So these are using LoRa (902-928MHz) as the actual frequency for text transmissions? This is really an exciting frequency that's been around for a while and has a lot of potential for disruption.
Fun fact--there is a completed WiFi standard for this frequency (802.11ah / WiFi HaLow), but as far as I know, there have never been any chipsets made for it.
Especially for cost--my group recently picked up a pair of 900 MHz Ethernet radio modems which acts basically like a 40km long line-of-sight Ethernet cable. Super useful for field research, but at US$900 per pair, they're above the "let's see what happens" budget for true experimentation. A $90 pair, on the other hand, even with a slightly lower range, would put them comfortably at about the price of a Raspberry Pi or other SBC, opening up a lot more possibilities.
I looked at the same thing for a project a couple years ago. We wound up writing a simple script using the TUN interface to send the IP packets over Digi's XBee radios (there's a 900MHz edition). Of course, sending all the packets through a userspace Python script was by no means efficient and the radios had only a 100K RF data rate to begin with, so the data rate was like 800bytes/second or something like that IIRC. Probably would be enough to cause some applications to timeout or give up on the network, but it was good enough for SSH and simple REST APIs.
There's no way 802.11ah will deliver 15Mbps, but if the standard is designed with that kind of data rate its gonna be a lot faster than what we were using (main reason is Wi-Fi is direct sequence spread spectrum while the XBees are frequency hopping, allowing a larger RF bandwidth). The one thing that is good about the XBees is they will support repeating messages for each other and they have a nice mesh functionality as opposed to Wi-Fi's client-AP model. Additionally, we were just using the small wire antenna on the XBee, which only outputs a fourth of the legal maximum power to begin with, limiting the range. The chipsets OP linked support the full 1 Watt output power.
As an aside, I also saw a serial modem using MURS that you could theoretically run a dialup program over. Also ridiculously expensive.
Yes to “LoRa”, but not necessarily on the 900MHz band. I have a few all using boards with 433MHz LoRa radios, most of the hardware supported is available in 433/815/915MHz variants. (In fact, I’m fairly sure most of the chipsets are capable of all those frequencies, and the only real difference if you buy the 433MHz version is the antenna they include.)
Yeah, this has been my experience too. Generally you use the frequency that has been legalised for use in the region you operate it in. I wonder if this automatically switches to the suitable frequency based off coordinates (when using a gps’d board) I’m sure it would be trivial to implement. Ah good ole opensource. This really is a great solution, riding around New Zealand the only solution is a SPOT tracker which charges like a wounded bull so I love that this is disrupting this space (and using clever tech - heck might throw one of these on the moto just to use it for security)
For Meshtastic, you upload a country/frequency specific firmware.
It’s perfectly happy running on devices without gps, so it’d be a maybe nice to have but not a thing the project could rely on to meet regulatory frequency compliance.
For your motorcycle use case, keep in mind this is a small group of people mesh thing, not something where one device can expect to connect to unknown other people. The SPOT tracker is still the solution you want if you’re looking for safety/emergency contact.
What meshtastic might be useful for on motorcycles (and I’m considering building some of these) is a medium range mesh network between you and your riding group, with super simple super limited Comms UI. I’m thinking 3 or 4 different colour old school video game type buttons that you can mash wearing gloves, that light up a bright-enough rgb led to see in the sun, with 3 or 4 agreed meanings for the colours. “Pull over and check messages”, “Wait for me”, “I’ve crashed”, “Cops!” type messages, that can be sent/understood without needing to read a text message on a phone screen (but still having that text messaging available as well).
The low bandwidth and the mesh topology put a hard upper limit on the number of nodes in a channel. From memory its limited to 250 nodes in the default settings.
"I suspect up to about 30 nodes per channel would be fine, I bet the channel will be almost unusable once you reach 100 nodes."
As for voice - I think in default "very long range mode" the mesh bandwidth is down around 20 bytes per second. It's possible to get higher bandwidth at the expense of range, but it's really not what you're looking for if you want voice comms.
Australia only has half the regular international ISM 915 band, we sold the top half to telcos. (Meshtastic knows and deals with that if you select the 915MHz _au firmware.)
