should probably not place anything near the resorts.
Passive avalanche beacons use 902.85Mhz.
Maybe even a bigger problem in the backcountry? Does anyone have a beacon to test with?
BC might be okay since the active transmitters use 457 kHz.
We are super far from those ranges.
RECCO is usually much larger equipment that's only operated by helicopter or ski patrol.
oh ok
Tho still, if I was wearing an avy beacon, I would probably not also wear a meshtastic node. It's too much risk of interference.
Phones that operate nearly a 1000Mhz away can cause huge errors when searching. (50-100% increase in distance error if too close to a beacon)
That seems not ideal
This is delightful for trying to see macro scale topology.
"show me the really tall stuff only"
How far is FPR from the paved roads?
I'm wondering if it's feasibly to hike it before the snows get deeper.
Anyone in here has the node SouthUtahRelay01 (!3108d267)?
And now I want bourbon 🤔
You should have a pour. Neat or ice?
Ice, for me.
I can jive. Usually neat for me.
I don’t recognize that node, and I keep pretty close watch. On our LF 51 test, or LongFast channel 20 (default)?
Just curious about this node. I have a mountain top node above Cedar City. Just picked up this node in our valley mesh (through the mountain top node) that's clear on the Arizona border down outside of St. G! My mind is blown right now.
Wanted to see if location was actually correctly advertised
Nice! I don’t think we reach anywhere near that far south just yet. Glad you popped in. These guys have been trying to grow your direction.
We've been building out Cedar City and Iron County for about 8 months now. It's coming along well but nowhere near what it's like up in the SLC valley.
Still on LF 20 here. Tried out MF for a bit but switched back
I’m gonna try to spend some time on the Skyline to Moab/Indian Creek node.
Found a friend to try a winter excursion in the Henries with me.
The skyline node shooould have LOS to salt lake mountain nodes if we sort out channel congestion.
I did some basic estimates and Denver would be possible with 6-8 hops. 😂
Anyone have thoughts about what mesh configs we want to standardize on?
I'm trying to plan what I should do with my remote nodes. They are currently on MF/45 and I'm trying to sort out if I should swap them to LF/51 or if we want to test swapping the routers back to MF or MS.
My vote is LF51
Want me to simulate some more mesh traffic on LF51? I could bring most of my nodes online and see how NPR + POMT handle it.
I think that would be great.
Sounds good, just as a heads up this will be a bit of a load test. I'm gonna setup all my nodes to report location more than they should. (to simulate what our mesh might be like with many more nodes)
experiment starting now
- 7 nodes
- 4 hop limit
- LF/51
- 5 min interval GPS broadcasts
Okay, all the nodes are configured. Might be good to test this for a few days and see what congestion looks like for NPR + POTM.
My nodes usually have 100% ACK rates from NPR, so let's see how this changes things.
okay, all my nodes are spamming away on LF/51.
It looks like there's a significant amount of traffic on that node now. My packets are bouncing around the valley before it gets to it but on the way back it's direct
"deviceMetrics": {
"batteryLevel": 101,
"voltage": 0,
"channelUtilization": 12.926666259765625,
"airUtilTx": 3.3204169273376465,
"uptimeSeconds": 2631
},
^ here's what my home node is seeing. I suspect NPR is under slightly higher load since it can see yours too.
This is the only client node, it retransmits if NPR doesn't ACK.
I'm very tempted to test true Mesh utilization with these spammy configs on different radio settings. 8 nodes seems to be enough that there aren't transmission problems, but it's definitely enough to be visibile on the utilization metrics.
I can bridge the local channel 20 to 51 if you want more, it's easy
How?
Open the flood gates. Let's try to see how hard it is to stop ACKs from functioning for NPR.
I have a radio on channel 20 that posts to my local mqtt server, then another radio on 51 that repeats it. They dont care about the modulation, just the channel name and the key
It would be useful to know what the limits are for the routers.
That is what I was saying to you guys about backhaul, a Pi Pico running mqtt might be able to do this so the router backhaul nodes use their own modulation
kk, standby
If we run online backhaul nodes we could probably even get away with SHORT_TURBO, but it defeats the point of it being "off grid". 😕
Not online, a "local" mqtt on a pi, at the node that hits the valley
I can draw it
ahhh okay
So we just backhaul using separate frequency + yagis + high capacity link configs?
I'm currently able to get traces between NPR -> POTM.
I suppose our goal should be to see how much spam it takes to break that connection?
I actually had traceroute to POTM hop through NPR. I think they are communicating better than they have in the past.
I was having mine traverse using BP_ROOF.
I love how we try to break it and it gets better. Why can't it be like this when we operate on LF-20. 😂
I'm seeing pretty consistent 5-6% channel utilization from observing NPR.
I'm apparently able to induce about 4-10% load with this test setup.
Ha interesting
I'm gonna try to make it worse by putting 2-3 nodes in client mode. They are hopefully gonna collide on some broadcasts, making it more spammy.
Maybe our problems have been related to multiple clients re-broadcasting concurrently.
Okay, just rebooted everything into CLIENT. Let's see how bad this gets.
Seems much worse for me, but I'm contributing to the problem by running traceroutes
I bet the signals are gonna bouce around like an echo chamber at my place until my Yagi takes a turn to retransmit.
48% was the first reading after reboot. 😂
Channel 20 nodes starting to show up for me. Anyone else yet?
They should all say (mqtt) after their name
It's definitely closing in on non-fuctional to me. I didn't get this message on my RPI
The ones in my house might see each other, but not be seen by NPR. Which are you looking at?
channel util. I'm seeing
a lot of them are from 10-20 minutes ago. I bet they are worse now.
mine are nodes 6-11, 16, 19
Npr still responsive for me
Hopefully it settles down after the influx @leaden crow
I'm very tempted to figure out how to TX arbitrary waveforms on the correct frequency with my HackRF.
Oh, we are trying to break it currently. 😄
My incuded load is:
- 7x nodes - sharing location every 5 minutes instead of every 60-120min like they probably should be.
I'm wondering if we really have enough capacity on the frequency to handle a big influx of people using the default settings.
We might need to keep LF/51 semi-private and ask that people be polite about only emitting metadata every once in a while.
also-brb trying to knock over NPR with a function generator. 😂
I think it would take a lot to recreate the LF20 problems. They had so many routers. Honestly we probably did them a service moving a bunch of nodes to a different channel.
And even with those problems, it still mostly functioned
YOLO
I feel like this might be the most realistic test of "poorly configured".
okay, yell at me if I forget to change any back after the next few hours.
hahaha, constant transmit after all the resets.
well it's not broken. slcmesh still receiving my messages.
Yeah, even at constant 15-30% channel utilization I'm able to get DMs in/out.
This isn't the fowarding node tho. I'm still very curiuos what this is doing to NPR
I just got a trace back from it, and it's still showing only 5.9% util to me.
I'm screwing around with some ShortFast configs locally. Load should be off NPR now.
The max I've been able to reach actively spamming around with SF has been 7%.
Alright, trashing the mesh complete. reverted all the configs.
To MF? Or just the telemetry reports
Should I turn off the channel 20 traffic?
Just say the word when the test has run its course
I've reverted all of mine.
TBH, I stopped looking after a bit. Anyone record NPR airtime + congestion?
This is all the telemetry I received from NPR today that included channel_utilization.
edit: oops this has older entries. I'll decode the epoch timestamps and see how many I have for just today.
I have a ton of logging for NPR, but it doesn't seem like it sends the channel_utilization very often. only that last entry is for today.
If you'd like to look at other nodes, I could send you my whole debug log.
@leaden crow I couldn't watch things too closely last night while we were testing, I was busy with family stuff. Did your connection to NPR start working with any level of consistency?
The traces were inconsistent during testing, but it seems like sending / receiving messages was reliable. I also think the rpi node has something weird going on where either it doesn't receive messages while a trace is running, or they don't make it to the web UI. I'm talking specifically about a couple of messages received by my roof node, but not the rpi. The other possibility is the messages just hit their hop limit at my roof node and didn't make it one more hop.
I wish they could make the messages persistent in the webUI. I wouldn't even mind if they stored them in cookies or something. If I have time tonight, I will look a little closer at the debug logs. I should be able to tell if the packet is logged, but didn't make it to the web UI.
I ordered E22-900M30S modules today. I'll need to build a hat, but if they work out, I'll probably replace my solar/RAK roof node with an RPI and one of these 30 dBm modules. Having debug log at my roof node would be nice.
So the node on the hackerspace is a thing now. Youll probably see it after thursday. I got her set to channel 51.
poe worked flawlessly
got upgrade scripts
restore config scripts
all the creature comforts
so it should be a solid node..
This is from running rtl_power all day today
https://temp.ӂ.net/d3-waterfall/example/
warning, takes a long time to load the data file is 245M
I think I will turn off all my meshtastic radios and run it on a narrower band overnight
Additional ignored node options, they might rework old ignored node:
———————————————————————
Could be related to the old config loss issue?
What are some good values for CLIENT_MUTE on a rak4361? What are the caveats. Is Bluetooth usually off or is it wake on packet?
I read some of the meshtastic page and docs but its generic no suggested values for devices and roles
Trying to configure the one in my car not to suck juice all day and hopefully get the solar panel to feed it well enough. Any ideas?
