#US - Utah
1 messages Ā· Page 6 of 1
sorry for light spam today
not thinking i got the straight cables when i needed a 90 degree sma connector ... lesson learned ...
Hey everyone! New to meshtastic and love the idea of it all. I have a pretty decent location near the mouth of Little Cottonwood Canyon that I could set up a node at. I saw some google maps screenshots with line of sight overlay or something. How do I do that to see how this location would be? Or if that's not needed at all that's cool too.
I have so many cables, adapters, etc. from ordering the wrong things over the years. The worst is when you ordered it from Aliexpress and waited 3 weeks for the thing that doesn't work.
Put your node on Lone Peak. That would help out. š
Im waiting on a zillion g2 station and rak wismesh pocket v2 and antennas and bullshit
Alot of my ideas are dumb and based off parts i have on hand and alot of my fun money each month i budget for stuff like this
I'm guessing you can't techincally place nodes on public land right? But do people do it anyway?
Yes, maybe be more subtle about it than the people that made the news putting large Helium rigs up. It is possible to go the legitimate route: https://www.blm.gov/programs/lands-and-realty/communication-sites
Ali express please donāt fail me
where is the NPR node
Are you asking if itās responding for anyone else, or are you asking for its exact location?
I havenāt received any messages form it since 2pm yesterday
Oh yeah I'm just wondering roughly where it is
The individual that placed it intentionally didnāt give its precise position, but itās in the vicinity of the other radio equipment on Nelson Peak. If youāre familiar with the commercial installations, near there.
What are you trying to do, maybe we can help?
I just heard people trying to reach it as a benchmark for having a good presence and didn't know which mountain we were talking about.
Ah, yes. Nelson Peak has (had?) the best coverage for most of us. The POTM (Point of the Mountain) router also has very good coverage.
NPR seems to be down, for me, right now. I canāt get a message to it. POTM āisā responding, and has very good reception. If you are in the valley my bp_roof node is good to test against. Send it a āhelloā message, it has a bulletin board service running on it. Itāll respond quickly, so a human doesnāt have to be looking at their radio to help you test
It can even give you stats about what itās seen recently
We are on LongFast, but not in the default channel any more; weāre now on 51 instead of 20. The default channel was getting very congested. (edit for clarity)
ok cool. I created an encrypted default channel but I didn't set the slot so maybe it's in a slot that nobody relays
Channel 20 is the default, but will show as zero with default configs. Itās confusing if you ask me. That has a ātonā of radios on it right now. If you set channel 51 then NPR and POTM should both relay for you, even if they donāt know your channel.
I cannot say they will always be that way, but they are right now, and itās the default behavior.
If you are still on 20 there are lots of radios to pass your message, it just might not make it because of the other chatter on that channel.
ok fun I'll experiment. My brother has one in WX and I'm in Alpine so a lot of things have to line up.
Is anyone running the tc2-bbs? Mine keeps crashing because the radio reboots so I can't make it stay running
I am
what is the node called
bp_roof
ok I'll try to drop a note if I see it
It's working fine for me today
Route traced towards destination:
!359efdbb --> !0e81d07d (?dB)
Route traced back to us:
!0e81d07d --> !359efdbb (-3.75dB)
Earlier today I ran a couple of traceroutes from work in NSL to my roof node and they both completed and bounced through NPR.
Should we be occasionally transmitting messages to ask people to upgrade firmware and change to client mute? What else would be causing congestion on the main channel? We could suggest changing location to 12 hours for fixed nodes maybe for people who want their fixed station to transmit it.
We can make suggestions, but folks donāt need to do what we say, and a ton of them respond being complete dicks š
If you watch the default channel traffic fully half of it seems like 13-15 year olds, and I know it canāt be. A lot of the radios we never hear from are plugged in somewhere and never updated or messed with.
Simply put, the current flood routing design isnāt intended to support 100+ radios on long fast. The devs have said that even 30-50 are likely too many for long fast, which is why we tried medium fast.
I could write a short book on what weāve discussed in this channel, and what the devs might do to develop this for the future.
In the end, it was easier to move our group of folks that are communicating to a different Lora channel and try to get everyone on the same page.
Your suggestions are correct, and the newer firmware already rate limits telemetry if the mesh is busy. Mostly you just need them to upgrade and use client_mute
I must be to fat away the time I tried changing to slot 51 I just get max transmission reached
If you check the pinned message, you can see what the LF51 coverage looks like on slcmesh
the pinned message on discord? i know its a dumb question haha
found it just had to open my eyes
quick question do the Utah group has a group channel
You found it Mista š
Is NPR down? I was able to hit it last night but can't seem to right now. I put a temporary roof node up and I think I should be able to reach it. I'm in Sandy, about 86th south and 13th east.
I realize you probably meant Lora channel. We are using the LongFast named default for chatting. Because we are on Lora frequency channel 51 we dont see the rest of the devices in the valley.
What is the name of your node? I'll try and trace to it. If you dont want to say here you can message me.
It's "Meshtastic 40fe"
I don't have super great LOS from my roof, but I should be able to hit NPR I believe. Not sure what else I should expect to reach. Still kinda figuring out how this works. I'm on channel 51, longfast, client mode.
Do you have line of sight to point of the mountain?
No unfortunately. I'm right near Peruvian Park Elementary and here's the line of sight from about there:
Also not sure exactly where POTM is, but just guessing for the green pin.
Not sure how accurate the mesh map is, but it looks like the south east part of the valley is a bit sparse on nodes (except POTM of course).
I should "just" about be able to see you directly
What radios do you see right now?
Are any at all showing up for you?
If you are using a mobile client, disconnect and reconnect to the radio will force and update of the contact list
Let's get an obscure client node up on flat iron, that should cover your neighborhood š
Oh yeah, I'm just barely in the red.
Here's the mesh map screenshot:
Yeah I thought about putting one up there. I think there are plenty of places to hide one up in a tree or something.
See if you can traceroute to my bp_roof
or send it a "hello" message, it's running a bbs so it'll respond
It helps for testing
Looks like nothing unfortunately. I just threw a sensecap tracker t1000-e on my roof with a solar panel so maybe it just is too dinky of a device to get any good reception.
Honestly, the t1000's are decent
Both NPR and POTM traceroutes are returning for me rn. I have 40fe in my node list, but I'm not getting a trace back.
Hmm, you're correct, I dont see you yet
As good as the t1000 is, you might need a node with a proper antenna for it to be reliable at that distance.
Ok cool, thanks for helping. I might stick the t1000 on a long pole to get it up higher just to play around but I'll order something a bit beefier for a permanent install.
This thumb-sized LoRa dev kit, featuring the XIAO ESP32S3 and Wio-SX1262 LoRa module, is an ideal starter kit for Meshtastic and LoRa/LoRaWAN projects. Supporting WiFi, BLE, and LoRa, it includes a built-in power management chip and can be extended via IIC, UART, and other GPIO interfaces, with compatibility for Arduino development.
Did you test them on LF20 before switching to LF51? I would guess with the number of nodes on LF20, the mesh might work better for you right now. IME something like this would most likely reach NPR if you have LOS.
#1197577977781821541 message
These are pretty cool, but I think they are really borderline solar capable (you'd need a pretty big battery and panel)
https://discord.com/channels/867578229534359593/1329096340700921898
I would test on channel 20 first, then if that's working, carry the node over to flat iron and test on 51. If you are still good, only then do you need to consider more hardware.
I can get immediate responses from npr and my roof node to my t1000-e when I'm over on the east bench.
I can just easily hardwire a roof node to power so I'll check that one out!
I'll test out channel 20 and then 51 from flat iron and let you guys know what I find!
If you decide to go that route I have one other option; I could make you a RPi hat based on that E22 1 watt module:
#1197577977781821541 message
You would need a Raspberry Pi. A Zero2W w/ header or 3B+/3A+ with a PoE adapter would be perfect.
What's the benefit of using the raspberry pi? I have a spare pi 3B sitting around that I could probably use. Do I just need to connect an antenna to it and load up the meshtastic software and I'd be rolling? I assume you can also ssh into it and access your node from anywhere too, which would be neat.
Sending and receiving messages over shell isnt well developed for end-users. If you're already a code monkey it's ready for you
If you have an android device you can connect to the IP of the pi with meshtastic running, but Apple devices cant connect to the IP, only bluetooth
I'm a software engineer so I'd be pretty comfortable using lower level interfaces, but I'm guessing the main purpose of a roof node would be to serve other nearby clients right? I'm having a hard time keeping a bluetooth connection to the roof one as it is.
Exactly. My roof has power, and I dont connect directly to that node much. I do use the IP for the BBS docker to connect to. It's a good use-case for me. I run two nodes in client mode. Any of your mobile clients (used in the valley) should be on client_mute. They dont need to contribute to the mesh. Having a client on flat iron would likely be more than enough to server your surrounding area, to get them to POTM or NPR, and to one another.
Yeah flat iron does seem ideal to get LOS to lots of good places in the valley. I'll shoot for that and see what I can get. Maybe temporarily I'll just go hang my tracker up in a tree and see what we get then I'll order something new to replace it with long term if that seems to work well.
Oh but as for channels, I don't know if I understand very well. Should I do 51 or 20? 20 didn't really seem to help much on my roof. I couldn't hit bp_roof or NPR or POTM or anything. If I hang something in the tree I just want to make sure it is set right before I climb up there.
pi3b would be perfect. you need a hat with the transceiver, I have been making them (see linked post above).
Before that Femtofox was released, a Pi was the only feasible way to use the Ebyte 1 watt module (much more powerful broadcast than other devices you can buy).
With an rpi, you just use the meshtastic web interface (need PoE or wifi connection). There is a curses based client too, but I haven't tried it out. The command line is also nice for troubleshooting (sending many traceroutes, etc)
LF20 is a different frequency than LF51. You won't be able to reach NPR, POTM.. but it has way more devices. 200+ last time I was on LF20. You might also have line of site to Lake Mountain with is on LF20
Ok cool, maybe I'll play around with both but 51 seems better suited to hit those other key routers in the valley. I'll let you know what I find.
I have friends that live on Julho behind Flat Iron park (the trees at the top of the park are actually on their side of the fence). I bet if you build a node, they would let us place it in their trees. It would have to be a solar node.
That would be pretty sweet. That's probably great placement with a much lower risk of getting taken. I'll take a look at LOS from there. It's right down the street from me so I'm sure I'd be in range.
I live on a hill, my roof node is above 4500', and there are a number of houses higher than mine. Eventually I'll put my roof node on a friend's house up higher (closer to 4700'). In client mode it assists when trees or homes are blocking reception to NPR and POTM for those around me.
If you havent read already, the primary difference between client and router mode is how quickly they respond. The clients wait a certain amount of time after they see an ack request before replying, giving the routers time to reply first. Routers should be placed in very strategic locations. Anything moving around should really be in client_mute, as we dont need them to repeat.
If you're out hiking with family and friends, all that changes. You might as well have everyone in client mode, and on channel 20 as that's the default. That way any other radio out in the wild would also be able to help you (and you them) to reach the rest of your party.
Your placement seems good to me. I always get direct traceroutes back from bp_roof. I do wish my work node in NSL had direct to bp_roof. That would give some redundancy should NPR ever go down. Maybe a node at 4700' would do it?
@leaden crow thanks for mentioning the pinned messages yesterday. That video from the comms channel was a good one. The link for the coverage map wouldnāt work for me it just says 403 forbidden. I switched back to LF51 but havenāt picked up any other nodes for the 45 min Iāve had it going. I may have to build a node I can stick up on the mountain south of Provo canyon.
If you use a VPN that may be your problem. I noticed slcmesh does not work for me from a VPN. Probably Geo-blocking non-US address space.
Highly likely š
Iām not running a vpn. Iāll have to try another browser besides edge
I had fun climbing a tree today. Turns out I have a pretty big pine that was the easy button so I tried that and I can hit both POTM and bp_roof directly. "Meshtastic 40fe" is the name if anyone wants to try hitting it. It's on client mode right now, I'm thinking that's the right setting for this kind of node?
Oh and it's channel 51.
pi@dsr2:~ $ mtr.sh 40fe
ā 4 ā Meshtastic 40fe ā !40ce40fe ā 40fe ā TRACKER_T1000_E ā n71jei/uNpIqJdRZ4ZF5fVdjUOJ8adaFjQKrqOiZfhY= ā 40.6061° ā -111.8568° ā 1425 m ā 101% ā 22.79% ā 0.13% ā -13.75 dB ā 1 ā 0 ā 2025-01-17 14:13:23 ā 8 mins ago ā
No Serial Meshtastic device detected, attempting TCP connection on localhost.
Connected to radio
Sending traceroute request to !40ce40fe on channelIndex:0 (this could take a while)
Route traced towards destination:
!02d7e081 --> !0e81d07d (?dB) --> !40ce40fe (-2.5dB)
Route traced back to us:
!40ce40fe --> Unknown (?dB) --> !02d7e081 (-13.5dB)
!0e81d07d is NPR, so you have direct to that too
Nice, we're in business.
So that's just your T1000E up in a tree?
I need to play around with mine more. That's really impressive for an integrated antenna.
Yep that's it. I had a spare solar panel just sitting around and plugged that in and strapped it to a branch then dangled the T1000E from a branch with a good view of the south/west part of the valley.
I saw some youtube reviews of the T1000E vs other devices with external antennas and those people said the T1000E can hold its own against other devices pretty well.
Iāve been thinking about a lone peak router too. Anyone have experience building nodes that would survive in that harsh of a climate? Iād hate to install one then have to do maintenance on it.
The unfortunate reality is that you should build it planning for maintenance. You can remote admin settings, but not firmware. It's likely future firmware will need to be updated (3.0 will likely have breaking features).
The real concern is batteries. Very few are rated to be charged below freezing. Some people are using these:
https://store.rokland.com/collections/batteries/products/molicel-21700-p42a-4200mah-45a-battery
But if you read the specs, they are still not rated for charging below zero. Honestly with the low charge rate from solar over the winter, I think they will probably work, but at some reduced life.
LTO batteries solve this problem, but you need a special charge controller for them
A lot of the mountaintop nodes we have now; Onaqui, FPR, NPR are mostly using standard 18650 or 21700 cells, but all of those nodes, this is their first winter. We don't have long term tests yet of these things.
I plan to place one near desolation peak next summer for wasatch crest coverage. I will start with the Molicel 21700 and see how they last.
#solar-power has a ton of advice and builds, you could probably read for days
Yeah that makes sense. I donāt mind taking a hike once a year or so but would hate to have it breaking every month or something.
Iāll do some reading and if I get serious about a build Iāll check back in!
I wonder if a box with a clear cover would do the greenhouse effect
Uhhh
Well we can test if ya want
Put an environmental sensor in that and monitor temperature outside in the day
Super easy with RAK. They are less than $5
https://store.rakwireless.com/products/rak1901-shtc3-temperature-humidity-sensor
Im waiting on batteries but ill make sure to add that
The front cover is going to be covered by a solar panel in my stupid mind
āTestingā
https://www.seeedstudio.com/Grove-Temp-Humi-Sensor-SHT40-p-5384.html what i would be installing
Grove -Temperature&Humidity Sensor (SHT40) is a digital Humidity and Temperature Sensor which has an accuracy rate of up to ±1.8% ~ ±3.5% RH and ±0.2 °C, Ā and an operating range of 0% to 100% RH, -40°C to +125°C. It comes with a SEEEDās standard grove l2C interface enabling developers or even beginners to easily communicate with the STH40 sensor...
What is the wattage on that panel. Iām thinking you could just add a heater. š
Its rated 12v 20w Iāve gotten 13-14w in my testing
I got a bunch from other projects they are shit chinisum but i got a bunch
Sold as game cam or deer feeder panels
That is a good idea, run an oversized panel and battery heater.
Victron does it for their RV batteries
Iāve honestly been really impressed with the device so far. Has great reception.
Anyone here own MeshTimStick?
Anyone ever set any nodes up in national parks? One or two well placed in Zion would give pretty good coverage of the main popular areas.
Several members of the Utah Amateur Radio Club have expressed interest in setting up Meshtastic Nodes. Does anyone have concerns with the club recommending Frequency Slot 51 to club members?
I think Iād recommend channel 51 to anyone thatās willing to watch this video, see what we have already done, not put up needless routers, and leave their mobile nodes in client_mute. š
For today's video, we'll be going over some Meshtastic Deployment Scenarios and what settings are best for them using this map we made (link below). We'll also try to bring awareness to some issues with Routing and overuse of CLIENT and how CLIENT_MUTE may be a better choice in some scenarios.
I have not used mine much yet. I would not have guessed how well they work. If the position of Peruvian & Jupert nodes is accurate. They shouldn't have good LOS to my roof node, and are ~5 miles away. I am still getting direct traces.
Link? Haven't heard of that device.
No, itās a node name
I would be hesitant to do this. I haven't looked that closely, but I wonder if you could find a decent position near the observation point trailhead, just outside the park? It would be amazing to get node(s) above Goblin Valley and The Maze. Those areas have little or no cell service.
Good call. Iāll share it.
Peruvian is my tree one. It's probably ~40 feet up in the air. Jupert is my other tracker card that I usually have either in my front window or just in the kitchen or somewhere else inside. I'm actually surprised you're getting direct traces to Jupert.
So on Frequency Slot 51 I have noticed that messages get received but confirmations often get lost. Is any one else seeing this?
Theres an exception that introduces a problem. The nodes in the denser dead spots in the downtown area. It can introduce a hidden node problem but if say somethings in a cellar and you have two devices and the cellar device isnt getting much having the other window node be a client might be helpful right?
I agree, itās not a strict rule. There are times client mode in a mobile node makes sense. I think some client modes in the downtown also make sense. Any time Iām camping all nodes are clients, to make sure everyone gets messages.
I see this āRepeaterā often in my trace routes. What is it? Why doesnāt it show up in my node list?
I should say, I googled the repeater role and it makes sense but I guess it still doesnāt show up as a node?
They're not actually repeaters. I don't think there are any on LF51. They are nodes with firmware older than 2.5. They appear that way in traces.
#general message
Oh interesting, I guess that explains why they donāt seem to show up in the node list. This ārepeaterā probably actually is in the node list somewhere but since it isnāt actually a repeater I wouldnāt be able to tell which it really is.
The mars station keeps logging aircraft with active transmitting radios. This is "Sky Router" (a station g2) and sent texts at each marker. Flying at 35k ft.
Took a stab at putting together a solar node, hopefully Iāll get it set up this weekend and see how good it works.
Nice @opal igloo! Will it be in the salt lake area?
I've ordered the parts for one (minus the solar panel and batteries) and will work on it soon too. It's my first and I think I'll mostly do the same as you if you have any pro tips or things you'd do differently.
@long crown Iām down in Orem so Iām pretty sure Iām the furthest south on LF51. Im far from a pro haha, I kinda mixed the build from my first node and the solar portion from a video from the comms channel on YouTube. I would say the one thing I would change would be the adhesive I used to glue the solar panel on. Itās taking forever to cure. The old comment section on YouTube suggested 3M 5200 I think if I did it again I probably would give jb weld a try or some adhesive I could get my hands on from work.
Very nice. I just completed mine and will get it up on the roof soon.
Coming back to this @leaden crow, I'm tentatively planning to put this solar node I'm building up on/near Lone Peak. I've got a brother in Cedar Hills I'd love to close the gap to.
It seems like a -4 F charging temp rating would probably be pretty ok. It will obviously get colder than that, but I'm thinking most of the time during sunlight hours, in direct sunlight the batteries would be warmer than that. Do you agree?
Seems like LTO or other types of batteries just don't have as straightforward of a build and lithium really seems to have an easy build.
Based on how NPR and POTM are performing this winter. I think you'll be fine. I'm tentatively planning that I will just replace the 21700 cells annually in my Wasatch Crest node (when it's in place). Although maybe that will change if I replace them after one winter and the old ones are still fine.
If you're buying the Molicel batteries, you can find them at better pricing than Rokland. Like here:
https://www.18650batterystore.com/products/molicel-p42a
Ok sweet, I'll look around a bit then!
Is there any way to identify the āunknownDBā
Does this mean that there is a problem with the Nelson Peak Router Database?
You could try asking on the public channel... Who is operating as a repeater? I think a small group was using ch51 for a private channel before we started. One of their repeaters constantly intercepts and relays messages to me even though I have direct connections.
There is no problem with NPR, it is on firmware 2.4, which does not add it's node number on intermediate hops. I think NPR and POTM are the only devices on LF51 pre 2.5 (everything else has the lock).
They also happen to be the best routers, so almost everything hops through one or the other.
#general message
GUVWAF is one of the Meshtastic Devs
When Gdane placed those last year, a lot of the 2.5 firmwares were a little sketchy. So it was a safe play to leave them on 2.4
Got it. Thanks. Learn something new everyday.
New here, down in south Utah county. Got a few radios, hitting routers and nodes all over but thereās never any chatter. Am I missing something
People aren't usually chatty on the public channel. At most you'll see test messages once in a while.
I have created a channel called 51chat, PSK = 1akfONiXDy5pXV8oLhvOo/ir4WVH+semS5O6jbuGiwY=
Hopefully people will feel more relaxed about communicating since the open channel is forced on everyone, and this is voluntary.
I propose starting each message with an increasing number based on the previous msg. This should help people understand if they are missing transmissions. If you are responding to a comment, start your message with #@# (other msg), e.g. 5@3 Deepseek watched all of TikTok and no longer functions.
Reset from 99 back to 1.
Have anyone have the t-deck or the t-deck+. I just know with is better and is it worth it to invest in one?
I wouldn't buy one until at least the touch preview is out. You can build it yourself, but it's not really useable IMO. I built and installed 2.5.20 last weekend and it still reboots getting messages. Touchscreen is laggy and imprecise, scrollball is just frustrating--clicking often scrolls then clicks for me.
I like the device, but I wouldn't be surprised if touch isn't a useable beta for 6+ months, and at that point there may be even better hardware.
@main stag Can you edit your pinned message and add FreeBirds channel to the bottom?
Otherwise I think we would have to find an admin to have it added to pinned messages.
The 51 chat I've put it in over and over still not working
Touchscreen is laggy and imprecise
I've noticed that after a doing a screen calibration, and until the next reboot, the screen becomes noticeably more precise. BTW, I'm from #1202833898376138752, not Utah.
Nice, thanks! I'll have to try that.
You can't send/receive messages, or you can't add the channel?
Just making sure everyone knows, the default channel can be removed without issue. If you remove it and set your own private channel in the first position you just have to ensure you have the correct Lora channel number, in our case 51. On 51, the default channel āonlyā includes people on that Lora channel.
We just donāt chatter a lot on the devices, itās more efficient to discuss here.
I only tried to add a channel from a QR code once and it replaced my default channel. Now I always use the cli.
meshtastic --ch-add 51chat
meshtastic --ch-set psk base64:1akfONiXDy5pXV8oLhvOo/ir4WVH+semS5O6jbuGiwY= --ch-index <Number returned from first command>
It might make sense to use a much shorter PSK like LongFast's. I wasn't sure of the exact format requirements and used the automatically generated one. Sorry people are having trouble connecting.
Well we wouldnāt want anyone to crack the encryption š
I think it is fine, good learning opportunity. Especially if we could get the messages on the grafana dashboard so people can check what messages they are missing.
Iāll say this only once more, and Iām am not attempting to tell anyone what to do. We donāt really need a new channel, respectfully.
The default channel is āonlyā radios on channel 51, us. We also do not āchatā on it, we chat here.
If everyone wants a chat channel on Lora then by all means, itās simply not a need.
I feel like Iām saying folks shouldnāt have fun. Please donāt feel that way. Enjoy the hobby the way you wish. ā¤ļø
Apparently I can't make the message longer since I'm not 'premium'. We'd need to pin a new message to go with the meshtastic dashboard one.
I can pin one š
I agree, but if people are more comfortable testing on a non-public channel, have no problem with that either.
Who is PeruvianPi 2843?
It appears your fixed position latitude is a bit off. š
0.0262° ā -111.8568°
Oh thatās me! Haha yeah I noticed that, will fix soon!
Iām going to replace my t1000e on a string with a pi setup.
Anyone have problems with Ethernet wires in trees? Iām wonder if squirrels will chew through it and if I should run conduit. The pi will run by poe.
This stuff might be easier than conduit
https://www.techflex.com/heavy-duty/rodent-resistant-sleeving
I think that is probably overkill. I have never seen squirrels around here chew through utility cabling. UV would be a bigger concern. I have seen cable jacket fall apart after a couple of years in sunlight. Maybe get outdoor rated cable?
Ok cool, yeah Iāve got a big spool of outdoor rated catv so Iāll probably just run it and cross my fingers that the squirrels donāt find it too interesting.
I ran across these and was wondering if you had details on building the hat available? I was considering putting a couple together for fun. It looks like a basic SPI + a few GPIOs.
You define the pins you use in meshtasticd config.yaml
I just used the same as MeshAdv
https://github.com/chrismyers2000/MeshAdv-Pi-Hat?tab=readme-ov-file#configuration
I didn't care about GPS, SMA or the headers on the Adv hat. So mine is just the E22 module and the 2 filter caps on the 5v
Great, even simpler than I expected. Thanks!
Doubt it is actually 12000mah but it was what i have on hand
Gps is not in waiting for grove cables
Wait, whatās wrong with the waveshare hat?
#meshtasticd message
They were like 50/50 on ones that had problems (could only send short messages, or not), but with the PSK change in 2.5 the packet sizes increased and way more people were noticing a problem.
I thought my waveshare was fine until I started noticing the problem in the debug logs, and then did more testing and it definitely failed at sending or re-transmitting messages.
The theory is that they use a XTAL clock that gets hot when sending messages and causes a drift in frequency. Meshtastic has longer broadcast intervals than regular LoRa, so is more likely to cause the problem.
Better transceiver modules use a temperature correcting (TCXO) that doesn't have the problem
Dang i just built one with the waveshare hat. I had found a guide elsewhere that recommended it but I guess that was bad info.
I gave away the last 2 E22 hats I built, but I have one more module. If you want it, I can probably build the hat this weekend.
Yeah Iād love that @leaden crow if it isnāt too much trouble. Happy to pay for it too.
Ive got this piece laying around too if itās useful at all.
I'm not worried about the cost. It's about $11 in parts. Spreading a few of these around the valley will make a huge improvement in LF51, so I'm happy to give some away.
I'll DM you when it's done.
What connector are you using to mount that antenna @rustic bolt? And where is that beefy boi going?
Sma to type n all with a rubber seal and some all weather hot glue
Its going to be placed on the roof for testing if it handles weather ok then you might see boxes like these placed in the north west side of the state
i still need to finish it up but its mostly ready to go
and 9/10 instead of a g2 i will go rak cause power consumption on the g2 is alot more for near same (w/o using the pa) performance
Yeah it looks sweet! I'm working on my first two builds at the moment and learning lots. One will be in a big tree at my house with a Pi and the other will be a Rak hopefully going up towards lone peak once it warms up.
I have too many radios, I need to make one of my RAKs into a solar node
They are acting as decorations currently, poor use of them
Hey everyone. I'm working on some nodes down here in STG. I was wondering if anyone else in here is too. If so, would you be interested in collaborating. Thx.
I was just biking in STG last weekend and noticed how few nodes are down there. I only saw ~4 on LF20.
You better hoard them. Who knows what's going to happen with the tariffs? š
I just panic bought more E22 modules on ali yesterday. Hopefully they make it.
well tariffs will raise costs the usps freeze wont affect most of the shippers as places like ali and dhgate set up wear houses and last mile devlivery a while ago its just prices we will see go up unless the trade war gets hotter
I was mostly joking. Definitely not to 'sky is falling' yet.
Yeah there's not a ton. Almost have a blank slate to work with š
I've tried to put in the key and it's not working
Do your settings look just like this?
Thanks
What amp are you using? I've seen some mentions on other channels on this discord of cheap aliexpress amps performing worse than without.
I would probably leave my stuff on LF20 too until you have other options. I have a couple of friends that bought T1000E, but I left them on LF20 until they get roof nodes or something else improves their connection to LF51.