This is the device
I doubt you even need to change the defaults. I would just run it for a while and see if that solar panel is not maintaining battery level.
Is there anybody still on MF? I connected to my MF node and the node list is empty. Worried that there is an issue with it
Got it. Yeah I'll try that. Benchmark it. Then turn power saving on .. Benchmark it and see if its worth it. Its sitting in my car window facing the sun all day so we will see how well it does
I expect the GPS is going to suck juice. I really need to find an acceptable trade off with GPS updates
Maybe use smart location if the node moved.. Update. If not.. Sleep? Is that achievable?
If anyone still is, the main routers are not.
Main routers as I understand remain on LF 51? Right
@white field tried sleep with the RAK GPS module. I think it is not possible.
If the node is going to live in your car, you might consider hard wiring into your car? I think it would be trivial to run off a kit like this: https://www.amazon.com/Upgraded-Hardwire-12V-24V-Protection-Detectors/dp/B0BZYPW2JW?th=1
I could. Its a good bit of current draw though when GPS turns on and the device fully wakes. Well beyond the cars parasitic draw so it would have to be able to survive the heat, cold and have battery to spare to survive when the car is off.
I def can wire it up though. I got my own higher quality 12-48 VDC to USB C power supplies ive built before so I could use one of those.
I agree that a higher quality buck converter would be nice. I would make sure you implement some kind of low voltage protection (cheap hardware kit has that going for it), or at least carry a noco boost or similar.
I'm going to plan to migrate some of my nodes to LF/51 this weekend.
I'm kinda hopeful LF will be able to reach NPR with one of them when finish. I'm kinda paranoid that the batteries might be dying for onaqui since they are standard lipos. 😂
Skyline might just be a PITA to get LOS without driving 2 hours. It was intended to rely on a mountaintop router to get signal from our area.
We’re using a T114 device and developing a custom app that collects data from the receiver via Bluetooth, then sends it to an MQTT server. Currently, I’m using an MQTT proxy, but it's not functioning as expected. Are there alternative methods I could try?
We’d need to know more about what your setup is.
I think I will take my SDR to work tomorrow and see what it looks like from there.
901 - 902 MHz seems like the least noisy to me
sssooooo I lost one of my nodes and it's the one with my admin creds. I have backups of the private keys, but I think I'd have to fork the firmware to spoof my node ID to properly impersonate it from a new node.
Has anyone seen the last location of 9c33 mobile:1 (🐋 emoji icon on node list) it had a builtin GPS and it's last known location per any of my other nodes was timestamped 2024-11-06 10:41 AM.
I suspect I left it at a coffee shop somewhere between Sugar House and Downtown SLC. I'm trying to see if anyone else got more recent locations from it to hint as to where I left it. (it was running LF/51 when I lost it)
Battery was probably dead by 2024-11-08.
Last I have
As far as I know, you can just paste the keys in the new radio and it should work
Oh maybe it’s the other way that doesn’t work. You can’t change keys on a node known to the receiver or it rejects messages.
Either case I’ll bring RAK it’s never seen.
Thanks, same coordinate I was seeing. Have a timestamp?
I think using my other nodes timestamps I’ve determine that it’s definitely not at my house. I have a timestamp from 11:38 from a pocket node.
All my home nodes have a timestamp from 4 hrs earlier. Implying it was out of the house when that last event was RXed.
From google maps history I think I have determined where I was. Now to go visit that coffee shop. 🎉
2024-11-05 02:38:49
If you leave your new node on for long enough with the correct keys, I believe the remote nodes will eventually grab the new key. I could be mistaken though
I have the same last position, but at a slightly later time:
{"channel":0,"from":468491315,"hop_start":3,"hops_away":3,"id":893012462,"payload":{"altitude":1296,"latitude_i":407896064,"longitude_i":-1118830592,"precision_bits":11,"sats_in_view":12,"time":1730896855,"timestamp":1730860007},"rssi":-54,"sender":"!eb4c89df","snr":5.5,"timestamp":1730896863,"to":3947661791,"type":"position"}
Your time zone: Tuesday, November 5, 2024 7:26:47 PM GMT-07:00
damn, close call!
I broke my hand mountain biking a couple of years ago and almost had to cut my titanium wedding ring off. Switched to a cheap silicone band after that.
Also, this went into effect 2.5.13.1
Lost my titanium ring this fall, moved to a 3d printed TPU ring, so far I'm loving it.
TPU is tough as hell
One of these days I'll buy another 3d printer. Both filament and machines seem miles better than the one I had 5 years ago. Currently below a milling machine on my want list, so it might be a while.
Lol! Sorry about that
No need to apologize. We've all done more embarrassing shit than that before.
This is rtl_power listening for 5 hours at my work in NSL. 900MHz is definitely more busy at this location.
https://temp.ӂ.net/rtl_power_work.png
Couple of notes; I turned my LF51 node at work off for only a couple of hours. You can pretty clearly see at 914.6 MHz when it was on/off.
901 - 902 MHz seems to be the most quiet again.
That IC looks like 54160
https://www.ti.com/lit/ds/symlink/tps54160.pdf
It looks like just an iteration of this board:
https://avdweb.nl/solar-bike/electronics/ku63-motor-controller#6vlightpowerboard
Probably with different resistor divider for 5V on your board
@real cedar Is this the new roof node at 801 labs?
3 │ Meshtastic 53bf │ !67e012a7 │ 53bf │ RAK4631 │ 82M5qMVVYF+In2y1K0h/vmQQIJ74sJLVQt0XeJQ5/AA= │ 40.7653° │ -111.8808° │ 1327 m │ 101% │ 0.98% │ 0.62% │ 7.00 dB │ 3 │ 0 │ 2024-11-18 23:26:12 │ 5 mins ago │
Yeah but I broke it 😂
I managed to some how send is managed true at somepoint
Gotta nuke and pave the firmware but we had to close the space haha
nice! at least it's working
Yeah its got a view on POTM and NPR so should only be 1 hop. You'll be happy to know I talked with the people interested in building nodes and drove home router vs repeater vs client vs client mute.. The why.. the when. The how and what not to do
So should have a lot of client mutes showing up at some point but I told them this would be the mostly planned out channel and that they are all in this together. Dont make it not work for everyone haha
One of my friends bought 3x T1000-E today, and I'm guessing he'll have a roof node shortly.
Its on a 12 ft mast above the building. It should be a solid client node. It will go down and up as we tweak things but anybody know how to run a web interface that shows live chat feed of rooms and the map? We would love to display it at the space
What are T1000-Es?
Edit: actual product link
https://www.seeedstudio.com/SenseCAP-Card-Tracker-T1000-E-for-Meshtastic-p-5913.html
You could ask @main stag what is required to run a Grafana dashboard like his (pinned).
That might be a lot for the space to maintain bit if hes got an easy way I'm down
It looks like you have at least a couple of options:
https://github.com/cordelster/mesh-metrics
https://github.com/tcivie/meshtastic-metrics-exporter
In either case you'd need some machine with postgres, prometheus. It'd probably be ideal if you build an RPi node. I'm waiting on parts for new hats, but I could probably donate my old waveshare hat. It has the known problem of sometimes not being able to send long messages.
Yeah I haven't open sourced mine at this point with some of the setup. That 2nd option dragonslayer linked looks like a fancier version of mine.
Almost all nodes I see claim to be 1 or 2 hops away, but I can successfully traceroute to only a few - both clients and client_mutes. Any idea why this happens?
This may not be your problem, but I have been noticing this for a while; I can reliably traceroute some nodes almost all the time. For example POTM and Mill Mesh. My outgoing trace usually hops through NPR, but return almost never does. Most of my NPR traceroutes are not returning. Although sometimes NPR traceroutes are very quick, and reliable for a short period of time. A lot of nodes that I reliably get telemetry from, if they are one hop (through NPR), I am not getting traces back ever.
Maybe I will script some traceroutes on an interval and see if I notice any patterns.
I see the same behavior from our routers. It is as if they put themselves to sleep for a short period after passing a message and are unavailable for the response.
Maybe this doesn't affect typical messaging, but I would think that confirmation of message delivery is similar to traceroute behavior.
I know they came out with the new t114 v2, but did they ever fix the T114?
I want to build one today, but not if it's a paperweight.
They will exchange them for free, from what I read
I just ordered the V2, just sucks that its no good
I made a thing and it works!
Nice!
Now the hard part. Obviously not going to try and run this off solar. I'll need to run a circuit into the attic space in my garage, run the antenna cable through the exterior wall..
what is that
Most of the transceiver modules in Meshtastic devices are regular Semtech SX1262 that broadcast at 22dBm. This eByte module has an integrated amplifier, it broadcasts at 30dBm (1 watt).
Well shit I used the wrong unit at the space
hahaha
This guy sells a hat with the module, but it has been out of stock for a while. I just bought some modules on Aliexpress ($6) and put it on a proto hat.
https://www.etsy.com/shop/frequencylabs/?etsrc=sdt
That's why I was telling you to just get a hat for the RPi, and ditch the RAK.
or the module itself
yep, that's the one
Pi Hat?
he posts in the #meshtasticd channel (the guy who made it)
I don't think anytime soon
I could see 2.4 increasing data rates but also I guess as a backup to if 915 gets gobbled up
I think you know more about radio than me, but my simple understanding is that 900MHz has more range than 2.4 GHz
Im aware of that though Im willing to make the trade off for short range stuff
There is also an audio PTT implementation that has yet to be really tested or used. 2.4 GHz will provide the bandwidth to make that usable
I'm excited for the day where that is functional
This is off topic, but any hams in the Northern Davis/Weber county area interested in trying Lora APRS on 433.775Mhz? I'm going to leave an iGate up for a while. C'mon, we've all accidently ordered a radio for Meshtastic that turned out to be 433mhz and don't know what to do with it...