Not good. Hopefully we get several new mountain routers this year.
You still have people on MF up north?
First I've run across this, figured other might have missed it too: https://meshtastic.org/blog/meshtastic-site-planner-introduction/ & https://site.meshtastic.org/
Its pretty decent lol. I came across it when looking for alternatives to heywhatsthat
Also man the Utah job market is awful rn
At least for engineers
Might be better for other professions
I think it's going to get worse before it gets better. We've had a few entry level positions open and were inundated with applicants. One of my friends echoed the same sentiment. He's getting very senior people applying for entry level jobs.
Yeah it's because most of them want out of the crud that is anything over a medium sized company.
Unfortunately the large companies act a lot like DOD contractors. In that they are the integrator and not every engineer wants to spend his time doing that type of work
I would say this is more dependent on specific employer and not company size. Have worked for F500, Gov and small businesses. Current small business job, I wear like 5 hats, do 50%+ more work. The grass is not always greener on the other side.
It's because of automated application submission and AI.
Those senior people are not actually interested in entry level jobs. Their bots are just ruining it.
I think this is only partially true. In my friends case, he works for a really desirable company. Some people just want a foot in the door.
Ohhh I've been in each as well but I've always liked the hustle and bustle of small to medium. Sure I would do more work but I'd enjoy the work vs come in everyday like "shoot me. Another day another dollar"
I see your message on slcmesh, but did not get it in Holladay or NSL, and slcmesh did not get my reply.
Do you want a 1 watt hat? I have a spare right now.
slcmesh saw my reply. it just took a bit to show up.
Did you order them off Ali? I am still using the waveshare. I should try sending a long message out of mine to see if it has the issue others have seen.
I ordered the modules off Ali
https://www.aliexpress.us/item/3256807016621535.html
I've just been soldering them on proto hat boards. I have one build and a couple of modules to build more if you want one.
My waveshare hat seemed like it was okay in testing long messages, but I don't think any of them really are when channel utilization goes up and they get warm.
https://www.youtube.com/watch?v=PND1GlMSrEM
Ah, I see. I sent a few max length messages and that was totally fine, just as you say.
Damn, those are a great price. I might make one after all.
I've been using this proto board
https://www.amazon.com/dp/B01MY8TYSD
and use the Mash Adv pinout (so the config is compatible)
https://github.com/chrismyers2000/MeshAdv-Pi-Hat/blob/main/V1.0/Schematic/Schematic_Mesh-Adv-Pi-Hat-V1.0.png
Yeah I should try that as well. I was looking at the new planner site provided by Meshtastic and it feels pretty accurate given the ground clutter around me and where i have my antenna on the roof. I really only have direct los to the capital and south above point of the mountain. NPR / SLC is blocked by some houses right next to us that are taller so it seems I don't consistently see traffic from there. If I got the antenna to the peak of my roof I'd be seeing the nodes up by Grandeur peak west side.
IME the 1 watt radio can better overcome broadcasting through houses / trees. The one I have in NSL is inside the ceiling of a commercial building that has more structure than residential stick framing. It consistently reaches NPR. The previous T-beam I had in the same spot/antenna couldn't reach anything. Offer stands, you're welcome to the one I have built rn, and I have a couple more modules to build more if anyone else wants them.
I ordered a couple of these to build Zero sized hats:
https://wiki.seeedstudio.com/wio_sx1262/
But they don't have the amp the E22 has. Probably won't be as good.
I've been using one of the e22 hats as well and it has been great. The node using it is PeruvianPi and it's mounted pretty high up in a tree at my house.
If anyone wants to use it for testing I set up an auto responder bot that you can send any message to. It's maybe a little too snarky of a bot so if I need to take it down a notch let me know haha.
801 Labs has a Meshtastic Node Building event on their events calendar for March 13
#745157117337403512 message
A couple of people expressed interested in meshtasticd nodes, so I'll most likely go and take a few hats to give away.
fun
The slides I used on Saturday for Utah DCC: https://docs.google.com/presentation/d/1PYlcqHhvzJZOTWwN_vaB-AK0Zji_hNcWag889cMgBhg/edit#slide=id.p
New seeed nrf52 device coming
https://www.seeedstudio.com/XIAO-nRF52840-Wio-SX1262-Kit-for-Meshtastic-p-6400.html
So cute and little
Are there any nodes in Morgan, Utah? Planning to install a router there soon.
I might finally get motivate to place another node on my chain to Moab if this weather stays nice.
I'll be in Moab the last weekend this month. Happy to test anything I can reach.
I'm still 3-4 nodes short given what I've learned about radio interference and node range.
Actually, I take that statement back. Some of my nodes are live, but they are too far apart.
I'm really curious to see if the node I placed at 39.036834, -111.577865 is still alive after being on a summit at 11k ft for several months.
My summit node on onaqui is still alive even tho I didn't winterize it's batteries. (5aH 3.2v pack from AMZ) Apparently it restarted 3 weeks ago and forgot it's name. (probably a crash since the charge never drops below 93%)
Has anyone seen traffic from Fool Creek Peak recently? It had an intermittent link to my skyline node at 39.036834, -111.577865.
Have you thought about testing a cavity filter? From what I have read they can be a huge improvement in high radio interference areas.
https://acasom.com/products/915mhz-5-types-cavity-filter-for-helium-network-filter-waterproof-lora-indoor-use-high-out-band-rejection
Unfortunately these ones have a pretty wide band. Some guy bought up all the tuneable ones on ebay and had them tuned to a really narrow band.
I have not seen Fool Creek recently, but I reset my nodedbs less than a month ago. Can't remember before then how long. Onaqui is Meshtastic 38fb right?
They problem is that NPR is the crucial node getting the interference. š¤£
Skyline can see NPR, but NPR cannot see skyline. I also get flakier than desired perf between NPR <-> onaqui. (yup, Meshtastic 38fb now that its name got wiped)
My tentative plan for this spring is
- place an intermediate node that supersedes onaqui (closer to NPR so I can get a rock solid link) - ideally in radio quiet spot so it can hear a far away-next hop
- place a node in a quiet spot that can see my onaqui replacement + skyline (this node makes a link from SLC to Henry Mountains possible TBD how reliable it would be)
- after this node goes live, I'd move onaqui to client mode
- place a node in the Henries - this expands coverage out to Moab or Indian Creek (assuming the range goes that far š )
- move onaqui to an overlook in Indian Creek to put rock climbers on the SLC mesh š
The real trick with this many hops is that all of them need to be pretty reliable for things to work. Messages from my house to IC would be 7 hops.
95% reliabiliy for each node would mean it fails to send 35% of the time. (and fails to ack 70% of the time)
https://meshtastic.org/blog/meshtastic-2-6-preview/#improved-device-state-file-management
This might also help with the config loss
Also, 2.6! š
I used these map tiles (OxedOSM USA (-AK +HI)) for my t-deck. It makes the map usable:
https://www.reddit.com/r/meshtastic/comments/1j1chem/meshtastic_26_map_tiles_with_higher_zoom_levels/?share_id=A2UdvVu_0wA4tOlJwIJj6
Warning, it did take almost 24 hours to copy the 826,695 files to an sd card.
I moved my roof / pi to the 1w module last night. So far I'm seeing better traceroutes and was able to get some messages quickly routed to @vital hemlock
I loaded 2.6 on my brother's two tdecks
"id": "!d36b38fb",
"longName": "Meshtastic 38fb",
"shortName": "38fb",
We have a new "router". Does anyone know who that is?
That is onaqui, it got reset over the winter. Jacob mention changing the role when the snow melts and he can access it again.
Nice!
I got my first traceroute back from you ever
!359efdbb --> !eb262843 (?dB) --> !0d4f215b (?dB)
Route traced back to us:
!0d4f215b --> !eb7b95c5 (6.25dB) --> Unknown (?dB) --> !359efdbb (-12.5dB)
Ah that's awesome since we're so close. I have mostly los to your general area.
I think we must have some obstruction. The traces I'm getting back are usually hopping through NPR or something else. I get mostly direct traces to dsr3 (in South Jordan) and bp_roof which are both much farther away. It's still a huge improvement.
Yeah there are some buildings in the way. I also think there is a bit of a rise between us too.
Do anyone use the 51 chat I was trying to see if it's working or not
On long fast, with the default key? Yes
My long name is hambone lol
Oh, the special chat channel, I do not. Default only
From Ogden to Pleasant Grove
I sent a message on the primary LongFast channel, did you see that?
I'm using heltec tracker
Ah, north. Might be too far north for me. Set your hops to 5, from the default 3
I recommend it for the price and don't worry about buying a gps
I also have a few heltec v3
If you double-press the user button it will send a location ping to the mesh. Other radios should start to see it
I have one of everything at this point. I have too many radios š
My tracker is in a square little case I made for it, so I could stuff a huge battery in it
Lol yea I don't want to get to that point
Did you increase to five hops? Iāll send you out another test
No new radios yet. Do you show up on the dashboard?
#1197577977781821541 message
Slcmesh
Nope
I checked my node in NSL. I only see bp# message and bell. No hambone unfortunately. You are in Ogden? Mista_G
I drove to Logan the other day and tried to run some trace routes as I went. It wasnāt very eventful outside of salt lake valley on channel 51.
There are a bunch of radio towers on the mountain right between Davis and salt lake county. I wonder if it would be hard to get something placed there.
I meant my ensign but Francis peak sounds interesting too! Would that one have connectivity to NPR?
When Francis peak was up last year on LF20, it could connect to Lake Mountain in Utah County. It should reach NPR.
I don't know if it's your antenna or LoS, but your tree node is better than my roof node. We swapped out the RAK with an E22 hat last night at 801 Labs downtown. The traces I get back are going from your node all the way to downtown:
!eb5de3e2 --> !0e81d07d (?dB) --> !eb262843 (-9.5dB) --> !6782905a (-7.75dB)
It's got to be the LOS right? We compared my Ali antenna to your Alfa one and the Alfa one was significantly better I think.
I measured the height and it's 55 feet tall so it get's a pretty good view of the valley!
idk, I don't think the SNR and RSSI values are always accurate, or the whole picture. The guy that placed Lake Mountain had interesting insight that higher gain antenna performed worse for him on the mountaintop node. My roof node might have too high gain.
So
If I wanted to make a Station G2 store and forward
I'd probably have to be router_late to be as client like as possible huh?
Yes it will because of antenna aperature. The angle in which is optimal for the omni to receive. You probably heard this as beam width and as you go up in gain your beam-width gets smaller. For a mountain top node I don't think I'd go past about 6 dBi
I need to be able to see downward quite a bit
I think im going to do it for now guys. I don't usually but I'm on the hill at the high point and am the Only node with the potential to see Ogden (I see them now and then). I'm going to do ROUTER LATE to prevent being an actual router unless absolutely necessary.. Reading the docs it seems to suggest this is the desirable behavior of the mode. Then I can enable store and forward and at least give the massive dead spot intermittent connectivity.
Hell yes I can see NPR AND I can see a lonely node in north Ogden. Success so far. Chan util only 5% Max so not bad. I expected neighbor info and store/forward to use more
Ahh yeah there's the whole valley. Giant explosion of notifications š š„³
So if yall see Aurora Jovialis that's this node
Note that store and forward doesnāt work on the default channel.
You also have to have a private channel in the first slot.
And I think you can use the server role, instead of forcing another router. But I donāt see an issue with either
https://meshtastic.org/docs/configuration/module/store-and-forward-module/#is-server
This module allows you to resend text messages after a device has been temporarily not in LoRa range of the mesh.
I wonder if that lonely node in Ogden is @maiden wraith or @tired stirrup ? I see your two Aurora nodes on my map up North now.
I was using it for a private chat yeah
HAHA yes. Success. If they ever get a channel in common with us in Ogden my store and forward will work to them
Makes me wonder. What if we in a coordinated manner made a non default valley wide chat solely for the purpose of storing and forwarding something public? Would that improve things or congest things?
Yeah I was just reading that. So wait. I'm curious. What happens when something's a router and a server?
@vital hemlock hey I read all the stuff in the antennas section. I'm kinda curious what antenna really is best for the mountain tops. I have the desire to throw the signal as far as I possibly can get it and receive from as much area as possible. This is more a lazy request. I could draft it in keysight ADS but don't want to open my vpn for it haha
A lot of folks have used the Alfa 5dbi, but we all know that a true 3dbi would be better for anything high. I think the Alfa is actually lower than 5dbi, and generally itās been tuned really well.
ALFA Tube in one image shown for demo purposes only, this is just for the antenna. This is a heavy duty outdoor 7 inch tall mini 5 dBi gain antenna for direct connection use to a N connector. This has N-male connector and is designed to be connected to a device or pigtail with N-female connector,...
I tried sending a message to aurora jovialis from Sandy but I donāt think it went through. At least no ack
So you guys aren't a fan of the rokland 8dBi up high like the bay area?
No sadly nothing
Normally I can see DSL NSL or whatever but I can't
I suspect my issue is probably the buildings around area that are getting taller. They are building a 7 story massive apartment complex right in my view of north salt lake from here
@vital hemlock I am trying to cobble together RPI Zero 2w based nodes for up in the mtns. I've got a build idea with the right sized solar panel, charge controller and lifepo4 batteries. What I am deciding between is either the rokland 8.5 dBi 40 cm length antennas which I've had extremely good luck with but they only have a 20 degree beamwidth or something like your suggesting. I wanted the VSWR to be under 1.5. My goal was to listen to as much as I possibly could and transmit to as much as I possible could. Ben Lomdon is ultra prominent
My plan was Raspberry Pi Zero 2 W, Meshtastic Advanced Hat for Raspberry Pi Zero, 20W 12V Monocrystalline Solar Panel, 12V 30Ah LiFePO4 Battery, 10A PWM Charge Controller, DROK Buck Converter 12V to 5V 3A, Gratury Junction Box IP67 Waterproof 11.8 x 9.8 x 5.5 inches, Renogy MC4 Solar Panel Cable Connectors, Krylon Camouflage Paint, (ANTENNA undetermined)
I bet I could make that survive a week without sunlight in - 20 F. That's what it's sized to do
@leaden crow had an idea to put an LTE hat on a pi to do updates remotely too, that would be sweet so youād never have to hike up to it.