There's also vastly less noise on 900 MHz compared to 2.4 GHz
@primal topaz
There isn’t much on Amazon, I’ll have to slow boat
Alright first or second week of December I’ll have two nodes on hand.
I'm sure somebody is doing it better than me, but I don't really like those grid plates. I usually make my own from a sheet of acrylic or polycarbonate. Use the grid as a template for the box mounting screw holes, then mount all of my components to the plastic sheet with standoffs.
No, the boxes I have bought have a removable grid. It looks like maybe yours does not?
This is probably a better picture. I drill and tap the polycarbonate sheet and screw the standoffs directly into it. I think you could still do this and use the mounting holes around the perimeter of the grid. You might have to add some spacers?
I have used small zip ties through the grid. That seems almost impossible to do if you can't remove the grid and get to the back of it.
I like this idea if you can get the zip ties threaded through the back of the grid:
https://www.reddit.com/r/WLED/comments/183l1j8/tips_for_mounting_components_in_work_box/kawhpn1/
These look useful: https://www.amazon.com/dp/B0BRQ8T4RY?ref=ppx_pop_mob_ap_share
You could screw into a grid opening with a short screw to tap threads
Product information: - Product Name: PCB Board Mounting Feet with Screws - Material: Plastic, Metal - Color: White - Size: As Shown Package Includes: 60 x Motherboard Fixed Plastic Feet, with Fixing Screws Features: - The motherboard standoffs and screws are made of high-quality plastic and metal...
😻
Channel 51?
Finally?
Hell yes
Saves me money I'll go do black mtn or Ben Lomond at some point. I'd be careful about Francis though. There is a monitored radar station there
If you want a better cold weather 18650
https://liionwholesale.com/products/cold-weather-protected-3500mah-10a-18650-button-top-battery-wholesale-discount?variant=40574021992517
I find this to do super well down to about -35 C
The LiionWholesale Cold Weather Protected 3500mAh is a PCB protected button top battery. This is a 3500 mAh 10A battery that can be discharged all the way down to -40°C and is one of the best batteries we offer for high discharge flashlights or other devices that need a high capacity but medium discharge protected bat
I can get my snowshoes backcounyry shovel and skis out if you want someone else to make the trek 😂. But yeah I get ya. I'm trying to avoid moving to medium fast. I desire longer distance but medium fast would make sense in a city with a lot of nodes. I really want to be able to reach SLC with my nodes. Right now they are beyond useless past North Salt Lake
Oooof I feel that. Sucks when you buy something and can't use it .. I definitely have a license but I quickly got disinterested in HAM due to lack of encryption. I dont have much interest in clear text comms these days. I was pretty avid for a bit but quickly got bored. (Probably just a young guys perspective)
I ran a traceroute every 15 minutes to NPR for 24 hours, then did the same for POTM.
I got a 25% success rate for NPR
47% success for POTM
Curious if anyone else has done this?
Q about the default power setting ..... I have a RAK and a LilyGo, and I flashed them with 2.5.11 a couple weeks ago. I connect to either of them via iPhone. Today I discovered that both of them have the power set to 25 dBm. I'm pretty sure it used to be 30 dBm, and I certainly did not make any changes wrt power. Any ideas ?
Their max power is 22dBm (ish) anything above that won't matter
Both of them ?
I meant to post in 'firmware', how did I end up in Utah 😬
Well, thanks, it seems they both do 20-22.
😂. I went up there today in the storm as long as you get up there in the next 4 weeks you should be able to get most of the way up and trek the rest. But yeah if come January you havent I might haha. Depends on the conditions.
I got NPR about 15-20% of the time at the university with direct sight so yeah I have a lot of issues with NPR
All the others I get clear signal nearly 100% of the time in town and successful trace routes
Yeah also you probably know this but higher gain will do you better than more power. The more efficient you can send and receive the better your sensitivity is. The FCC uses ERP in an effort to include the antenna and its gain factor in the equation. Your transmitter usually compensates for an antenna to an extent but its your responsibility to stay in the bounds of the band operational limits. In the past the laws simply stated Peak Envelope Power (PEP), which is 1W for the 902-928Mhz ISM band. 1W is the base rule for the band. Let's do EIRP for my setup st the space. EIRP is calculated using Ptx+Gant-Lreturn. Given a transmitter power of 24 dBm and antenna gain of 6 dBi, we account for the VSWR mismatch loss due to a VSWR of 1.1. This gives a minimal return loss of about 26.4 dB, but since the mismatch loss is negligible for such a small VSWR, we approximate it as zero for simplicity. So, the EIRP is relatively close 30 dBm.
This isnt super accurate but I cant do all the math in discord. If I transmitted at a higher power I.e above 24 dB and I had a 6dB antenna I'd be violating FCC regs (in theory) so power isnt everything. You need less than you think since you can usually achieve more range with a better ability to listen. I dont know your situation but I suggest higher gain antennas and lower power to compensate for lack of being able to Chuck out 30 dBm on a omni directional (and probably crappy) low gain antenna. So that's something yo keep in mind. Ive been able to achieve some great stuff with lily go and raks with higher gain and lower power.
I ran traceroutes overnight to DC801C6 and finally got a few back.
Route traced towards destination:
!359efdbb --> !ac7f43c3 (?dB) --> !0e81d07d (?dB) --> !77272d1f (?dB) --> !452b53bf (-8.0dB)
Route traced back to us:
!452b53bf --> !33676da0 (-14.25dB) --> Unknown (?dB) --> !ac7f43c3 (0.5dB) --> !359efdbb (3.25dB)
The routing is so strange to me. Outbound hopped from NPR > POTM
Inbound went through Mill Mesh to that elusive Unknown
Yeah POTM should be the main LOS router to it. However that's going to and soon. There building a high rise soon
So its mostly just gonna server as a client in a dead zone of the city
That's too bad. I wonder if we could get a node near Ensign peak that would have LOS to downtown and davis county? I ride Bobsled a lot in the spring. Maybe I'll try to explore above BST for good locations.
Snow report for you at Thurston from me up there today
Base hard pack 19 inches under stormfall
Base hard packs about 20 inches thick
Very trekable still
Temps ~22 F
No not -
~
Around
So around +22 F
But yes get some of those lions I showed you
The cold weather 18650s they have are freaking amazing. Honestly if electric cars up north use a 18650 cell they should use that one hahah
Not feeling like making the trip yourself eh 😂?
I'm headed out of town for thanksgiving. I'd say make your design good. Test it over the winter on your balcony to make sure. the last thing you and I want is to make the trek only to need to return. How and where do you plan to mount it?
Howdy Meshers. I'm finally getting some good results. I'm in Clinton.
Love the old Amiga emblem. 💪
I suggest a tree. Tbch Francis peak is highly trafficked. Tons of teenagers too
It may be best to be off the beaten track on this peak
Channel 20 I'm assuming?
Most of us are on 51 with no 51 up north
Yet at least
Correct. I have since switched to 51.
Yeah I can't get much up in the Roy area of 51. I can't see NPR. You'd think with direct LOS I could but never have I ever been able to get it this far north
Has Utah ham reached back on anyone's efforts to put something up at Farnsworth?
I'm seeing NPR.
Are you in Roy or south Ogden or clear field cuz I put 50 ft out with monocular vision of it and still nothing and its the only one that doesnt show
I'm in Clinton. Node info is incomplete.
Yeah debug shows me theres just not enough signal strength and I get a lot of corrupted bad packets from it
I wish I knew why. Its kind of weird and it happen on multiple professionally designed and built stations. Pro grade not consumer
So I'm sitting here thinking.. Why is this the case. 😂
I just dont get it. Ive tried a variety of things. I even measured lobes in an anechoic chamber and literally aligned the lobes. Ive done filters. Ive done all kinds of stuff. Right south of the base on the hill side that is still part of clear field. I get nothing. No buildings in the way etc
Its kind of frustrating.
I feel your frustration. What is your setup? I'm using the RAK 4631 in the RAK solar enclosure with a 5.8 dBi N-Male Omni Outdoor Helium 915 MHz Antenna mounted on the roof about 30' up
I assume you are !15e09ec4? You popped up in my node list, and on slcmesh. You must be doing something right to hit NPR. Can you get any traceroutes back from it?
RAK4361 + 8 dBI + lightning arrestor .. 45 ft up
Its an N male omni as well btw
Connectors and stuff and lightning arrestor are all pasternack. Antenna I'd an L-COM tuned for 902M to 928M
Vswr of the entire feedline is 1.1
That's my mobile node, a LilyGo T-Echo currently charging in my basement 🤔
The base unit is !88ee0e65. I'll see what traceroute data I can extract from NPR and report back.
I would love to see us try to get something like a Station G2 node up there linked to an RPi with internet so we can run diagnostics and do remote updates and tests.
How far away are you from NPR?
Just use the new firmware and do OTA?