You can use an admin key to admin a node from Another node 1 hop LOS
Its way cheaper than a LTE node and I've found it to be reliable
I have an LTE hat you can have if you want to figure it out. I havenāt used it for a couple of years, but I bet it still works fine.
I thought about that for a Lake Mountain node (they have power in the shack up there). Iām not 100% what pins the LTE hat needs. Might need to make a custom pinout e22 hat to make them both work
You would want to setup a reverse ssh tunnel that runs on startup and probably have it auto reboot at least weekly.
Admin key can only do settings though. If we had LTE on a meshtasticd node, we could do firmware. NPR has been stuck on pre 2.5 all winter because nobody can get up there.
It wouldn't be too hard. Setup LTE. Setup an antenna. Setup a wireguard endpoint or tail scale. Voila
I checked. This is the LTE hat I have:
https://www.waveshare.com/wiki/SIM7600X_4G_%26_LTE_Cat-1_HAT
pin definition looks like it would have no problems with the MeshADV pinout, and it has pass through header on top, so they would just stack.
Like I said it's been a couple of years since I used this. I know for sure that I couldn't get it working with a Verizon SIM. Verizon supports the module (SIM7600G, but not the device). You can actually spoof any IMEI number, so I wouldn't be surprised if you can get it working with Verizon. I think I did try that, but maybe a bad IMEI I got from a generator.
AT+SIMEI=354893099621849
what this does is spoof the module to be seen as your tablet or phone IMEI
Yep, number of ways to deal with it being dynamic address assigned by the carrier.
Tailscale is lowest bw needs TBCH
@leaden crow
I figured out my problem up north. Theres a HUGE 7 story apartment condo unit going up that obstructs the Fresnel zone between all things north of it north of Layton. So while I can broadcast to you. Yall can't get through that darn building
Its also an eye sore
So yeah rufio and I at the space are planning grandview after the mud and I Might just get impatient and put something on antelope
Nice, think you'll build it at the hackerspace? You should just build up the spare RAK you have now. I might have some hardware to donate if you don't have everything you need already.
For mountain top nodes I desire that hat you built. It's second to none bro. Well fucking done tbch
It has an extremely good oscillator and reference
Its not the wattage its the stability
haha, well I don't deserve that much credit. I only built one initially because meshadv was not in stock and I copied his schematic.
I kept making more because they are so cheap. My last order from aliexpress, the proto hats were $2.25 and the e22 modules were $6.37.
It really is an awesome module
I have zero complaints.. No issues out of any of them so far
Would love more built.
If it's so cheap we should get the space on the idea of a RPI zero 2w and a module
Its way better than a heltec haha
If you want to go that route, I think we should figure out the cost of getting femtofox batch directly from JLB. Depending on quantity, it might be cheaper. This just dropped:
https://www.youtube.com/watch?v=lOPbck4ysos
This video goes over the Femtofox which is what we're using in our latest solar build for Meshtastic. This video is an overview of the Femtofox and Solar build. We'll get into the actual build in the next video. For now, if you want to purchase the parts needed, see the links below
Femtofox Github
https://github.com/femtofox/femtofox
== F e m ...
The luckfox pico mini are $5 on aliexpress
Let do it. You want to RFQ them?
Ill talk to the space people
Confirming a hunch dragon. Will see if I can get around that complex
Its in a natural vision blind spot
You'd actually have to know where to look
I just caught 47 NODES!! Hell yes. In the morning. See if you see my 10-2s from each device
HELLLLLLLL YES.
Description Provided 902-928MHz N-female/N-female Cavity Filter Why do we need a high-out-band rejection filter? The first function of the filter is to filter out interferenceĀ signals. Therefore, the higher the out-band rejection is,Ā the closer it is to the working frequency band, the better. A filter without marking t
PRODUCT OVERVIEW This is a small, cost-effective band pass filterĀ centred at 915 MHz with a typical insertion loss of 2.5 dB in the pass band 902-928 MHz. This filter has excellent rejection specifications. Rejection at 880 MHz for instance, is typically greater than 40 dB. The filter also provides power handling of up
š¤š¤š¤
Good for raks and lesser power devices
I got your tests on DC801 and public channel. Have been tempted to order that Acasom.
Haha so dude I must be in a quiet zone. I heard your router and its relaying my stuff all at e way up here
I did for giggles. I'll let you know how it goes with it.
@white field thank you for a relay down into the bowl š
Where ffff
Wait what
Onaqai
Isn't that behind toele LOL
(South)
The ffff on the route traced back to you is NPR. That happens on old firmware (pre 2.5). It doesn't report itself on hops and appears that way.
Ahh. If it's a rak I can offer to take a laptop up on a mini bike š¤£š¤£š¤£ and upgrade it. Jk.. I can only do that on the BST. Not sure about that area
I wonder if this is achievable with femtofox's, a good amount of solar, an 8.5 db antenna. Longest link is 116 miles. I have a need and want
Haha
Its taken me a while to find this gap. Mind the gap
Should a really stable remote pi install be using a readonly sd or ssd?
I don't think so. Meshtasticd will want to write to it's internal database for node list, etc. I would just get a high endurance sd card. I've had some rpi running for many years on sd cards that are not read only, but don't write very much to the sd card, and not had sd card failures using sandisk or samsung.
Oh btw my Utah desert bound wip
nice!
No. There's no point if it can be read.. Security wise you should design the node to store as little as possible and so if someone catches it or screws with it they get nothing. If your a troll like me. Tack on the ability to monitor air tags and ultra wideband and get real time position updates of the thief LMFAO
Love it. Where's it going?
Or do you mean for use out there?
I'll add your site survey to the collection I have if you got a good location so I can keep track of what's active and mesh the maps together
This is at the fake mars by Hanksville, but we could either use one of our local hills or get it onto Mt Ellen if sdarc wants to work with us. TBD
What is the tallest peak that can see the wasatch front from its per view?
@real cedar the mt ellen repeater might see a larger part of the region, but it's not going to reach. It's not even actually mt ellen. It's a lesser peak to the south at 38.072586, -110.795150
I'll run it when I get home and check it out
This all the nodes i got so far
Hehehe you'll get another soon. Aurora Neptunia is hidden DEEP on antelope islane and is Router_Late to facilitate Ogden and box elder county communication if it can. Its more of a torture test. Can a RAK with all the lowest power usage settings enabled.. Survive with a 7.5 watt solar panel in a tiny tiny grey box on some peak
Its literally the lowest power node I've ever made. A solid 3 dB PCB antenna and a small filter. Its credit card sized. Its basically a chip, a tiny solar panel, a filter and the rest of the space is a flat lifepo4 cell
And it's on a cliff you have to sketchily climb to
@pallid isle
Oh are you on channel 51. A lot of those names are channel 20 names
And channel 20 is dog crap
Well I don't know if I did channel 51 right because it like the only one so I think I put in the key wrong š
I just amazed how my nodes I picked up
When we say "channel 51", we actually mean frequency slot 51. You don't need a key to change your frequency slot, it should be in the Lora settings.
Ogden and Layton and Clearfield areas can now reliably reach salt lake
FFFF is NPR
C6N is Aurora Neptunia on Antelope Island
H and C are nodes that are in my apartment complex that serve as a catch all because they are partially blocked. They see each other. They only see 50% of the valley each and it's entirely seperate sadly
But this is my attempt to get connectivity up north until I can put something on Ben Lomond and Rudys Ridge HP
Sorry new to meshtastic lol
Ok so I'm going to ask some stupid questions here
So I changed the frequency slot and tried it on the default room and the message don't go through. What I'm doing wrong
Well I know one thing. My RAKs panels don't like the lower sun angle and don't charge until about 12 until about 6 to 7 pm. I guess I should consider that normal. I can tell they are still shaded unfortunately. They got some sun though
We will see how they do by the end of the day.
They each have 2 to 3 days of reserve so š¤š¤š”
Each panel can supply 10459 mah in 8 hours after 75% sunlight testing so they should pop up? Hopefully?
Might be time for a different set of enclosures. The larger ones. Sadly. They aren't coming up yet. Neptunia will probably sink and die. Oh well. For the science.
If you guys have a known good router config or know how to config a rak4361 for the lowest power. Could you place it here? I need a known good router config. Basically I'm looking for all the intervals to reasonably balance locationing (necessary to know if stolen) and intervals to send stuff at and if neighbor info and all that stuff should be on or off
@vital hemlock did the wiki's ever detail this?
Have enough people flashed 2.6 to know if it makes much difference with routing?
Not many solar rigs run gps. Itās a huge power hog, unless you turn it way down. Which, Iām sure you have.
The solar channel here on discord used to have some pins with a lot of good info. If they donāt now, itās still a safe place to ask. Lots of smart folks in there. šŖ
Only on direct messaging I believe. And we could run some tests.
Thanks I'll go take a look around. And actually I don't have it turned super down. I have smart broadcast on using the auto algo with the Zoe m8 module
And smart actually works. It doesn't turn gps on unless it has moved
Which is hella nice. It will get a fresh time stamp and location every 5 hours though
Utah definitely a decent candidate now. We have a network with more than 3 hops from one end to the other though I do intend to try and fix this in the summer.
Also I ran some calculations. Ig you wanted to cover the entire area from the goshutes through the raft river through northern Utah to log back down through the entire front and over to mt Ellen and down to moab (currently an island) you could still do with with strategically placed mointain top routers on medium fast without a hop limit. Would be beyond epic tbch. But I ran the calc mostly to see if we in the valley could change to a higher config. Short fast is doable in just the front. Medium fast still gives you long range benefits. And I ran some testing on a repeater I have on channel 37 that sits up on Adams Creek trail rn high up. I was able to see it in toele no issue with either setting but short fast gets a little spottier. It's fast though. Like 1.5kbps fast instead of 0.2 š
Many of us had a good experience on MediumFast, just not everyone.
I am glad to put a radio back on MediumFast if we need to test more
Also, made this. It's hopefully going to replace my 1262 tbeam on my roof. I'm tired of climbing up there to updated it every six months š„
https://www.printables.com/model/1244203-meshadv-pi-hat-enclosure
**yes, the heat sink on the SD card is a requirement and not at all for shits and giggles /s
I had a bunch of rpi 2b heatsinks laying around and slapped them on everything, even the buttom side of the board š
I'm currently testing the Alfa 915 on it, then I'm going to try a 9dbi Rokland. I don't expect good performance out of the Rokland in this scenario, but I'm going to test it anyway. The amazon specials we were getting is what my roof node has now. It says it's a 5dbi, and honestly it's tuned really well. If I'm not happy with either of these I'll move this back to SMA from Type-N and use what's up there now.
Anybody doing some serious testing against NPR?
Itās responsive for me, but itās been at 24.96% channel util. today.
I've got a script sending out a message every 5 minutes today. I don't imagine that would give high load though.
Holler if I should tone that down a bit though and I'd be happy to.
I think for now I'm gonna try to get really solid additional routers built for long fast then test medium fast only cuz I need to get one out of desert haha. It can see npr and onaqui and every other useful peak
The only trouble is getting there really
Then once deseret and Ben Lomond are also up. I genuinely believe medium fast should cover mostly the same area no problem
At least so the calculations from the lora setup of medium fast would actually lead me to believe
I think it would take much more than that to be pegged at almost 25%. I wonāt be worried unless this goes on for days, but 25% channel util. is supposed to be when it will start dropping packets.
I'll stop my script by night fall for sure.
@real cedar I'm wanting to get one up on Lone Peak this spring/summer. I'd love tips on exact placement though. Obviously the very peak would be ideal but I worry that it would just get stolen. That would get Utah County connected into LF51.
And this talk about those little power efficient linux nodes maybe with LTE all running on solar sounds super cool.
Uhhh was that you last night. All 3 of my nodes registered over 37% utilization for a brief moment last night. Let me get the deets
Mint Mobile is your friend where T-Mobile exists. Some nodes it may not be worth it. Remember LTE also isn't low power. The way LTE works is actually interesting. You have a reflection plane. Your transmitter modulates a signal back onto its antenna at extremely low power that is slaved to the carrier. But it doesn't actually transmit to reach the tower. In fact it doesn't transmit at all. It controls the reflected signal from the phone Feedline.. Its rather clever. So instead of 4 watts to talk. You simply need a couple hundred milliwatts in a circuit. Its why you always hear "the tower controls your recieve strength.." Its cuz it actually does. It controls the power broadcast out and thus the total expected reflected power. Your transmitter modifies the carrier wave by modulating a wave with nulls and peaks that destructively or constructively interfere with the reflected signal
Cool huh?
Its something very little people dig deep enough into LTEs actual hardware to figure out
I found a better mountain
Will be headed out there with a node in the next couple months. I promise ECTO.. I'll get you a reliable connection from Ogden š
This should also connect a few friends in the raft river range.
No, I just started it at about 9am this morning and killed it just barely.
Someone went wild trying to transmit to DS NSL but It was dagrons nodes haha
Its where all the dips in my battery were too. Impressive that it sustained it for at least a few minutes
@leaden crow or anyone else. How do we use a t1000e ebyte module on a pi to store gobs and gobs of metrics, store n forward messages etc. It's kinda what I want my home telemetry node to do. Feed all the telemetry to somewhere I can analyze it
Little interesting tidbit, with my script running my old man was able to pick up some messages from me (through NPR I'm assuming) out in the desert by the salt lake here: https://www.gaiagps.com/public/0rTVCAjrWP1AJbifE8mc0JPH/
Turn on debug logging in /etc/meshtasticd/config.yaml
It will capture everything. I don't remember when I reset mine last, but my current /var/log/meshtasticd.json is 187 MB.
Yup the farther you get out that way the more and more you can see NPR reliably. Same going north or northwest. Kinda why I want to get an extra solid base station out there. Not only to relay all of northern Utah but you know cuz I'm in the desert a lot and I'd like to not be signalless
Its not a friendly place for humans at all
How do I map neighborinfo after that though?
And 187mb. Pfft no issue. I'll attach a 1 tb ssd to that node. Space problems be damned
I'm not sure, have not spent that much time looking at it. All the telemetry data is plain text, but any payload you'd have to get your keys and decrypt (assuming they are messages you have the key for).