We are in the Roy area by hill afb
Oh, I mean firmware version updates. Not just config updates.
oh yeah, its only ~40mi. I've done links that were 75+. to be fair, my testing was in areas with zero interference.
If you want to range test your equipment. You might also be able to get direct LOS/signal to onaqui router on MF/45. It's 110km (68mi) from Roy.
If you can get signal from onaqui it will rule out antenna/distance/RSSI as the issue. (not seeing probably just indicates LOS issues)u
I think it's configured to broadcast node info every 1hr.
Once I get over this cold, I'm planning to swap it over to LF/51.
line indicates roughly the path of the LOS for onaqui to Roy.
37 miles.
Last I checked the raks can OTA but I mean you can definitely try. I suspect the pi might suck a lot of juice.. I really dont like raspberry pis. A whole lot of extra power I dont need to run a node. Which Adds to size and cost cuz of solar and battery but if it works out id love to see it
Might be antelope island that's messing with stuff. The elevation drops a lot for every few hundred meters west. But it opens the possibility. Hiding a transmitter on antelope island Haha. That could be another way to ensure we can reach NPR haha
Way less hiking and snow too
Catch is. Its a state park
I'd hope we are not placing nodes on antelope unless there's an existing radio station abandoned up there. Too much foot traffic who might take photos and post online with a ⁉️ . 😬
pro tip: bring a few different size clamps, and be prepared to shim with a stick if it's still too small.
@real cedar I'm gonna send you some sequentially numbered test messages. They should all be coming thru NPR.
I'll include a sequential number and a timestamp.
These shouldn't be deprioed like traceroutes might be.
Is your node !452b53bf?
setup:
- yagi pointing at NPR with a Station G2 in client mode (rdge d968)
- T1000e in client mute connected to my laptop with meshtastic CLI (lyre 73d8)
- hop limit 5 (I saw an example of it taking 4 hops to reach back with the ACK so, I'm going to give it +1 for buffer)
The test is gonna be
- send this message:
"ping test - instance: ${count} timestamp: ${timestamp}"- DM NPR to check if the message gets ACKed
- DM DC801C6 to see if message gets ACKed
- wait 30-60s
- repeat from count 0 until count == 19 (20x)
Results:
You should end up with a list of messages counting up from 0-19. (which should give a sense of reliability to RX from NPR)
okay, test complete. I ended discovering that the CLI doesn't support destination ACKs. @real cedar I ended up sending you 21 test messages. 0-20 via NPR.
Let me know if you want it routed to a different destination node and I can re-run it.
Theres an issue with that I'm out of town for the weekend
My main nodes and car nodes are with me in Idaho rn 😂
But when I get back we should try again
Sounds good
fuuuu, my onaqui router is stuck on MF/45 until I can fork the firmware to spoof my old macid and nodeId.
Just tested with a second fresh device with old admin node's keys.
CLI is showing that I'm getting NAKs. T_T
views out here are pretty tho. No regrets
I have finallllly figured out my problem. Myself or OCR typoed the private key and the firmware doesn't validate that the public/privates match.
That took entirely too many attempts to figure out. I eventually ended up writing a python script to brute force the correct private key from a "mostly" correct one.
I'm going to be updating onaqui and skyline to LF/51. (both are going to stay as routers even tho I think they would still practically act as routers even with client configs)
Skyline is still alive 🎉. It’s now on LF 51
I think it’s also meshing back to SLC. I’m seeing traffic come thru. 🤩
Do you have location on on your nodes Jacob?
I got my roof node swapped out today from the solar/rak to rpi with the e22 module. So far traceroutes seem more reliable. I cleared my node db, so still waiting for it to finish populating.
Wooh, MF to LF fixed onaqui. I can now ping my house from Faust
Skyline seems incredibly flaky, i really want to reposition or supersede onaqui so that something is between POTM/NPR and it.
I think NPR and skyline can connect, but it’s a 100mi span so it barely works.
They are in the private spreadsheet
Temperature reports inc from 11,000 ft. 🤣
It's this sensor?
https://store.rakwireless.com/products/rak1901-shtc3-temperature-humidity-sensor
I think you're gonna have battery problems below -40C 😁
On channel 51, adding routers is meant to be a group decision. Do you know that these nodes will not help your connectivity as CLIENTS?
What is this private spreadsheet again?
Help me understand why hiding a node's position is important? Positions can be fuzzified to an arbitrary degree. Already, it's somewhat annoying that we have no idea where repeater ffff is. It's currently playing a significant role in our mesh, at least for my location, MM.
Both of them are true summit mountain nodes with good area coverage.
Even if I configure them as clients, they can only typically see other similarly placed mountain nodes 99% of the time. (Unless someone is in the backcountry and needs the signals bounced in/out) Therefore, they would be functionally working as routers since they will always repeat what they hear
The spreadsheet is just a personal list I’ve made to pair with my coverage simulation for the mesh routers. (Locations have frequently not been public and I have the precise coordinates for all the routers in both)
DM sent with details
It unfortunately seems like my nodes are also pushing the limits of what’s possible for range. At 30-40 miles.
onaqui is only ACKing signals a fraction of the time.
This is a similar problem to the issues we had with NPR - POTM. (I have a yagi pointing at NPR with a rock solid connection and those messages are spaced out so throttling isn’t a problem )
With the E22 module, I had 93% success rate on traces to NPR overnight. My previous test was 25% success rate.
I got traceroutes back from onaqui
Route traced towards destination:
!359efdbb --> !0e81d07d (?dB) --> !d36b38fb (3.0dB)
Route traced back to us:
!d36b38fb --> Unknown (?dB) --> !359efdbb (-14.75dB)
The problem with LF20 was that multiple routers used up all message hops talking to each other. I just think that we should experiment with client nodes first, and then make a case for changing them to routers. I suspect that changing POTM to a client would be functionally nearly the same and might help weird behavior we see from NPR, but I could be wrong. I understand that remote nodes are vulnerable and in some cases illegal, but giving them a 15 mile radius at least can tell us if a node is north, south, east, west, nearby, or 50 miles away.
Some people may not know this, but you can set a static position from the CLI. I intent to do this placing nodes next year, manually fuzzing the exact location by a ~mile.
Ex:
meshtastic --set position.fixed_position true --setlat 40.676965 --setlon -111.817369 --setalt 1400
Yup need to do this.
Yeah I can configure mine as clients to avoid timing collisions
They are still always going to a reply everything they see. They can barely link to our best routers, but intrinsically bc they are about as far as possible from the rest of the mesh
Understand how to configure channels on your Meshtastic device, including private channels, encryption, MQTT, and location precision.
meshtastic --ch-set module_settings.position_precision 14 --ch-index 0
That would set the precision to just under a mile.
You know I really want to make a meshtastic weather station now but I want to be upscale about it. Like I want it to do a lot more than the basics.
I want to have a thermometer, barometer, hygrometer, anemometer, rain gauge, pyranometer, pyrgeometer, UV sensor, wind vane, visibility sensor, ceilometer, lightning detector and locator array (using delay lines to make it smaller), soil moisture sensor, soil temperature sensor, leaf wetness sensor, snow depth sensor, air quality sensor, CO2 sensor, ozone sensor, particulate matter sensor, solar radiation sensor, infrared temperature sensor, ultrasonic wind sensor, frost point detector, evapotranspiration sensor, cloud cover detector, dew point sensor, GPS module.
😂 how to do this.. Idk yet
Missing background radiation. 😁
https://radmon.org/index.php
They just added a geiger counter to the firmware support
Its more of a nice to have on the mesh part.. Lora would do fine too
I just really want to get that project restarted.. The lightning detector and ranging part is actually going to required some skilled crafting. see typically to rf locate lightning you need a large array but theres a lot you can do to make something electromagnetically larger but physically smaller
As in delay lines to make the particular antenna in an array to seem farther away lol
Be able to make it compact. I'd love that and then post its output and a map on my personal page lol. It would be an amazing thing to have.
I'm so sad I installed onaqui before I had temperature modules.
skyline (3805 solar:4 ) is apparently too far for us to get mesh traffic from it's weather data. (I set it to broadcast temps every 30min)
do we know if Lake Mountain is on 51? That is the only way that I’ll be able to reach out into the valley if that one changes. That’s kind of my lifeline :-). If that one switches, I will switch mine in Eagle Mountain in a heartbeat.
I don't think there is a 51 node in lake mountain
Jacob has one up there at one point but idk what happened to it
I think that was @Toshi that had a node on Lake Mountain. It's probably still on MF.
I don't think his node on Lake Mountain worked very well. @dim bay Lake Mountain node works really well, but it's on LF20.
My node in Lehi gets intermittent connection through the NPR node. Depending on your location it might bridge you occasionally😅
Your eagle mountain nodes might be able to see onaqui, but again it’s kinda flaky in its ability to link to NPR. (20% or lower acks)
I remember seeing traffic from them when I was hiking up.
I was thinking about trying a node near Bismarck Peak and then move onaqui to Deseret.
Bismarck would be better candidate to get skyline weather.
I really wanna make the links reliable out towards Moab. 🤣
I get super fast traceroutes to you through POTM
Route traced towards destination:
!359efdbb --> !77272d1f (?dB) --> !67e2f2c2 (-1.25dB)
Route traced back to us:
!67e2f2c2 --> Unknown (?dB) --> !359efdbb (-4.25dB)
Is that to jfi0 or jfi1? 0 is Lehi (home) and 1 is work in Bluffdale.
Looks like that’s the Bluffdale one!
jfi1
Yeah I could probably use binoculars and see potm from my node at work easily.
Some observations from the remote nodes I've been setting up:
I can get really long links from remote node to remote node (presumably bc there is very low noise environments) - I've had very reliable 75 mile symmetric links when I was in the backcontry.
While configuring skyline, I noticed that I was able to RX traffic from nodes in the city, but absolutely no traffic was being ACK'ed by the city. Onaqui is 1/3 the distance of skyline and currently still has very flaky ACKs to NPR. (a fraction of what I've been able to acheive with other nodes) I think what I'm seeing here is that routers within the area of SLC are in high noise areas and cannot RX very weak signals. However, the remote routers I'm trying to link to them are find RXing because they are in low noise areas.
Drawing some conclusions from this, any node wanting to be the "initial hop" away from a city must be very close to that city router (10-15 miles max, so the city node [POTM | NPR] can hear it) After escaping the noise of the city, the gap between nodes can probably be much farther. (3-5x the distance)
I'm going to try running some range tests for this over the weekend on Sunday to try to get a better intuition about this.
As a point of reference, what's the farthest anyone is reliable able to link with NPR or POTM? (near 100% ACK rate from DMs sent to those nodes)
(including antenna + hardware + distance would be helpful)
One counter to that; I think at one point, Francis Peak to Lake Mountain had a pretty solid connection. That was over 50 miles across the whole Wasatch Front (noisy). Also, that was with pretty nothing-special hardware: #1197577977781821541 message
#1197577977781821541 message
That sucks. Oh well I guess.
Kinda wish people were more cooperative at radio sites. Often the people that run them can be quite.. Insular at times. For some reasons that are understandable and others merely territorial
I'd love to see Francis or Thurston get a node that can connect to the core of the city. I know there are some planning it. Its warmed up .. Can still traverse up there.
The last few times utahham commented, he was unfortunately super busy. If you look through his history on this discord, he potentially has access to a bunch of the site the hams use. Earlier this year planning or scheming for places to deploy nodes.
Yeah. Ive just ran out of budget to spend on mountain top nodes for now :/ I got more friends into it. Showed them the ropes. Theory of operation. The modes etc. One of them apparently is planning one near Rudy's flat on the ridge high point or bountiful peak. That alone would be a huge help for me
Nice! I have hardware to place a couple next year. I'm not hopeful they'll be super useful until I get more than 1, or maybe do some testing and move them around.
If you want me to review hardware and settings I'm down. If your part of 801 labs and want me to stop by while you have it on hand.. That could probably be arranged too
Like what would be super ideal is a trio of routers. Bountiful/Rudy's... Francis/Thurston... Ben Lomond.. Would extend the coverage to tremonton easily
I went to a meshtastic meetup th3ph3d organized a couple of months ago, and joined their discord, but not a part otherwise. It's a cool spot. I will definitely stop by there again at some point. I have too many family and work obligations to drop by regularly at this point.
I taught it along side limitlessmatrix the other week. I wasn't prepared but its a solid easy place for me to meet at so I figure if you wanted to run through your hardware before you send it up. I could do it in person there
I may offer to teach more there and be able to open and close. Then we would need not worry if its open or when a class is. But that's still up in the air. I need the time first
I was chatting with a friend who might be able to a roof node in downtown SLC setup for LF51. (Probably 10 stories up)
Anybody know who Dev420 is? That node does not have location info, but I think it might be really close to me. I have direct connection and reportedly high SNR:
!359efdbb --> !e4ba990d (6.0dB)
Yeah, I couldn't make that. I'll try to watch the other discord and go if I can make it next time.
😂 yes I know who that is
Friend of mine. Different forum. Different nick. Withholding for privacy
She constructed a solid node we were chatting about it.
Wait 6 db
OMFG LOL
well now I know where you reside
😂😭
My roof node is dsr1, the exact location it reports is the Skate Park at Holladay City Hall. I'm actually East of that a bit.
Yeah the reason I know is I have a solid idea of where dev420 actually is. 😂 so yeah that was a funny unintentional dox of yourself but dont worry it will be out of mind the moment I see my girlfriend or something
That would be epic. Is it in the avenues or nearby a high spot?
hiya
Not used to seeing direct connections. Most nodes are 1 hop through NPR or POTM to me. Was curious if it was someone on here.
not too far from ya i see
😂
@leaden crow
nice, but mine is bigger
Listen guys. That's very phallic and all but get a room
🤣
Is that the 15 dBi?
New for 2022, Rokland introduces our large profile tuned 8.5 dBi 915 MHz 40" fiberglass antenna with N-male connector and U-bolt bracket mount. This is certified and designed for outdoor environments and is weatherproof. See photos for mounting instructions.Important note: this is different from our low profile 8 dBi a
🥵
Oh its a half wave 😂
Now I get why its so massive
I got the same gain much shorter with a quarter wave that's on the hack space. Would be cool to see if we could hook them up to VNAs again and see what the performance trade offs are
I don't really want to pull it off my roof and test it anymore, but I did check the tuning, SWR
Pretty standard graph. What I expected. What I was actually wanting to characterize is the donut. I have a setup where I can put an antenna on my SA or even a VNA. Move the vna at a set speed around the transmitting antenna capturing a two tone signal... In a perfect circle and get the radiation pattern relative accurately. The only issue is I need a fairly uniform square room or anechoic chamber to be accurate but my guess is the vertical beamwidth is a bit smaller than it is for the quarter wave which is very advantageous for longer range links
Maybe if I make it to 801 labs for another meetup, I'll bring it along. If it's motorcycling season at that point, it's going to be awkward to transport.
Hah yeah. No worries if you cant
WAIT
im 3 miles west of the park
I don’t really trust the reported 6 snr, but assumed you were close.
It could be a node below v2.3 forwarding traffic without hop_start, tho I doubt any of yall are doing it wrong on LF/51.
Woah, just saw this on the Bay Area Meshtastic announcements channel. I don't see any news reports about one hitting, but there was a 7.0 earthquake in the ocean about an hour ago.
I wish more people took the time and engineering to do stuff like this
https://shop.uniteng.com/product/meshtastic-mesh-device-station-edition/
The hobby modules are cool for like.. A week until you figure out all the not so nice things and are like man I should have done more research
Is that basically COTS (off the shelf)
Got a BOM? I'd love that
All P no S right?
huh sweet Now I know what to look for
Fair enough. Thanks for the offer. I might take you up on that at some point
I wonder how small we can make a mesh node. Like I kinda want one that's tiny, ip4x, concealable on a bike that can just get wired in to a bat pack directly that has weather and environmental sensors and store n forward
I might try to spin up a stupid small rak (since the schematic is open source) and try to see just how small I can make it. 😂.
Its a MTB focused in reducing weight and being cross country. its got a fizik seat. So fat chance even the thinnest of things gets to blend into the seat 😂 but I get what you mean. I just want a smaller waterproof form factor is all that I can perma attach
I would look up some of the things people have devised to hide an airtag on a bike. Obviously it's going to be larger than an airtag, but something like fitting it inside the steerer, and integrating the antenna into the steerer cap. You can look at tools like this to get an idea how much size you'd have for everything:
https://www.oneupcomponents.com/products/edc-v2-tool
My bike has this access panel for cable routing that would be perfect, but not a lot of bikes have something like that.
was gonna try updating the remote configs for onaqui tonight display location.
I'm not getting any ACKs from messages I'm attempting to send to it. Currently, wondering if the smog is negatively impacting performance.
If the dew point is within a couple of degrees of the temperature, it's fog...usually. where i I'm at on the Ogden bench, it's clear, the dew point is 26, the temp is 29. As you get lower in the valley, the temp drops to the dew point and fog develops, along with trapped smog. Moisture in the air can definitely impact 900mhz.
Thinking about it again, the link I'm dealing with is from 8000 ft to 9000 ft elevation between two peaks. Hop from my place to NPR is still solid.
Might still be some impact that isn't well understood. 🤷♂️
I got a traceroute back from onaqui (1 of 4 in the past hour)
Tbch I love the fog a lot. I ain't got no issue with it driving, being outdoors in it etc.
I dont think its messing with your signal much. Chirps can usually cut through fog pretty easily
Especially at 915M
So this gigantic 10 dbi omni antenna is working really nicely.
Way less obtrusive to have than a yagi. (even tho it's still nearly 4 ft long)
Is it the 45" Rokland?
hahah yup
Will report back with more test reports after I finally get time to get back out of town.
Nice! I almost bought it too. I did some napkin math and thought the 12 degree vertical beamwidth might not be good for me with the elevation differential from a couple of potential mountain nodes. I was probably being paranoid. Not much difference from 14 degrees.
some one forgot the negative in their gps and can see china atm
@leaden crow dsr2?
lol, I only had that powered on for a minute. I’m going to swap out the tbeam at my work in NSL with another E22. I’ll fix it before I place the node. Thanks for the heads up.
Hi from Springville. I've got a couple of lilygo devices that are populating the node list with dozens of nodes, and I've successfully sent and received one or two messages, but usually they fail to send to other nodes. I'm interested in hearing the local chatter and helping build a reliable public network on wasatch front. What should I do next? My guesses:
- join the MQTT server? I saw something about it way upthread but I don't see a way to search this post. What would be the appropriate settings?