Got ya. I'd love to build a map of our routing. The node space is just too poorly placed to do that. But man a network wide map of how the network is actually routing. I think that would be epic. I've got the space. I've got racks of servers in my closet. I definitely have the connection for it. I'm open to hosting something but only if bullet proof and easy to maintain
I unfortunately did not have a chance to try different antennas at work this week. This happened in my office on Mon.
Big fucking ooof
Happy to report everythings still transmitting. C6C has this LOS and a 6db antenna and filter. It sees everyone with 1 to 2 hops max
What happened?
Not sure if I should talk about it because liability is in question and our insurance company thinks lawsuit may be possible. Fortunately 99% of my work is pushed to repos, stored in cloud, etc. but my MBP that was floating around in my swimming pool of an office is still a huge inconvenience.
Don't do anything that would hold you liable. I won't ask further
When it's set and done maybe share?
The liability is between two third parties. Not me or the company I work for. I'm guessing insurance will work it out and I can share. Even being a huge pain in my ass, it's still kind of funny.
I mean, I always laugh when stuff like that happens. Some people get mad at me for it, but I say, when shit's happening, why make it worse? Might as well laugh at it.
Earlier I didn't see the play button and thought it was a picture. Now I'm absolutely rolling with laughter
I'm sorry dude
But the way that door seal worked
Whoever did that seal
Give him a raise
š¤£ššš fucking wall gave way before the door
I can't even
Engineer approves of the door š
And the office printer goes
https://www.youtube.com/watch?v=NUYvbT6vTPs
Cat Vibing To Ievan Polkka - Street Drummer With Vibing Cat Meme (full video) New Cat Meme Cat Vibing To Ievan Polkka (Official Video HD) Cat Vibing To Music | Cat Vibing Meme
instagram : @bilalgoregen
Bilal Gƶregen iƧeriklerinden ilk siz haberdar olmak istiyorsanız http://bit.ly/3r5bfbj linke tıklayarak resmi YouTube kanalına abone olmay...
@vital hemlock
I understand y'all tried GPS before. Try the ublox chip versions and turn on smart broadcast to 0 and 0.. Broadcast location every 10800 and acquired every 5400. I've gotten the tiny rak solar modules to stay alive no issue in the cold with that if south facing with the panel. It turns off the GPS. Most of the time and if in 5400 the smart function if the m8 doesn't detect movement it won't bother at the 5400 mark. At least from what I am finding on 2.5.20 on rak4361s with the 19007 baseboard and Zoe m8q sensor block
YMMV
It drains the bat for like 2 minutes and then nothing for many many many hours
Sometimes a whole two days before my power metrics module detects GPS turning on if stationary with a good view of the sky
The Mediatek GPS units are dog shit. They might be cheap but they have none of the nice features of the m8. Also over serial you can reprogram the m8q if you know what your doing. I have a TON of experience with it cuz we used them on animal trackers
If the devs need a guy who knows the m8qs and GPS well (I designed a lot of ground stations world wide military and civilian). I might be able to see what can he tweaked? Idk time willing of course
What are you using for your linux native setups? I want to setup a linux native node. USB would be preferable but a pi hat would work too
We had some of these made for us
https://www.etsy.com/listing/1849074257/back-in-stock-meshadv-pi-hat-v11-fully
Loving them
I have mine on a Pi Zero W 2 and a RPi 3b
If you need lower power use the femtofox community hardware repo and run your own bom out to osh or jlc
I love what opensource country is doing so definitely donate if you can
I got rods boys
Big ones š. Maybe oversized a smidge š¤
I'm testing with my 31.5" 9dbi
It's doing "alright"
It's tuned really well, according to the nanoVNA, but it has worse SNR than my roof node. I'm not an expert, but I assume that means it's receiving unwanted signals "better". I may need a filter.
Right now the SNR from NPR to my roof is 2.5dB
And the SNR from NPR to my other test antenna is -15.25dB.
Not loving that
So that is simply a matter of its lobes
You need a radiation diagramn and then see what its beam width is and where its main lobes are at
then orient it to make sure its aligned with the primary stuff you want to send and receive from
@white gyro hey bud so I ran some rudimentary line of sights. If you want to connect moab to wasatch
Loafer, Bull Mountain, Mt Ellen. I would suggest 8.5 dB well matched 20-30 inch half waves. haha
but it could be done and I definitely think eventually such a link would be awesome
Is the only way to get a radiation diagram from a spec document?
Typically yes unless you have a silent RF env or an anechoic chamber
its only an approximation at best though
@junior hill do you all have data sheets for the larger outdoor antennas you sell? Thanks!
Yeah that's gonna be the easiest way. I mean I know some places that do testing but I dont want to pay. Tell you what though. Find your angle to see the other antenna. Angle your antenna so it's center can draw a perpendicular line straight to the other and odds are you probably would be on the center lobe haha
What you want is the horizontal and vertical radiation maps. Entenna height and gain do correlate to a degree, so you do kind of want to optimize these things.
I'm guessing it's extremely similar to this @vital hemlock
I think 116mi will likely be unreliable unless those hops are from less noisy mountains.
For instance, my skyline drive node (on Maryās Nipple) can RX traffic from SLC mesh but never get TX acks. (NPR should be in LOS except itās too noisy a station, onaqui should also but my placement was on the wrong side of summit)
Up to 80 miles has worked for me in backcountry, but we see failures to link routers with NPR with less than 30mi gaps
I already have a node linking over Manti ridgeline, bull mountains still has some interspersed snow - was scoping it this weekend. Should be g2g soon
We need to link my Maryās node with the SLC mesh.
(39.0362178, -111.5777100)
Filters might be our friend here.. I know @leaden crow is a huge fan of 1 watt modules cuz they are legit good but I don't think we need more power or more gain. What we need is a good filter.
RAKs might only put out 23 dBm and he's right that this is weak for TX return so ideally you want both sides to be a 1 watt
But here's my issue. Where do I find a good 1 watt cavity filter that beats this value
https://gpio.com/products/915-mhz-ism-band-pass-bandpass-filter-with-26-mhz-bandwidth
That's really what it may be down to. If we filter and drop the noise floor (this actually raises the base Noise Temperature since we have added a component.) We should be able to make much longer jumps. I'm trying to find cheaper economical filters but these these are things that aren't usually super cheap
PRODUCT OVERVIEW This is a small, cost-effective band pass filterĀ centred at 915 MHz with a typical insertion loss of 2.5 dB in the pass band 902-928 MHz. This filter has excellent rejection specifications. Rejection at 880 MHz for instance, is typically greater than 40 dB. The filter also provides power handling of up
Tell you guys what I'll dig through some massive parts catalogs I have access to as I work in this stuff. I'll see what I can find. I think a filter under 40 bucks doesn't hurt And if everyone can filter and it's highly efficient (high Q low insertion loss).. Then we could filter and do a far more efficient job at emplicification when the signal reaches our xmitters
With high gain and a filter.. I think we could have a lot cleaner of a link
If only there existed a board that had this on it https://www.digikey.com/en/products/detail/qualcomm-rf-front-end-rffe-filters/B39921B2672P810/10648486
Order today, ships today. B39921B2672P810 ā 915MHz Frequency General Purpose RF SAW Filter (Surface Acoustic Wave) 1.1dB 26MHz Bandwidth 5-SMD, No Lead from Qualcomm (RF front-end (RFFE) filters). Pricing and Availability on millions of electronic components from Digi-Key Electronics.
https://www.minicircuits.com/WebStore/dashboard.html?model=ZX75BP-915-S%2B
If you want a pro build. Mini circuits makes an awesome cavity filter
If your feeling particularly adventurous.. Run a Rogers or Isola PCB.. https://www.minicircuits.com/WebStore/dashboard.html?model=CBP-915C%2B
Put that on there and have it have the female ipex connector š and make a sandwich resonator that fits over the rak 4361 and then output the signal through an ipex male
I'm thinking of doing that and reporting back.. I could get some RO4003C
I have two Asacom's and two gpios shipped and on the way. I tried these first cuz they are cheaper but I know for a fact the mini circuits will do a great job.
I really might make my own well filtered version of the femtofox with a RPI zero 2w instead and eat the power needs. It might be the best way to integrate all that I want which on the mountain I'd like soil moisture, barometer, temperature, earthquake sensor and UV. I'd like it to actually be believable as a science sensor. Plus I think it would be extremely cool to pass that metric data over the mesh. At least I'd find that useful and would probably publish it
https://www.mouser.com/ProductDetail/Taoglas/DBP.915.V.A.30?qs=MLItCLRbWsyKal3qF5ESLA%3D%3D
Im going to spin some filter PCBs and see if I can make us some cheap 5M filters centered on 915
Notes if anyone wants to sanity check. Its in my private git repo for stuff
@leaden crow I just put the gpio labs cavity filter on my ceres node. I can jump to onaqui directly. The filtering makes a HUGE difference. I think LTE is interfering with our noise floor. Its using a 6 dbi flexiwhip.. the gain makes up for the filters insertion loss
TR: C6H -> 38fb -1.2 dB 38fb -> C6H -4.27 dB
(H has one too).. even more reason for me to make a filter and ULNA combo.
Best damn jumps I've gotten ever. Filter worth it
Guess Iām ordering more stuff.
I know the feeling. At least you dont need to order a lot like I did. Im pretty much sorting out what works. Then ill probably spin a PCB and if I can get that pretty darn solid Ill offer it out to people
I do think the narrower band filter idea and committing to channel 51 is actually quite smart especially if we stay in the center of the band like we are. It limits our frequency freedom but we can make tight filters for it. I think for an lna ill need a chebyshev filter so that will take me some time to model to get the transformation correct
@vital hemlock any idea on a beta release for 2.6. Its jam packed with so many goodies and im eager for a stabler build. I literally cant wait
I've been running it without issue on a handful of radios
It's been solid, no reboots or weirdness. I want to run it on my linux nodes, but havent yet
In all sweaty from doing this
Itās going to replace my roof node, so I donāt have to get up there every damn time. Ran a 15ā usb cable down to my office on the second floor
So itās that zero w2 + e22 module, with the rokland 9dbi. Iād love to buy a filter once you guys figure out whatās working the best.
I donāt have the expertise with antennas and radio that you all do.
Thats awesome and dont worry about expertise. Ill bring mine to the table probono .. I want the network to be large, reliable and painless is my desire.
Thats gonna be a great node
tell me you have that one lightning arrested?
A good 14 gauge from the arrestor to ground should do a great job at discipating any currents from a lightning strike up to 5 miles away (indirect strikes within 5 miles can actually burn out transmitters. Or at least have before)
They are cheap.. easy to do and run lines for.. cheap insurance
Also good if you do get struck. can tell insurance oh hey yeah this was lightning arrested. pay up
I never had an issue with the other node, but with this antenna I definitely was thinking about it. Iāll get on that this week. Anything else to consider I might have overlooked?
I assume the arrester threads onto the outer barrel?
I mean.. 1 watt module so you cant use the gpio labs filter and have to use the expensive acasom but if you dont need nor want to filter.. just a lightning arrestor will do and yeah mount the node as close as you can and your gooch
Personally if it was my main home node that did all my routing to the net I would filter it and make it a higher end box and store lots of metrics and make a map and stuff. But its a thing you can do without on a non router box
It certainly helps massively. it increases receive sensitivity a lot
You think a rando-Amazon arrester would be fine?
rokland is most reputable but it could be if the numbers are to be believed. sad thing is its not water proof
Iāll wrap it with rubber tape and heat shrink the whole thing. Thatās what I did on this one
yeah honestly it looks a lot like the ALFA ALR6
its probably a clone
heatshrink it with some good shrink and your gooch
Iāll check roklands arrester, no reason to cheap out
they sell bare copper. No reason to be that way about it .. get what cheap outdoor rated solid core wire 8-14 ga
run it down in whatever color your house is so its not seen
Preferablly to whereever your panel and meter ground themselves. then everythings common
Iām glad to support Rokland over rando-amazon, thatās what I mean.
ahh yeah.. they are kind enough to bypass some restrictions. sadly they dont stock everything
I needed a lot of the newer modules from rak and I had to eat the darn VAT
if theres anything I wish they would have the whole catelog of its the io and sensor blocks
Ordered. Iāll see what I can find copper wise to ground it once it arrives.
Have to hope Iām good until then š
yeah thats definitely not going to be a regreted decision when the occasional bad storm does run through and you get wonder if its safe. Its my concern with mtn top nodes but im opting for a different route with those. As insulated as possible
that way lightning just goes whereever there is least resistance
You want to know how to start a lithium battery fire? Get struck by lightning
Thatās the same side of my house as my power meter. Should I run the copper over there and secure it to the shared ground, or just go to the earth itself? š¤
oh yeah thats the easy way just make sure neutrals not in contact with it and your gooch. At the hacker space I just drove a 6 foot pure copper rod with it solder welded to the wire into the ground
Thats why despite its horrible location rf wise.. its got such a good ground plane and can hear weak things
I need to put a filter on that lol
Well, for now I still have the tbeam up there as well. If anyone wants to test medium fast I can switch that one over
That's at the other end of the house
Ive lost sight of lake mtn but still have a direct hop from clearfield to onaqui.. which is weird.. so I cant see the south area
LOL is that a sector antenna
No, it's a 5dbi hiding in there
ohhh nice
This is the one that's been up for years, no lightning yet š¤
99.9999999999% of the time it should be struck. I just like cheap insurance to beat insurance over the head with
It's been so reliable, honestly. It'll be a shame to take it down, but I dont need it up there anymore unless we are going to try mediumfast as well. No rush to remove it though
Oh well t beams might be reliable but they arent as good as a native linux client. Now you can help build a map and we can see all the weak points. Ideally thats what I want to do rn. Gather a lot of data. Improve whats here with hacker space friends and then put up routers in good spots
The routers should be able to hop to one another but cover a very distinct barely overlapping cell me thinks
Helper routers can be router_late or client. If theres a particularly good node in the valley with great SNR to everything else in the valley down low.. it being a client or store n forward router_late would greatly improve the robustness of routing
It's running neighbor info and s&f, I have to check if router_late allows wifi
My idea more or less is to try to at least give the router part a ton of stucture that way no matter what users in the valley choose.. i.e client, client mute and hidden, tak and tracker... they all can feed into the same priority nodes.
router doesnt allow wifi for bluetooth I believe
I currently am just thrashing my FRAM lol
The easiest stoire n forward might be a rak with an io block that lets it do so. Then just router late and have it store 40k records or something
I need to read more into it but it will be a key feature of all the router nodes I put up
Ill eat the power budget
Learn about and compare device roles such as Client, Repeater, and Router as well as other Device settings.
router late has wifi
4x6500 mah 21700s
oh sweet that will be great..
then just feed it to a server
we have a LOT of issues in toele
im trying to help two space members out but they only got one chirp out ever few days to NPR
I can send to them and hear them but they cant hear me ack so they dont know their message sent through me
its the weirdest link too
cuts straight past the southern tip of antelope island.. straight over kesseler ridge into erda and toele haha
just like onaqui
I'm going to test putting mine in router_late for a few days, then I'll switch it back
In theory it harms nothing. It will if seen by a node struggle bussing... route as a last ditch effort and if it cant then when it fails your device says it couldnt make it
but if somethings super advantageous like yours or mine.. then once alll the clients try it can try and usually succeed
Client would move it ahead and likely it would do more routing
but IDK how to store n forward in client
otherwise I would
s&f is a manually activated request, just so I know we're on teh same page
when you come back into the mesh you can manually ask what you've missed
how do I make it do so?
and you cannot do it on the public channel, forbidden
Oh lol im more or less doing it for dc801 chat which is a private channel
are you using apple or android?
android
can you see my bp_rpi?
yes
On apple itās here
I dont see it on my android device
I use apple for my primary handheld, and android because it can connect to IPs for the roof
yeah ill have to set it up on the femto fox
darn.. So I would love a way to store n forward everyones stuff but I guess that would need to be very algorithmic huh
tbeam was the first device to support s&f, I didnt think there was a way to make the rak do it
There is
nice
you need the FRAM module
Once you have that.. enable it in the firmware. branch off a git and make a build for your device
once thats enabled just enable store n forward and it lives in FRAM
Once you have it worked and you request your messages the s&f node does a direct message with the contents and a summary of count
that's how you know it's working
for Android it is required to send it a direct message containing the text "SF" (without quotes). The
GOT IT
SWEET
ITS WORKING ON A RAKS FRAM
phahaha
so much for only tbeams

FYI the entire tutorial for making that work is on raks website. they tell you how to make it work in firmware
Okay this is beautiful. I need to teach this at the space. Get everyone to have their main node store n forward
then just SF it
remotely with whatever device you got
the BBS is honestly the best option
Especially because you can chain them to have them update each other
yeah ... your right. So where would we setup an autonomous BBS node? do we have a really good place that can see us all? or how would this work
That's the nice part, they dont have to be centralized. I have one, it can update someone else's if they wish
It sees your node come online and says "hey, you have mail, nerd"
That would be epic. Got links and a tutorial?
bookmarked
Thank you all for all of the feedback, bug reports, and feature requests! Many of those have made it into this latest update and is what we're we'll go over in this video.
Check out the project on GitHub:
https://github.com/TheCommsChannel/TC2-BBS-mesh
Looking for Meshtastic Gear?
https://store.rokland.com/p...
If I have a large array of nodes with a common telemetry channel can I have this link over lora to pull from each of them as bbs source node or no?
I'm going to update mine and put it on the bp_rpi
like I want a big net for it is all
I need to get my femptofox up in the window
Well, what's nice is there are shortcut commands you can pass, and they can just post to a bulletin board
you could read that programmatically, I believe
Watch that last vid, when you get chance. He made quick commands, they are pretty sweet
I will definitely do so. Honestly using ATAK and stuff has shown me that applications are more robust than the mesh app itself
Which is a testament to the network
but not so much the app haha
sorry
I love the app dont get me wrong but man was atak so much more reliable. I guess its had more devleopment
hence why I am going to feed drone telemetry over it in a small protobuff
I do want to make an ADBS B app too over lora. I think it would be cool for me to ping a node with "FR" (flight radar) and have it send a bunch of protobuffs that detail all that info for atak to display
I am getting creative but you can do a lot with very very little
also supports js8call
Yes
I would love someone to relay that into a public channel people can observe
it would be even cooler if it say 801M_relay: On behalf of "someones message" with your call sign or whatever but IDK the legality of that
it would be beyond epic to have a HF world bulletin available over the mesh ahah
I just dont have the money for an ICOM
moved back to client from router_late, the late confirmation of direct chat receipts feels sluggish and I dont like it š
Got the bbs working right on the bp_rpi now. Feel free to screw with it
FYI, router_late is bad š
That's by design but yes I understand
Its there to be a last ditch effort. I think I mentioned client was better
The way I might go is once the femtofox is up. It can host a BBS mesh.. Then I'll just do client for almost everything that will route to the greater network and hidden for everything else that doesn't need to talk
I think it will clean up my traffic alot
@vital hemlock re-reading the blog made me think
https://meshtastic.org/blog/choosing-the-right-device-role/
Can someone give me a list and rough idea of where all the router and repeater nodes are at. I think we need to optimize this spring more than anything
Here's a though. Clients do rebroadcast after routers. What we need to do is the ideal route optimization problem
We want each router to see each other but we want them to see a different subset of users and not talk over one another. Similarly we don't want a repeater too close to another router because it will rebroadcast ahead of the clients so if anything router clustering is preferable to router-repeater. Cuz you won't drown out the clients as much. You'll cause extra hops
So NPR. Definitely a router. Deseret definitely a router. Because onaqui is lower if we wanted to keep it.. Client would be smarter believe it or not. It sees lake mountain which taking another look.. Router. Know why? You have POTM turn into a client. Potm can still help route to npr for those who don't see NPR but NPR handling all of the bowls traffic is a lot less congestive.. I bet the routing gets interesting down there rn..
Now say someone puts something up on one of the ridgelines to say grab the avenues.. Client.. (Rudys ridge).. Then.. For bountiful.. That one should be a router. Cuz it would mostly grab the north and parts of toele and not really Interfere with NPR. If I get lead mountain up.. Bountiful would see it just fine with filtering. And lead mpuntain could facilitate the hop to bountiful or NPR whilst handling the traffic in Ogden. Does the hub spoke make sense?
It would be acceptable to make Ben Lomond a router at the cost of some congestion if I put one up there because that would easily grab the whole cache valley haha
I'll draw a map ish in minute but the thought dawned on me when I was looking at FAA sectionals
Okay check it. This is the greater map view but Ill zoom in. Tiny square are basic client mutes and clients.. nothing special. Triangles are strategic clients.. they are positioned amazingly and you would initally think router. NO lol.. let them do the job as a client. If the routers trully cant see each other let the helper client do the work.
King general symbols are full routers
Red dash are router to router links.
Purple dashes are router to strategic client or strategic client to strategic client
Now ill zoom in and show how each has its unique cell
This would probably be the most optimized you could get it
We should save this LMAO but this is me spitballing a plantogether for when I plan to put more than a few routers out there. They all have a distinct cell. Sure theres a bit of overlap but take a strategic client in the overlap.. say near a high point and put it up there and it can help route in between if say it really is a more reliable route. Especially with 2.6 coming for DMs but I imagine the routing will get tweaks eventually too.. having very optimal placement would make this usable and not break it
Turn the lower routers that are routers or repeaters now into really good clients.. less high gain. say 4-6 dBi and a filters enough. Am I making sense?
I deep dived into how all the routing works to build this rough idea of how it would go. Especially if it got to the point where we were taking data and stats and saw the strategic clients actually doing what I think they might do.. we could broadcast on long fast who we think should be a client.. and say the rest client mute. I think that would be huge.. not saying people have to do it cuz we ask but it would be "optimal"
I for one appreciate all the planning, this looks really good.
Yeah no problem. I have plans to utilize it. And you know if someone doesn't want to maintain a node but wants to contribute. Honestly my most likely mountain nodes are going to he ebyte modules or raks. If ebyte it will either be adv hat RPI zero 2w or femtofox lol
So I'll always take hardware for the cause or old solar panels etc. Anything I don't have to put extra money into I'll make look pro haha
Here's my new trace route to neptunia.. The db increased from negative to positive with the GPIO labs filter. The Rak modules are so much stronger. I bet this goes well with a tbeam since they don't do more than 24 dBm right?
I hate everything about this rig up
Lol don't do this at home
Testing is iterative, looks like progress to me š
I had it on the vna and sa but then decided to hook up the node for tests
It matches super well
Its got great compression points
I mean the rejection outside 902 is 47 db
Outside 928 its 39
Props mad man who owns that hobby radio page
He made a really good filter
Things like this make me smile. Someone cared to be honest about their work and they under promised over delivered
Lots of RF stuff is snake oil.. So I guess I'm jaded
So the best most efficient solar panels you can buy that are less than 2ft.. Is the 12x16 renify 50w.
Otherwise in testing I've found on a 8400 MAH pack.. The soshine 12W does just fine even today it charged under snow and rain
I will likely put the renogy either on Ben Lomond or lead mtn. Probably lead mountain. It's cold. Extremely remote.. Behind promontory point and even farther back than the real lead mountain.. With the sister peak higher than it and messix... You have to drive across the lucin cut off to stand a chance of getting there at all. The north side takes more than a day
High end:
https://a.co/d/6mRH6BI
Mid:
https://a.co/d/2DH9TFD
Works
https://a.co/d/7zo8dqY
Renogy 30 Watt monocrystalline solar panel is ideal for recharging small electronics while away from the Power Grid. This particular panel is small and lightweight, making it easy to store, reposition when needed. Featuring a junction box, connecting this panel to the charge controller or additio...
If it's gonna be really cold up where your thing is at I'd go mid. If it's just a node you want to put out somewhere.. Works.. Is definitely the tier.. And if it's cold and remote and your gonna need snow shoes and skis to service (personally I'd love this idk about the rest of you lol)... Go high end..
No regrets can be had if you can literally "fast charge" your batteries lol if you made your own controller.
Filter on Ceres node 2.7 miles away. Now fully in the multiple positive db up from - 13.5..
Confirms gain without filtering us useless. Gain with filtering us powerful
Mind you Umbriel doesn't have a filter. And I've got its panel antenna on the PCB faced in the general direction. So a whole 4.5dB more consistently is nice. Probably more on higher power nodes
I also did some tests and calculations. For the RAK4631 running Meshtastic LongFast (250 kHz BW, SF9, CR 4/7):
Nominal Sensitivity: -131 dBm (SX1262 w/a reputable antenna)
With my setup (Cleaned RF feed line): (factoring filter loss and antenna gain) is now in the range of -147 to -159 dB for long fast which dramatically increases how much I hear. Will see how this works out
Sweet, Loweās has 10ga stranded on sale. Iāll rattle can it the right color š
LOL that works. see if they just have the right jacket color though
ace usually has the same stuff in many colors
Good call. Iāll do that
With all my nodes filtered I can hear lake mountain again. I can see someone in saratoga springs LMAO. and I can see ghost and poly directly in toele. The problem is what I know is probably true. They cant hear me because they are unfiltered. Im surprised at the difference it makes when you have 10 and 8 and 6 db antennas with the filter. I highly recommend them now. I definitely want to see what happens if all routers get filters
and 6-10 dB antennas
everynode in visible north to south lines along i15 is one hop. everything that spreads is 2 max
My SNR and RSSI numbers are up a consistent 6-7 dB so thats quadruple the power and in this case its useful not just indiscriminate (non filtered)
I suspect theres an upper limit here when a signal gets too strong it will just be attenuated by the amplifiers 24 dBm P1db Compression point.
Do you have a final recommendation for which filter to buy? I think you were testing some different ones or something but havenāt followed the entire thread.
Yes so for 1 watt nodes you want the Acasom. Sadly its a bit expensive but its really really good. You could use it on less than 1 watt nodes but theres a better option for that https://acasom.com/products/915mhz-cavity-filter-for-helium-network-amplifier-filter-sma-type-high-out-band-rejection
For less than 1 watt.. 25 dB and lower I recommend the GPIO labs https://gpio.com/collections/filters/products/915-mhz-ism-band-pass-bandpass-filter-with-26-mhz-bandwidth
Lastly if you are as committed to FS 51 as me you can design one narrowly for the frequency slot. I will probably do this with the hope we dont move from this channel.
PRODUCT OVERVIEW This is a small, cost-effective band pass filter centred at 915 MHz with a typical insertion loss of 2.5 dB in the pass band 902-928 MHz. This filter has excellent rejection specifications. Rejection at 880 MHz for instance, is typically greater than 40 dB. The filter also provides power handling of
If you make your own I recommend either a dielectric for economics or a cavity saw from minicircuits
Filters dissipate excess power as heat. So be sure to mind the power limits
This is supposed to be the best, but not available, and tuned for LF20
https://shop.airframes.io/products/lora-915mhz-filter
I could never find a consensus if the acasom or callboost one is better. They cost about the same.
You could test. See if I made a filter for a certain channel id probably rev a PCB because id like them to be a max of 10 to 20 bucks and channel adjustable.. (change the rev)
Right now im staying broadband because I suspect we arent dealing with much Lora interchannel interference as much as LTE
#general message
Ahh ... well it depends on the width you get. Either way its a trade off. I figured id give people broadband first. Nobody really gave a nod that we are super committed to channel 51
cuz if we are.. then yes of course we should look for as narrow as we can get
I don't think we are committed to LF51, but it definitely has some traction at this point. The more people we get on-board, the more difficult it will be to switch again in the future. I still think we may be better off with either MS or MF on their default frequency slot. I don't have strong feelings about it, I'm just happy with how well it's working rn.