- I have a couple of ESP32 boards. Buy lora boards and make some mode nodes to give to friends?
- I'm tempted to buy a bunch more t-echo devices and give them to all my friends in the area for them to leave plugged in or fool around with from their phones. Is there a better hardware choice? What setup should I do before handing them out?
- I bought a Linx ANT-916-CW-HW-SMA for the t-echo and that maybe helped a bit. If I wanted to buy a turnkey node to stick outside for better reception, what should I get?
- I hoped the t-deck would work out of the box like a blackberry for messaging, but looks like the keyboard does basically nothing so far. Is there some firmware I should be loading instead? Presumably I'd also need a case and a battery.
the tdeck needs firmware installed after you put it together
Yeah I've managed to update both devices to the latest stable
I would disable MQTT unless you're configuring it for some private use. The default LF on the Wasatch Front is already a crowded mess, MQTT causes a whole bunch of problems.
Unfortunately adding a bunch of devices like t-echo to the mesh isn't going to help much. They really should be configured as client-mute in a crowded mesh. See this for more details:
https://www.youtube.com/watch?v=htjwtnjQkkE
If you want a roof node, I would buy the RAK enclosure with solar built in:
https://store.rakwireless.com/products/wisblock-meshtastic-starter-kit?variant=43884034621638
https://store.rakwireless.com/products/unify-enclosure-ip67-150x100x45mm-with-pre-mounted-m8-5-pin-and-rp-sma-antenna-ip-rated-connectors?variant=42861623738566
The touch firmware has not reached the point they are releasing binaries. If you look at the pinned message on this channel:
https://discord.com/channels/867578229534359593/1137655366411370557
It has instructions for building.
I've only ever received like 3 messages on LF. How do I get to where I can read the crowded mess?
I would build the roof node. If it has line of sight to the Lake Mountain router, you should be able to send messages to the whole Wasatch Front.
Would an outdoor node also be table stakes for friends in provo etc. if I want to give them something that they can get online with a minimum of fuss? Is hitting lake mountain the basic thing I should shoot for?
It is going down a rabbit hole thought. I built and then upgraded my roof node a bunch of times over the past 2 years to get better signal. A good antenna will make a huge difference. This one is amazing for the money:
https://store.rokland.com/collections/all-helium-antennnas/products/alfa-aoa-915-5acm-5-dbi-omni-outdoor-915mhz-802-11ah-mini-antenna-for-lora-halow-application
Absolutely. The area around you would benefit from having a high quality roof node if none exist. Especially anyone that has line of sight to it. Topography plays a huge role here, if you are on a hill or bench, even better.
One thing of special note, leave your roof node as client. A huge problem with LF20 in Utah is that too many people configured routers. That Comms Channel video explains it well.
So shopping list would be that starter kit, enclosure with solar panel, and a battery if I want it to work at night? And then I assemble all that, stick it outside and connect over bluetooth with my phone? Or do I need another module with mcu and radio or something?
And then an SMA to N-F adapter if I get that antenna? Any pointers on battery?
Correct, but the battery is not optional, it's not going to run reliably off just the Solar. I would browse:
https://discord.com/channels/867578229534359593/970723761013800970
A lot of people are testing these cells because they are rated for low temp:
https://www.18650batterystore.com/products/molicel-p45b
You're going to need battery holder (or spot welder), BMS, and the optional 'Battery Connector Cable/5 pcs battery wires' with the RAK starter kit
If you buy that Alpha antenna (or one of the bigger ones on Rokland), you'll need this adapter:
https://store.rokland.com/collections/antenna-extension-cables-pigtails-connectors/products/uflipex-ipx-mini-pci-to-n-female-bulkhead-pigtail-cable-extension-rg178
And I should be thinking 18650 or larger for battery?
I would get the 21700 if you're buying somthing. If you have 18650 on hand they will be fine.
If you are comfortable soldering this together, these are the parts I used:
https://www.amazon.com/gp/product/B08MPXHFJB/
https://www.amazon.com/gp/product/B0BSC61X69/
If you just want turnkey, there are several people that put all this stuff together and sell it on Etsy. Obviously you're going to pay a premium buying it that way.
Thanks for the tips, I'll read up
If somebody wants to dip their toe in without hardware, can they just download the app and join the local LF via mqtt? What would the server be for that?
I don't think so. You can get T3S3 on Aliexpress right now for $14.48 with free shipping if you just want the cheapest thing to try it out. https://www.aliexpress.us/item/3256804440825086.html
Make sure you choose the SX1262 915MHz option if you're buying those.
Went up several hundred feet above the valley floor to the shoreline trail with the t-echo and the linx dipole for about 20-30 minutes and still saw no traffic on LF. Traceroute to lake mountain took 3 hops through spanish fork. I got good signal to a few more nodes, but only managed the one traceroute. I'm doubting that a roof mount is going to do any better. Seems like something's wrong.
What do you mean by "traffic?"
The node list can take several hours to populate. If you are expecting to see a whole bunch of people chatting on the public channel, that doesn't really happen. Sometimes it can be days between people sending test messages or whatever. You can send test messages, but there is no guarantee people are watching devices and will respond instantly. I haven't had any nodes on LF20 for months, so I don't know if it's more chatty now than in the past, but I doubt it.
3 hops to lake mountain seems odd though. I could get direct traceroutes to it from my roof node in Holladay, which is much farther away.
@slate urchin You still have your stuff on LF20 in Utah County, right? How's service lately? You see much chatter on the public channel?
The reason why people can't hit lake mountain directly is because that site has line of sight to about south provo... then part of Lake mountain blocks it from going down south. That site was first an aredn mesh site that runs on 2.4 and 5ghz, and Spanish fork was always impossible to reach. We had to setup vpn tunnels over the internet to get it to work. Not ideal....
Pretty sure this is total overkill. My router on onaqui summit is running just fine off a $5 5000 mah lipo cell from Amazon.
I think we can likely get away with cheap batteries at this latitude
(To be clear, I would use the molicel batteries for anything truly remote)
I suspect you're right, but it's nice to have the rating. Even these cells are not rated for charging below 0C. Hopefully the low charge rates with solar in the winter make that not really a big deal.
For roof nodes you can use whatever. 🤣
They sell batteries that include the BMS on Amazon which was nice
Ah I guess it was more expensive than I remembered, but drop in usable with RAK hardware.
Convenience is probably worth it, especially building several nodes. I built a couple of those E22 hats and I don't even want to look at my soldering iron for a while.
Did a kind soul finally put a node up that let's me relay to slc
I can now see y'all from my apartment
And a crap ton of new ds xyz nodes
I switched out the node in my office (North Salt Lake). It was an old SX1276 tbeam, now it's an E22.
Yup that pretty much connected me
I'm thinking of strapping the station g2 to a ugreen 20k mah bat pack and calling it a mobile node 😂
Insane power for mobile though
T beams are trully not that great. I think imma buy an extra station g2 for the apartment and put the solar node on the roof.
I'll just lift it with my drone
Wait a minute, I love my t-beam supreme!
Whoa, what’s wrong with the tbeams?
Honestly, tbeams are extremely stable compared to a lot of the newer hardware, and also have more memory for things like store and forward. The 1262 tbeams are absolute rock stars.
New hardware is just new, so far they haven’t shown to be “supreme”.
Ive had well matched antenna feeds out on them and I get very poor trace routes where I can physically see the node and I pull out the rak. No problem. 6 dB
Its very consistent too. Sitting on the 9 line in slc
And I routinely test this at the 999 ride on Thursday nights too particularly after firmware updates. They just dont give me the same consistency in trace routes and sending messages as my rak does with an identical antenna feed
I have 3 of them and they have consistently been my worst performing devices. But stable.. And have store n forward
Just as a sort of anecdotal has anyone in SLC seen consistently on Thursday the node with the long name "Aurora_Jovialis"? Cuz thats my tbeam with a solid antenna feed. Ive never been able to get it to communicate consistently
I love the variety of hardware we have today, and the nrf52’s have such superior power advantages it’s why we mostly moved to them. That said, almost all of us that have been with the project for more than two years would tell you that tbeams are the “old reliables”, and are truly exceptional radios 99% of the time.
I have no objection to you or anyone saying they are your least reliable, but know that calling “all” tbeams unreliable is ignoring the sample size here. How many times have we seen folks frustrated with raks because of config resets? The sample size of folks with rak problems is probably <2%. But the number of them we’ll see on here or Reddit complaining about said problem is very high. Makes it feel very common.
Man, this sounds like an old man rant, so I’ll stop there. I just wanted to defend the old radios, thebentern recommends the Lora v2.1-1.6’s regularly because they can be found so cheaply. And they work! They are just out of vogue.
It’s so hard to convey tone over text, sorry C6, I didn’t mean for that to feel like an attack
Lmao bro I dont assume malice behind someones messages
Not unless they directly make that intent clear so no worries. Fair enough thanks for the perspective. Its probably a bit of new hardware making the old seem meh. I definitely will day. Tbeams are way better than any heltec I ever bought. In fact I just generally regret getting a heltec. I dont use it.