I think being off 20 keeps people who dont know much better away which is good. I think we should commit to 51 and if we do I can spin our own filter.. and spread the BOM. Im leaning towards these
https://www.taoglas.com/product/dbp-915-v-30-dielectric-band-pass-filter-915mhz-8-76-03-0mm-bandwidth-5mhz/
You dont need to kill off LTE
Enhance your wireless transmitters or receivers with the DBP.915.V.A.30 Dielectric Band Pass Filter from Taoglas. Designed to protect LNAs, it offers low insertion loss and high attenuation for improved receiver sensitivity.
you just need to drop it below the noise floor of the pass band
then it literally doesnt matter. so people are splitting hairs a bit filtering to the maximum degree
Im fairly certain if we drop LTE down 45 dB we dont really care about it that much
It's supposed to be. ROUTER_LATE is a "give way to everybody else" mandatory-rebroadcast role. This means that the packet should ultimately get through, but it will yield to other relaying nodes, which is where the delay comes from. The reason for that behaviour is to prevent the ROUTER_LATE role from interfering with the normal "routers then clients" SNR-based relaying logic.
So from what I saw in your docs.. really what Might be best is most people are clients.. seperate cells for routers.. but they can see each other.. and things that would otherwise help a lot being a router but are just not as high up like POTM.. being a client since it can see lake mtn? I was using the sight planner to try and make a map of the least routers needed for MF51
Clients only rebroadcast if they don't hear somebody else rebroadcasting first. ROUTER_LATE is a way to force a rebroadcast anyway, even in situations where a regular client would not.
Oh so its actually a good thing to have around if your strategically placed?
but say not as good as a router
Yes. The point is to use it in sites that are not optimal for use as a router, but that are required to relay around problematic terrain, patch black spots in coverage etc.
Bear in mind that it does contribute to overall channel utilisation, so you don't want too many of them. It's not intended as a general-use rooftop mode.
Okay so jovialis is in a window with a good view of Onaqui and NPR.. but cant see any of salt lake.. Ceres can see everything except onaqui. should I try making my ceres node a router late or jovialis?
For reference im an extremely long through for most people. My buddy in the avenues still cant get a message to me
I get DS's stuff
I'm not familiar with the geography in your area sorry. I just keep an eye on mentions of ROUTER_LATE to provide assistance with deploying it properly. I was the one who added the role, so IMO I have some responsibility to ensure we don't end up with another mess like the one that resulted in the removal of ROUTER_CLIENT.
Any value in adding a ground plane closer to the antenna, rather than just the ground 2 stories below?
Im up on a plateau the rest of the network is below me but far away and the routers are much above me and a bit farther
ceres with the best LOS overall
Ill test and find out haha
As a general statement, if your site must relay in order for packets to get from one area of your mesh to another, but it isn't somewhere that ought to be a router... then ROUTER_LATE is probably the right choice.
If there's a router on a better site that covers them, then don't use it. Let that router do the job.
Got it yeah I think I know which one to make that
Until I get something up on thurston cuz im sick of unreliability haha
wait
ROUTER_LATE
do they cause the same congestion problem with multiple of them
just want to clarify for me and the masses
Cuz I have 11 testing nodes up here and only 2 are any good for egress haha
So, that has a complicated answer.
-
They do contribute to overall congestion on the channel (because they rebroadcast everything that still has hops remaining). The degree to which they contribute is exactly equal to what that node would not rebroadcast (due to overhearing other rebroadcasts) if it were a regular client.
-
They don't contribute to congestion of the early contention window, where the regular routers live (i.e. adding a ROUTER_LATE will not interfere with other routers, beyond the overall higher channel util).
Oh thats perfect then Im going to try something weird.. and see if it works up here
It goes against my instincts but two router lates and see what happens. I am so sorry if channel utilization skyrockets
ill revert.. ill monitor
They won't bounce back & forth, if that's what you're worried about. It'll repeat any given packet once, and once only. It will not repeat it a second time if it arrives later from another source.
Okay so they actually might help
@broken turret thanks for clarifying. My node relays a lot of the traffic in the valley that canāt see our mountain top router momentarily. Itās been a client ever since we placed that node.
I thought Iād test the new mode, but mine is definitely better as a client.
PB are you in between NSL and HILL AFB
I just got a stronger hop through you than Dagron NSL office
Iām in Taylorsville/West Jordan on the south side of a large hill
They can. Depends what you're trying to solve š.
Near the top
LMFAO so far away from you
There are water tanks at the top of the hill, you can see them anywhere in the valley almost. They would be such a good spot. Sad
I literally route through onaqui more often than npr
I'm the northern tippiity tip
C6H sees onqui and nor amazingly well. C6C see everything but onaqui stupidly well but has less gain
I've got a big haul link and the rest of the nodes are stuck behind a 7 story apartment complex
I think NPR suffers from interference. So many radio towers near it.
I also suspect my nsl node does too. It's right next too a whole bunch of train service that apparently heavily uses the 900Mhz band.
I think we should BEEF up that node
ACASOM filter haha
1 watt module
cheaper femptofox rev
Working on it. I have a new antenna and cable. It's already 1 watt E22.
Oh thats YOUR node? dude
thats such an amazing spot
what a fucking solid youve done
I will buy a filter when I can make up my mind. The acasom has got some hate if you look at the history on #antennas
I need to redo my map now thanks to your new info. It changes the calculus a lot. I will have a full revision with the meshtastic site planner with all the radios overlaid soon
What is onaqui? A community? A cluster of nodes? A router? A single client?
It's a node on top of a mountain configured as router
What does it serve? If NPR can see everything except onaqui really well, then honestly... NPR is probably a better choice to have as router.
NPR is a router, they are both mountaintop nodes
A router
on a mountain top on my map
38FB
NPR is a router.. on nelson peak.
Remember I'm not familiar with your topography š. What does onaqui serve, that NPR does not already cover well?
Normally just toele a town only 30 miles from it but I have LOS cuz im high above others. so I can make the 75 mil jump
NPR serves basically all of salt lake
and part of toele
These are valleys in between very prominent ranges
Lots of blind spots in NPR covering the back side where toele is
Roger. So onaqui is one link in a long chain to serve part of toele. Does it do anything else? How much of NPR's coverage area does it overlap?
So I have something for you. This is a planned map. Dont take it as all these exist right now basically antyhing north of the lake doesnt but I was doing planning to cover all that area.. cuz we could with good filters and high gain antennas. If you look at the southern most triangle.. theres a cluster of routers there
https://discordapp.com/channels/867578229534359593/1197577977781821541/1355781488385462375
IMO, if onaqui can see NPR, make onaqui a ROUTER_LATE unless there's a compelling reason for why it needs to be a regular router, or its coverage area does not overlap much with NPR.
I think the plan is for onaqui to hop to fools peak, and then to mary's nipple (maybe one more hop in there I'm missing). Although we won't know until snow melt how many of those survived winter.
Ahh so this is like a router - switch kinda deal.. router handles all the routing but a router late acts a lot like a switch post router
Nope. Not a good analogy sorry.
Trying to figure out a good way to explain to people at our local hacker space. Got a solid analogy?
hmm is it more like airports? hub spoke? I guess im not understanding how the mesh appropriately strucutres itself. I know clients do it based on SNR (to an extent)
Maybe like crowd management. The routers are the PA system. They cover a lot of the crowd, and tell most people where they need to go. The individuals attendees are like clients. People ask their friends where to go, and someone pipes up with the info. Anybody who overhears shuts up, because the person who just spoke already said it. And ROUTER_LATE is the usher who actively goes to help any remaining unseated folks find their seats before the show starts.
LMAO how jank were some of them? I can try and salvage hardware if it got inundated with water
Got it so if we had a VERY strategically placed router like NPR but say Point of the Mountain.. its so much lower but on a ridge between two discontiguous sections of valley.. and can see a lot of both.. Router_Late is probably a good option as a okay clients couldnt route so heres a little help
POTM can see most of what NPR can see and a lot of what it cant
Not jank. I think they are the standard large RAK enclosures. Probably just batteries not making it. I think he used some molicel on the later ones he built, but onaquai is "5aH 3.2v pack from AMZ"
The key thing to remember is that the nodes all share the same channel. Only one can speak at a time, unless separated by terrain that blocks the signal. If you have two signals occupying the same place at the same time, a node at that spot has a high chance of receiving neither message.
All the roles are basically a way of prioritizing who gets to use that shared channel to repeat messages, and when, in order to ensure that messages are delivered to their intended destination as reliably as possible, and over the nearest geographical distance possible, without consuming excessive airtime.
Yeah that things a beast and good. what I meant by jank was ... water could get in. Normally I can recover that stuff in an oven and cleaning in the ultrasonic bath. So I was wondering if I should go get more solution ahha
Got it. I think I know how this changes my calculus
Clients can just do their thing. ANyone in the valley should just be one.
POTM being in my mind a really good spot to relay two discontiguous areas that NPR and it serve.. but NPR being better position.. Would be better as a router_late so it would talk after the clients havent figured out their CF downstairs so to speak hahaha and POTM can just rebroadcast and see what pops back
but wont broadcast ahead of NPR
the far better positioned router
Best to essentially order them in our topography (extremely prominent mountains and deep valley floors.. with glacial cut canyons... ) according to how their stratified by best LOS?
Set your best site in any given geographical area to ROUTER. That may not be the highest node (although it often is); rather it should be the one that provides the best coverage. If you have multiple routers serving the same area they will race a bit, and can prematurely consume hops. Depending on your network, you may be able to use the terrain to your advantage to have multiple routers without running into this problem, but it can be tricky.
For the sites that are necessary to cover what the really good router cannot, set them to ROUTER_LATE.
For dense urban areas where nodes can receive well, but nobody hears their transmissions, put a CLIENT nearby. That will relay anything that doesn't get as far as the router, and be quiet otherwise.
Everything else should be CLIENT_MUTE (or equivalent) if the mesh is busy, or CLIENT if there are a bunch of little holes around and the mesh isn't too dense.
My God, this is a huge help, man. I am going to go recalculate everything and figure that out because I think I got an even better plan now. Thank you so much for this explanation. It definitely is a lot better than the wiki, no offense to the people that put that together. I think it's great, but this explanation helped a lot.
I would love an updated blog post on the rolls and how to stratify them in different kinds of topography. In fact, I would love to contribute to such a thing if it's going to become a thing on the meshtastic documentation.
And of course anybody that doesn't read the docs will be a CLIENT out of the box. This is fine, and you can leave them that way. They'll automatically shut up if they hear rebroadcasts. Not as good as CLIENT_MUTE in a well covered area, but good enough for most purposes
I ended up leaving my mobile car node as client. It was useful for routing through the glacial canyons. I could park somewhere decent and my client mute could get to the valley
Mountains are a blessing and curse
Which modem preset are you folks running?
If it's still LONG_FAST, be very careful of how many mandatory-rebroadcast nodes you have. If the chutil gets too high, the network will melt badly. Recent firmware versions have done much to combat this, but it's still something to watch out for.
If moving the mesh to a faster preset is practical for you, it'll give you a lot more headroom, and also more scope to chuck extra rebroadcasters in places that they'll help, without making anything melt.
LONG_FAST
Most are 2.5.11-2.5.20 we have some 2.6.x
Yeah, keep a close eye on chutil then.
unfortunately our Mtn top router.. is a 2.3x š¦ actually a lot of them are
we are planning to move to MF in the summer
For a router, this isn't actually a huge problem - the behaviour of that role hasn't changed much. Although I'd recommend updating it when you get the chance, so you can use the new public key remote admin.
The problems with traffic that come from the old versions, happen when you have a lot of them. They are excessively chatty.
Yeah thats actually a great feature when you admin the router as you can admin it from anywhere
Great plan š
We use SF here on our mesh. Decided to move once, and pick a fast enough mode that we wouldn't need to do it a second time.
Ahhhhh.. alright that pretty much confirms. Run the calculus for the minimally chatty 2.6.x network in our topography...
The issue is we have a lot of long tall to tall jumps that need to be made
I fear SF may not make it
People here also told me it was tried once and went poorly
Our longest permanent SF link is 254km, and makes it easily. Basic Omni on each end.
Whether it'll work on your long links will depend on what's in the way, and how much noise you have to deal with.
If you have clear RF (not visual) line of sight, the path isn't full of water or terrain, and the noise isn't giving you headaches... then it'll go as far as the horizon will let you.
@leaden crow @vital hemlock SF test in the summer?
Instead of MF?
Full of mountains, huge salt flats, salt water, and lots of marsh lands and forests.. with a desert west of it
Ill run the math on the site planner and see if it can be done
If it's been tried and went badly, then there are likely some of the factors I mentioned above in play. Can fix that either with a slower preset (e.g. MF), or a denser mesh.
The difference in budget between LF and SF is 10dB.
Oh, sorry, missed a long discussion here. Iāll read the backlog š
If its 10 dB then thats not much. It might be that we all need filters on our Raks and our mountain top routers.. not even stupidly amazing ones.. just some filtration.. and we could maybe give it another go
If there's lots of evaporation, that won't help. I know the Hawaii folks have issues with that which precludes long SF links.
This has already happened. LF20 in our locale is useless. Hundreds of nodes, who knows how many on antiquated firmware. Dozen+ badly placed routers we can't get the owners to fix.
We tried Medium Fast last year and could quite get the coverage. This discord group has been on Long Fast frequency slot 51 since then.
Our valleys are notorious for inversion
Ouch! Good call on the frequency move š
lets strategically layout really good routers. Really good router lates and then give SF a go and if SF doesnt work.. MF
yeah moved to slot 51.. keeps Helium out of our channel too
theres a big helium net in Utah and they keep getting in trouble for big transmitters and 8W xmissions LMAO
Why the heck do they need 8W on LoRa?!
well its a crypto mining scam
the more nodes it sees the more crypto you get
FCC laws be damned for those guys
Overpowered TX doesn't let you see any further.
If you can crypto it.. people will break the laws for it.. its pretty much that simple hhaha
it certainly does not
its just rude
The Helium bros
https://ksltv.com/local-news/why-are-antennas-popping-up-all-over-the-foothills-salt-lake-city-seeks-to-solve-mystery/516749/
Better antenna, a good site, and maybe a LNA if the noise floor permits, will let you hear more. And proper filtering.
LMFAO
Right on the trail too
like didnt even TRY to hide it
I found one recently.. Recorded its loc.. I easily was measuring 37 dB 15 feet away. It was concerning
thats where a really good filter is going to help a lot.. its with an LNA. I thought about spinning an expensive lora node with these additions on a PCB but like we are talking me cobblign together a 120 dollar amp
I need to drive somewhere now - replies will be delayed for a bit.
no worries. thanks for all the information. its been a big help
You can get an ebyte 1W module that includes an LNA, if you're keen š
No worries. Good luck with the network planning!
Is it bias tee or does it DC block
cuz that will ruin filters if it doesnt block DC
No idea sorry. I've not needed to implement one. If it has a DC bias though, that can usually be switched off.
If for whatever reason there's DC on the line that you can't switch off, just put a block in the way to get rid of it.
LNA can help compensate for the signal loss through whatever you're using to get rid of the DC.
yeah I probably will do that. Guess i got some work cut out for me on filtration
I might actually try to make a good one for up to 1 watt with a DC block for all of us
We have at least 6 nodes based on the E22 (4 I have built). Probably many more this year.
you know whats problematic. Without filtration and high gain.. all they are doing is amplifiying the noise
š
they def can talk further though
Yep, that is probably the next easy problem to solve.
I say we try the wide band available to us cheap and easy from whomever that spans 902-928 or the GPIO labs on the raks and see how much that helps. Only worth going to the further expense if its still a problem?
E22900M30S is the one I'm thinking of. We have some custom bridges on our mesh that use it - has proven to perform very well.
Yeah.. those plus a decent filter.. and a 8.5 dbi rokland on our Mtn top routers and I think we are in Business
Careful not to use too high a gain on the high sites. Don't want to concentrate the signal over the heads of the folks you're trying to serve. 8.5dBi is pretty high.
#1197577977781821541 message
only problem is they need a rpizero2w and those bad boys are power hungry. Bright side is the femtofox has the build but its a bit expensive and beyond the needs of router.
Nice! You're already familiar with them š
Yeah do you know a really good 5.8-6.2 dB?
They don't. The hats you guys already have might be designed for the pi zero, but you can run the S22 modules on an ESP32 easily enough. Definitely power hungry though.
I've had some decent luck with these: https://a.aliexpress.com/_mPtcRwP
They are decent, without being amazing. But they are cheap š
And have been reliably consistent on the VNA, which is better than I can say for many cheap antenna models.
SWR generally dead on around 1.3 in free space at 918.875MHz, without additional grounding. Or if you use them at 868MHz, they're almost at 1:1.
good catch. Wonder if they can be ran off a rpi 2040 or 2350b now haha
Oh damn. thanks. ill get one and test it at that cheap of a price
@vital hemlock @leaden crow @broken turret We could do SF but we would need 1 watt, good filter for 51 (I could make one if we are a go for 51), and all standardize on 8.5 dB roklands. They have 40 degree apertature and thats enough for how high they are
This is with the STOCK calcs with all the accurate env data and without the increased sensitivity filtration would give us which technicaly should put me calculating this at -131dB for short fast but I digress I wanted the worst case. Heres what the site planner found but its gonna take a few routers. The only way the routers see each other but distinct cells is with 8.5 dB. If we use 6 there might not be enough gain for each router to see the other
But also we now know. This DEFINITELY would work for MF haha
Oh crap I forgot lake mountain LOL
and finally If all on that old connected map was done. A few picks for those who dont know the topography we are looking at.. I guess it really can be done but all the routers would have to be rather high quality hahaha
A simplified look. Red would be routers. Purple (Router Lates) surprisingly nelson has some shitty blind spots, Green baloon strategically placed clients that people would normally make routers but shouldnt. They would be helpful but overall are unecessary unless you wanted to slap some cool sensors on it and take some cool data.This might work for SF. Guaranteed for MF
R: Filtered 8.5 dbi 1 Watt
RL: FIltered Typical Rak GPIO labs filter.. well matched atenna
Clients; What ever with a good antenna.
Bonus .. I kinda want to make an underwater drone.. so I might be tempted to make a submersible client in the lake. If I do.. youll get picks.. if I cause a lithium explosion.. you will hear nothing
The way the mesh actually networks amazes me
So forget topo for a second. You have all these nodes. They hear.. and they wait to broadcast until the other clients are done but as they progress the increment count is only done as a node actually resends it out so some paths can exhaust themselves while others dont within the window? and then of course everything we discussed today ontop of that basic concept. I want to deep dive and know mroe because im sure im still wrong
Seriously though I bet with a really good filter a Rak 4361 could become a great router.. Anyways Ebyte Powerhog S3s it might have to be but im kinda bias for the rpis.. if we want LTE backhaul.. or someone wants to do that.. would be nice to have a linux client to ssh into. Cheap data only plan
SOme of these are gonna want more battery than others Im just gonna say it now haha
For what it's worth, I've found the site planner to be a bit pessimistic. At least around here, things generally seem to work a bit better than the meshtastic site planner predicts.
Any time a node rebroadcasts a packet, it decrements the hop limit by one. The version of the packet that it will rebroadcast is whichever it heard first - that's usually the one with the highest hop limit remaining. If it subsequently hears that packet again, it will simply ignore the duplicate, and that duplicate has no influence on the hop count at that node. Yes, some paths can be exhausted earlier than others - that is why it is important to site routers well, so you get robust coverage in the early window without exhausting too many hops.
Can confirm the 4631 makes a great router. All our high sites are standardised on that board.
I use 8x21700 cells in ours. >3500mAh per cell. If the solar is completely disconnected, they can still go for about 2.5 months on that battery alone, on SF. Note that the slower the mode, the more battery it'll eat (because the radio is required to spend longer transmitting). But even on the slow modes, they're really quite efficient, and will still go for many weeks š.
I have 4x at the moment for each plan I have. I think it will be enough because ive cooked up something hilarious.
So not only do I have decent sized solar panels but also My budy and I are 3d printing what for all intents and purposes looks like an anemometer but isnt hahahahah.
Inside youll find a very low friction permanent magnet 3 phase dynamo... spun by a Belt drive with a gear reduction of 22.71 to 1 ... a full bridge rectifier.. a capcitor bank.. a regulator.. a lion battery charge chip and another set of capacitor banks on a PCB LMAO
and it produces about 6-8 watts at 12-15 mph wind speed... we are optimizing the cup and moment arm for the desireable toque but essentially our back up is wind HAH
These things are hilariously powerful for what they are
24V 0.25A easily
For a RAK.. this definitely charges it.. but id like to see just how low I can bring optimal generation before making a permanent design
what would make it more powerful of course is if that ring magnet could be traded for a N52 as strong as I can muster into there.. that would be funny af tbch
@vital hemlock lets give the SF plan a try with filters. Honestly.. I know we want 1 watt devices and im game but its very economical to go with RAKs.. for now if we want to get things up quickly and want economical filtering
I ordered the acasom and cables for ds nsl. The antenna I have for it is from a slightly different listing than the starf erayd recommended (320mm starf rated 7dBi)
excellent. that should go well. Im honest to god probably going to put a RAK 4361 based node with a decent filter on black mountain here soon
thurston ahs to wait.. got fresh snow and its DEEP
Because I have two radios on my roof I can test anything whenever you like. Just let me know. I have former roof node on MF right now, channel 51
Just in case anyone pops on there
I can turn my rv node on and switch it to a different preset for testing. Itās a RAK with an Alfa. I think bay area mesh has the right idea with MS. Itās link budget is only 2.5dB lower than LF.
𤣠I can make an extra good FS51 filter but it could cost me. I could get a Teflon based PTFE PCB
Ceramic cavity from mini circuits
integrate an ultra low noise amp to make up for the losses
Add a chebysheve 8th wave transformer to deal with the transistor parasitics
and have an EXTRA tight filter+LNA but running it in my head its 120 dollars LOL
Honestly if all the nodes on the mtns are remote admin key changeable. we could try moving a router to MF but I dont know if I want to get snow shoes out for anyones broken router just yet
Its slushy up there now
and muddy
I figured out something about the high gain antennas. The beamwidth goes up as you go up to 1/2 wave size for the same gain.. so a rak or rokland high profile 8.5 dBi thats like 36-42 inches long will have a great deal it can see up or down.. as you get smaller the aperature and beamwidth and profile compress
so if you want a 8.5 dBi go big
like I did
Its why I can hear NPR and Onaqui with 1 hop
Like I can hear my friend in the cedar moutnains area with 1 hop.. but they cant hear me. Just too far
Like these are what I have on the router Late clients
https://store.rokland.com/collections/802-11ah-wi-fi-halow/products/8-dbi-n-male-omni-outdoor-helium-915-mhz-antenna-large-profile-40-for-rak-miner-2-nebra-sensecap-m1-bobcat-hotspots
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
Its lower gain counterpart https://store.rokland.com/collections/802-11ah-wi-fi-halow/products/5-8-dbi-n-male-omni-outdoor-915-mhz-antenna-large-profile-32-height-for-helium-rak-miner-2-nebra-indoor-bobcat
The rod from god: https://store.rokland.com/collections/802-11ah-wi-fi-halow/products/10-dbi-backcountry-n-male-omni-outdoor-helium-915-mhz-antenna-48-for-rak-miner-2-nebra-indoor-bobcat-hotspots
Rokland introduces our large profile 5.8 dBi 915 MHz 31" fiberglass antenna with N-male connector and U-bolt bracket mount. This measures 10" taller than our low profile version. This is certified and designed for outdoor environments and is weatherproof.Ā Ā ThisĀ antenna isĀ compatible with the RAK V2, Bobcat 300, Nebra
Introducing the Rokland Backcountry 10 dBi outdoor Helium & IoT 915 MHz 45" fiberglass antenna with N-male connector and U-bolt bracket mount. This is designed for very rural and remote locations and is not advisable for urban or suburban locations.Ā With this very high gain, the signal can receive from a very long
please dont use the rod from god
I can guarantee unless exactly 15m from ground height.. you will not have a good experience
If anything use the 5.8
Another great antenna
https://store.rakwireless.com/products/5dbi-fiber-glass-antenna-supports-863-870mhz?index=102&variant=43034795114694
@broken turret
@vital hemlock
I think i know our first new router.
Mt olymous
Its kinda OP and tbch would work really well with Point of the Mountain being Router Late
Then I'll just figure out a next best placement router up north haha. Cuz I overlooked Olympus when making the map.. I thought it was too close to people but reality is.. Its a mountaineers mountain.. Trully and boldly lol
Aint no easy way up and the only easy way down you don't survive š¤£
If I can walk out to the front cliff face and repel down to a lower traverse and maybe if lucky even find and undercling to mount and achor it to.. I could get it in a spot that even when seen aint nobody gonna try to go to it
Ill start seeing if my gears ready LOL
somwhere on these escarpments
I think there are plenty of places to put one and hide it high up.. just will take a bit.. ill probably bum rush the peak and slowly decent to find a good spot
@leaden crow since theres still a good bit of snow on olympus imma try something maybe this weekend if everything arrives..
Part of the way down bountiful. The mountain itself jets out more than the others .. if I can get one in a tree on a lower point there.. I think that would be super ideal and its 10 minutes from me.. plus a 5 hour hike up triumph
ALRIGHTY if y'all 801ers need a 3d printed meshtastic part I can print any engineering grade material under the sun. Willing to print stuff for this now that I got nylon, cf nylon, cf peek and PCTG dialed
If you cost me a nozzle with a material I don't yet have calibrated.. There will be a 40 dollar surcharge.. Jk jk. I've got a Polycrystalline Diamond Nozzle. That things a tank. I can print ceramics in volume
Routine numbers people want
Print size 220x220x250
Flowtech Hotend by Microswiss
0.6mm Diamondback Nozzle
Dual direct extruder
Enclosed and Climate controlled
Bed: WhamBam Woven Carbon Fiber Build Plate; Compatible materials=no known incompstible materials.
Other nozzle sizes owned are 0.2, 0.4, 0.8 and 1.0
Materials supported: all Base Thermoplastics, all high performance thermoplastics, all engineering grade copolyesters, all composites, all specialiy under 365C hotend temp
Send me an STL and I can have a slicer and idea of how much material and how long and stuff
Is Lake Mountain going to stay on channel 20 or can we get it to 51? I don't know if the owner of that one is on here.
@dim bay How does your Lake Mountain node look lately? Is it still pegged channel util. on LF 20?
You are more ambitious than me, bit I also tweaked an ankle last weekend in Moab. Probably not hiking for a bit.
Less ambitious than you think. I just am constructing my nodes small and discrete..
For mountain top nodes I'm using these to strap and structure them with pipe clamps (insulated and rubberized).. They are lighteright. You can easily carry 4 to 6 of them. Insulating fiberglass and UV resistant. Ridge enough for a 50 watt renogy solar panel. 100 lb weight limit
I don't need a ton of rigidity. But I do need them to be able to be in trees up high enough not to get screwed with
And I need to be able to construct something loose enough that the branch can grow and stuff but also rigid enough to survive years without issues
So that's my design choice. What y'all doing for mounting?
#solar-power message
I'm basically making a shallow small pyramid. Branches grow.. No issue as just a loose strap to the branch.. Top changing as the span moves.. Doesn't really harm it as it's loose up top.. And then I just mount in the center spring loaded.. And that's my dampener.. And it can buffet around
I'm kind of tempted to file down the teeth on the cam/clamp of my straps. They only need to hold a few lbs and it'd be preferable if they could slip a bit with tree growth. Also, kind of assuming it won't be an issue if I visit and adjust it 1-2x per year.
Sweet so tie down straps. I questioned how long mine would hold..Didn't want to be strap to the the failure lol. Mine just sucks to construct when up a branch or two with everything temp strapped and the weight up there haha.. Nice thing is its a bitch to get down too so easier than a strap to cut.. The average hiker Karen is what I worry about. š
Cuts it down.. Kicks it. Leaves it there. Pretty common behavior
I've considered the loose hanging node too. Cuz I'm designing an overkill 3d printed box. Aiming for IK10-IP68
TPU and IK10 is the easy part lol
Ip68 is the hard part. I want it submersible in case of ice
I'm trying to make a maintenance free version that has no issue free hanging for the life of the average tree up there
I scouted around last fall before crest was snowed out. I'm fairly certain it we be all but impossible to see up in a tree and ~50'+ away from a trail.
I'd use UHMWPE rope
Ahh yeah
I've just got a few spots where it's more those low brush trees
And those stout thick deciduous trees
The are more like an understore
That's what's out near Ogden.. Less piney