Oh my god the config reset happened to me on my car node and I was livid cuz servicing it is a pain in the box its in
😂
You have done a lot of work, and bring expertise to the project, and I appreciate it. We all benefit from it. ❤️
Working on more ATM. Who wants to see if we can get kennecot to be chill with a node on their smelter tower 😂 jk I'm not screwing with them but I do have plans for Farnsworth peak. I just found out someone at the space has ties to Utah HAM so I might ask him what the status is of getting something up there. My other friend is planning a bountiful peak router but she won't be able to get it up until after winter is my guess
I have another friend who lives at the mouth of BCC and we are talking about trying to plan a link through the canyon but it will nearly certainly require multiple hikes and me getting my ropes and carabineers out
I did a lot of this mostly to connect me and my friends. I'll probably leave it all in place if I move from Utah (non zero chance of going to space X after my masters in RF eng). I might just give them all the remote management keys and stuff so they can make edits OTA
I agree. The original T-beam and T3S3 I bought 2 years ago have been rock solid. Unfortunately they're both 20dBm SX1276. SX1262 is a huge improvement.
I had the hilarious thought that I'd have the ear of the company and might be able to smooooze my way into launching something but i gotta make my way up those ranks first
Question for the group. Does anyone have 21700 batteries they would like to sale?
I think our best hope is a SpinLaunch CubeSat. We need to convince them it would be a good test system once they reach orbit. 😁
#1197577977781821541 message
also
#1197577977781821541 message
Thank you!
4 if that's cool.
I'll DM you.
Can you reach my E22 node in NSL on LF51 now? C6 can, but I don't think he's as far north as you.
I ran traceroutes overnight from NSL to Holladay and 93% were successful (hops through NPR). Kind of interested to see how far north it can reach. So far I am impressed with the E22.
Agreed, my roof node is the 1262 tbeam, and it’s great.
@white field You might have something to point your Yaggi at. It looks like it's drifting East at a high rate of speed.
#general message
Nevermind, it landed already
I'm unable to easily send messages to someone in south Jordan but I think that will clear up once I have my station g2 and absolutely mad lad blast 30 dBm out
Might even finally hear NPR with zero hops
I'm next to Hill AFB
So I'm rather north
If hes in Ogden he should be able to use me as a relay
@maiden wraith Thurston and Francis are still trekable without snowshoes +latest report
maybe we can talk when my g2 arrives
I hear there is currently a lot of black ice from trail runners. Bring spikes
Although I didn’t get a picture, @maiden wraith is a badass! Thank you for the help today.
Nice how did the test go?
Im strapping mine to a ugreen 20k mah usb c pack. It will be in my backup with a solar panel that can charge the big pack
gonna put a 6 dBi flexi on it too so its easy to stash in a backpack and will have it on me most times. It will probably become the main node for me. Will put a rak out as my balcony deal. I dont feel like needing a lot of power there especially if my main is CLIENT mode
I am gonna use mine for the portable node at festivals on 20’ tailgating flagpole
Set up on balcony while not out and about
that is actually one of the better times to try. You want to know how it operates in unideal conditions
I gotta get out to one of those
I wanna see this madness
haha
Its gonna be quite the trek for me. I'll have to park my Ford focus on the lower ridge turn off before the split then walk the rest of the way. If you give me exact GPS coordinates we can talk about doing so if this storm doesnt absolutely bury stuff. Also how do you want it mounted and anchored to the ground to deal with the snow levels you would expect up there? Typically theres about a 80 inch base during winter so keep that in mind. If you've got a solid plan and I can get it into the ground with a back country shovel then I'm down but if we go up there with next to nothing it might be a fruitless trek 😂
What app / program is creating this "device metrics log"?
I would expect this with the number of users up North. When we switched to MF, I could see some people and get some messages, but traceroutes never returned. When NPR went up, everything improved, but traceroutes were still spotty at best. It improved again a bit when we moved to LF51, but traceroutes were only maybe 25% successful. Then I just kept upgrading my shit until they are mostly successful now.
Both android and iOS apps do it
😂
How badly does MF affect range of nodes and mountain top routers? Like its definitely going to be something to take into account
It depends. Some of us had no issues at all, others were very affected.
If you have line of sight, should be fine, even at long distances
I think it’s a -5db loss moving to MF, but I’ll have to check
Hmmm yeah I think your right on -5 dB of link budget subtraction.
That shouldn't be too bad. Once LF51 is super figured out north to south I think it wouldnt be a bad idea to move to MF51 in a coordinated manner tbch
Would help since I'm noticing the spaces network utilization often spikes to 25%
Though the cause of that spike is unknown to me rn
I really need to add remote admin keys to devices so I can just switch OTA. Its just not something ive done yet. I want things dialed on my devices before then
I think traceroutes are deprioritized, or maybe it's just the nature of a radio mesh (Acks less likely). I have sent and received regular messages when traceroutes were failing. I agree that MF did and can work. I think if we got even 1/4 of the people on LF20 to switch to MF, it would probably be the critical mass we need for the mesh to work for most people.
LF = long fast, MF=medium fast, what's the 51?
The channel. The default longfast is on 20, and medium fast I believe by default is on 45
Its the channel your on. we are on 51
Thank you for posting a screenshot. I don't see a frequency slot option? I'm running the Meshtastic android app from the google play store.
It's in your screenshot, under 'TX enabled' / 'TX power'
Unfortunately I think you're too far away to reach any of the people on LF51, but it can't hurt to try.
Duh, thanks
It's actually a pretty small (20') trailer. We usually do dry camping in places that are difficult to get larger rigs into.
Interesting, why the dual axel on a short rig?
Gah, wish I had space to park a rig like that
Tows better, has more payload. A lot of single axle trailers either have really small water tanks, or don't really have the payload to travel with full water tank. Since we almost always dry camp, I'm usually travelling with full water tank. I also just like a safety margin. I have 2x3,000lb torsion axles, so way more payload than I will use. I am now running LT tires, but another common problem with current trailers are ST (china bomb) tires that explode because of heat. Towing close to the tire/axle weight limits is more likely to cause this problem, especially combined with desert heat where we live.
@leaden crow picked this up this fall. The wife has basically never camped. Got her out a few times with our 9 year old last year in my Kodiak canvas, she fell in love with camping. She just wanted a trailer so it could feel more organized.
Got the side yard and side of the house poured right after this pic, had to wait for it to cure, so I just put it where it belongs a few days ago. It’s nice to have it all ready to go.
We hear soft sides lead to hard sides 😆
Kind of like motorcycle, leave them in the garage and they’ll multiply
Nice! We tent camped for years in a Springbar, which is esentially the same as your Kodiak. We bought the trailer in 2018 when our daughter was born. We were basically have FOMO about all our friends camping and couldn't deal with an infant in a tent.
This is a 2022, but they didn’t know what they were doing and broke things, some folks are just not curious. I want to know how everything works. I see cleaning staff empty trash cans and wonder if they have an optimized routine, I stand at a urinal and wonder about cleaning products and frequency of cleaning. They just figured they’d make sense of things eventually and did not.
My wife is itching to get it out, so spring time I’ll have my panels on it and a radio attached as well. 😁
We are alike in that regard. I have repaired, upgraded, improved an absurd number of things on our trailer at this point. If you want camping recs, let me know. I won't disclose my favorite places in public, but have no problem sharing with people in DMs.
I have 2 more E22 hats built. I would like them to go somewhere they can improve our LF51 mesh. Anybody that already has an rpi node with the problematic Waveshare hat, it's a direct swap with just a config change in /etc/meshtasticd/config.yaml
I don't have any spare rpis at the moment, so you'll need that if you're building a new node (plus sd card, power supply, antenna). A Zero 2W or 3A+ are fine for Meshtasticd. Instructions are here: https://www.youtube.com/watch?v=vLGoEPNT0Mk
Whoever claims them first gets them. I will probably build more.
How does one... Claim them
If you want one, you just did
How much?
Also wdym problematic wave share hat. Its like the best thing on my pi at the space haha
For PoE
The Waveshare hat in that youtube instructions has problems. You can get lucky and some of them work most of the time, but I noticed with mine it would sometimes choke on long messages and sometimes not.
Free for whoever claims them. This is $11 in parts. I don't mind building a few and giving them out to improve our mesh.
Sweet. Are you able to drop by the hacker space at some point for the handoff? Figure it kind of central and I gotta drive from up north
I will absolutely use such a thing on a pi node
Sure, you could also meet me at my work in North Salt Lake if that's less driving for you. Although I probably won't be back in the office until January 2.
There are more details about the Waveshare hat in #meshtasticd
#meshtasticd message
If you are using a PoE hat, I don't think it's going to work. I would suggest one of these style PoE adapters:
https://www.amazon.com/UCTRONICS-Splitter-Gigabit-Raspberry-Ethernet/dp/B07V35DH5F
https://www.amazon.com/UCTRONICS-PoE-Splitter-Gigabit-Raspberry/dp/B07CNKX14C
https://www.amazon.com/UCTRONICS-Raspberry-Ethernet-Chromecast-Compliant/dp/B0876N3PMZ
I misunderstood this. Waveshare makes a bunch of stuff. I'm talking about the Waveshare SX1262 LoRaWAN hat that is problematic for meshtasticd. Not a PoE Waveshare hat.
Where do you work in north salt lake?
heh, I can't DM you. It looks like you have to add me as friend first.
Worth a read, new role
Nice. Seems useful for roof top nodes haha. Though I get the distinct feeling this won't reduce peoples confusion with all the modes 😂 but thats a solved problem if people read the docs
This is amazing. It's exactly what we'd want for nodes that are in places like Millcreek Canyon where they might see some nodes in town (and also NPR), rebroadcasting, but still always needs to rebroadcast to push/pull things out of the city mesh.
Kinda worrisome tho about how long the delay is tho. If it's not long enough of a delay for a given hop_limit, I could see this causing issues with router nodes that need to receive ACKs from the original messaging chain.
- MSG: sender
- MSG: router1
- MSG: client1
- ACK: client2
- same time and now router doesn't see the message
- MSG: ROUTER_LATE
- ACK: client1
- ACK: router1
If ROUTER_LATE is suppppper delayed (30s+ it shouldn't be a problem)
When you say routers that need acks, what do you mean? It wont cause more utilization, as any node that has already seen the broadcast wont repeat it again. We simply might not get the ack from the intended recipient. Then again, if the intended recipient receives it 30+ seconds late and replies, the app "should" then note that they received it. What did I miss?
Isn't the point that this node will repeat even if it has seen a rebroadcast?
Looking that PR tho it looks like they covered my concern. (tho tbd if they are accounting for it with hop_limit 🤷♂️ )
Yeah, it will, as you said. I’m simply saying the other nodes won’t in turn start rebroadcasting again. (Broadcast storm)
I’m not disagreeing, I didn’t understand your concern and was trying to. 👍
Ah, it’s a concern that the node doesn’t quite wait long enough and it interrupts a different node trying to retransmit for the same message.
4 hops of client nodes could take 15-30s to get an ack
so it would need to wait at least as long as the slowest possible response for a message
It should be getting delayed further every time it hears another packet on the air, and it won't talk over existing transmissions if it hears them.
Are you able to clarify the concern you have regarding timing? I'm not sure I understand it. This mode isn't supposed to wait for an ACK - it's basically just a delayed version of ROUTER, and relates to flooded packets.
Thanks for working on this @broken turret
No worries 🙂
It should be getting delayed further every time it hears another packet on the air, and it won't talk over existing transmissions if it hears them.
It wont always hear neighbors without or signal.
Its possible a node will be too far to hear someone else transmitting, but they are still overlapping from the perspective of someone with less background noise.
I think I found a use my old Tbeams. Imma try out the other software. Meshchat and rnode on them. Its been attracting my curiosity in parallel primarily since they seem to have the user chat figured out a whole ton better lol or so they claim. So we will find out 😆🤔
Good way to use what I currently dont
I'm interested in how well that works. Keep up posted
I guess the cool part is. It mostly supports the same hardware and my 5 extra Tbeams won't become ewaste 🥳 . + best way to get through doubt's and questions 😂 is to test it all out
@real cedar is DC801C6 node down? I haven't seen it for 4 days. It's on a POE switch, right? Can it be remote restarted?
Not sure. I would ask in the hacker space channel
I mean its in a sealed enclosure and its on PoE so it can be toggled but idk about remotely. The hacker space does have a firewall
Thanks for clarifying 🙂
This just sounds like the standard hidden node problem, rather than something specifically applicable to the new role - it would apply to any rebroadcaster in that situation. Keeping channel util low, plus the random SNR-biased timing, should be enough to ensure that most packets still get through; there's no need to introduce massive delays (which would cause more problems than they solve). And obviously, make sure your infrastructure nodes have an actually decent antenna setup.
In wireless networking, the hidden node problem or hidden terminal problem occurs when a node can communicate with a wireless access point (AP), but cannot directly communicate with other nodes that are communicating with that AP. This leads to difficulties in medium access control sublayer since multiple nodes can send data packets to the AP si...
I'm curious if Lora would even support CSMA/CA - NAV (Network Allocation Vector)... In the military comms world its how mesh hidden node problems often get mitigated to an extent. Just a side thought not a try to do
It's a logical mechanism, not a physical layer one - so there's no reason why you couldn't use it on LoRa. However, it's generally implemented as a multi-packet thing, which meshtastic isn't - individual mesh packets are atomic.
The other way to achieve something like that without multi-packet would be allocated timeslots. But coordinating timeslots in a decentralised manner can get pretty difficult.
Ahh like like OLSR, AODV.. So now I'm curious. What is the draw back with meshtastic moving to multi packet?
I think RTS CTS would be very problematic when two nodes can't see each other but they talk over each other on say node b
Bandwidth. It already struggles. We don't have the overhead on the slower modes (in areas with a busy mesh) to add a decent RTS/CTS mechanism.
Why? That's the problem CTS is designed to solve... when node B sends CTS, then anything that can hear B and isn't the RTS sender shuts up.
The other problem with doing RTS/CTS on a meshtastic network, is all the broadcast traffic without a designated recipient. You can't use CTS for that properly, without having a CTS avalanche from multiple potential receivers, at which point you get CTS collisions.
Sorry I was caught up in making food and only finished a half thought. I was leading up to CTS collisions but yeah you pretty much covered it. IDK Id be interested in seeing what people do as a solution to the problem long term on the logic side
Might have over caramelized the garlic a bit..
Meshtastic actually uses CSMA/CA, but instead of an increasing contention window when packets fail it’s either based on channel utilization (for the original packet) or SNR (for rebroadcasts).
Yup - I was talking about the NAV bit (i.e. RTS/CTS or equivalent schemes, and also briefly about TDMA), not the carrier-sense bit.
Bountiful peak is getting one more when its a tad warmer and promotory point and the matlin mountains this spring is gonna be fun
Is there a formal planning process for where to put the high visibility nodes, or do you guys just discuss it here and then put them up?
So far folks have acted on their own, or discussed here. Nothing formal
There are a decent amount up, and more planned. Hopefully ones that will be useful to you. 🫡
Just ordered the rak wireless stuff you guys recommended a while back. You pointed me at https://store.rakwireless.com/products/wisblock-meshtastic-starter-kit?variant=43884034621638
https://store.rakwireless.com/products/unify-enclosure-ip67-150x100x45mm-with-pre-mounted-m8-5-pin-and-rp-sma-antenna-ip-rated-connectors?variant=42861623738566
https://store.rokland.com/collections/all-helium-antennnas/products/alfa-aoa-915-5acm-5-dbi-omni-outdoor-915mhz-802-11ah-mini-antenna-for-lora-halow-application
https://store.rokland.com/collections/antenna-extension-cables-pigtails-connectors/products/uflipex-ipx-mini-pci-to-n-female-bulkhead-pigtail-cable-extension-rg178
but now that I've paid, it looks like I could have gotten all of it at rokland and avoided needing to ship from china?
whoops, yeah. also if you want to order from rokland, including any batteries in the order makes it ship via ground. If you want parts faster you can split them up for a few $$ more.
ayy hell yeah. might want to go before the next storm if we are lucky enough to have it strike in the evening and not throughout the day
Thank you man now you and I can likely chat directly
it also opens me up to probably trying to see if I can stretch it farther into the desert where I am interested. Thought that will be interesting. Despite the long distances being quiet.. it may have to pipe back a good bit of traffic from the front
Gonna do the 8h hike in when the weather is a bit nicer i wanna hike in camp near the peak and hike back the next day after setting the node
I could drive up but it would be a fun hike
Promotory point is going to be ez just a long drive
The matlin mountains is going to be hell cause i need people at promotory so i can point the directional at it and have the local traffic node like i have a narrow line of sight from matlins to promotory so local antenna and directional to talk back to slc i need to test range and such but alot of hardware is being placed
Im out near the matlins alot when im out of slc i really want better coms out there mostly farm and ranches out there so not alot
The matlins are the best spot i could come up with for the west desert for my needs if anyone has better please share
@white field has a node in Onaqui Mountains. Unfortunately it's connection to NPR is spotty rn. I do sometimes get traceroutes back from it.
│ 10 │ 38fb onaqui router │ !d36b38fb │ 38fb │ RAK4631 │ pi8JCOlpFgJO14WF09SEtMsu0cW4KQ78gmFi5IKX0EQ= │ N/A │ N/A │ N/A │ 94% │ 1.94% │ 1.22% │ -12.00 dB │ 1 │ 0 │ 2025-01-07 20:36:47 │ 58 mins ago │
Route traced towards destination:
!359efdbb --> !0e81d07d (?dB) --> !d36b38fb (0.75dB)
Route traced back to us:
!d36b38fb --> Unknown (?dB) --> !359efdbb (-4.5dB)
Onaqui has kinda trash placement despite almost being a true summit at 9800 ft. It can only see north west to south east. (North side of a ridge on a bush)
I’d love to help with an alternative node that can help link to my skyline node
I have a really nifty one on Mary’s Nipple. It’s got low temp batteries, a temperature sensor, and is at 11k ft.
It’s got los to SLC mesh and can retransmit up to 50-70 miles if your nodes are in similarly quiet RF zones. Unfortunately, none of its transmissions are ever ACKed by NPR or POTM. (Which I think can see it)
I just looked an Onaqui wont have LoS to the matlin mountains. It would need to be on a ridge to the east of grouse creek to avoid LOS interference.
For planning your longer links. Learn from my mistakes, links longer than 10miles near cities can be really flaky. 😬 Once you're farther away 20-50 miles can be reliable. (50-80 can still work)
Is this running on LF51?