It's great but until someone comes up with something better than 'solder on a screen, 3d print a case' it's going nowhere for lack of critical mass. The best thing that could happen to this is if it turned from a $35 project into a $50 product. I know a bunch of people who find it interesting and would be into investing in deployment (as backup comms for natural disaster scenarios), but nobody wants to commit to soldering 50 of these things.
From a quick calculation (micro-production, 1 man operation, 50 units batch):
- Materials: considering the best supported module (T-Beam V1.1 w/ NEO-M8N /w SX1262) with an OLED screen, an upgrade to a decent GPS antenna (ceramic, 28mm^2), 3D printed case and 2 Panasonic 18650.
Materials cost: 48€/piece.
- Assembly: pipelining would probably be bottlenecked by display soldering (outsourcing the 3D printing). Batching soldering could bring that down to ~1 min/piece. Flashing/testing can be fully automated and parallelized with soldering. Final assembly should be an additional 30s/piece. Start up ~10min considering testing machine boot and soldering iron warm up.
I would buy one, as well as few naked one that I would solder myself.
Those things could become very nifty. Solving problems that mesh-wifi can't quiet get right for organic deployment.
There is hardware available with the same microcontroller (ESP32) in a thick watch style case, but I don't know if anyone has ported meshtastic to it yet. On the face of it, power usage looks a bit higher due to the LCD display rather than OLED.
Not very, but the point is that the necessity to set up an assembly line is a sticking point for adoption. You need to deploy a lot of these to make it useful so skill or time bottlenecks are a big disincentive.
This is absolutely brilliant. If this works as advertised, it'd be trivial to add mesh network capabilities to lots of the mountainous areas I'm frequently in while hunting or hiking or what not. That would be huge for safety, and substantially cheaper than a sat phone.
> safety, and substantially cheaper than a sat phone
I agree that it's a really cool project but I personally would not use it as a primary safety mechanism.
I've got a ham radio license and I see the same mistake with people who think a 5W VHF handheld or 10W HF radio will help them in dire straights. HF isn't reliable enough and all it takes is a little terrain to block VHF. You're also not guaranteed that there's even anyone on the other side of the radio if you really need help.
I would recommend a SPOT or inReach as a your primary safety device with this as a (neat) backup. All they need is a line of sight to the sky which is much easier to do than either LoRA, VHF or HF.
I'm in mountain rescue and I'll second all of this. It's why we will set up repeaters, or station people as repeaters if we have enough people, but more often than not, we are out of radio communication for hours on end, despite all parties having the necessary equipment. More often than not, we'll use inReach texting to maintain comms.
But projects like this excite me for future directions. Sometimes when we've been in areas with cell phone coverage (some wilderness in SoCal has coverage), we will be able to see tracks in SARTopo update live from teams in the field. It's really cool when it works!
I'm totally 100% on board with this as a secondary data/telemetry mechanism.
For all the open source software out there it's a tragedy that there's less of that culture in ham radio(which is one of the primary charters of the spectrum-carve out we have). Stuff like this and the M17 project give me hope that we might some day have something more than just the dumb AX25 packet radios(which are stuck in the 80s at 1200bps and zero FEC).
Not able to reply to the sibling comment from bsder, but yes, encryption is forbidden in the ham bands.
What vvanders is talking about, and it's a massive point of frustration for me too, is that all of the layer 2 open standards that have actual use are ancient. Like, we're supposed to feel lucky that we whave ARQ. Ridiculous.
It's not really anything to do with encryption or indecipherability, but all about spectral efficiency, and most of that really comes down to coding. I can pop open fldigi and generate a 64QAM signal really easily, but actually coding the data to ship over that pipe starts getting into some pretty arcane territory.
Think about just how good a modern cellphone is - 100+ megabit connection in the palm of your hand, competing for spectrum with hundreds or thousands of other devices, with a handset power of less than 1 watt (usually). There's nothing inherently magical about it, it's just not something that there's a lot of FOSS work around.
I've been working on deep diving on this a lot lately, I really think there should be more open source libraries around coding, but I have a funny feeling anyone who gets good at this gets snatched up by someone in the 3GPP consortium.
> Stuff like this and the M17 project give me hope that we might some day have something more than just the dumb AX25 packet radios(which are stuck in the 80s at 1200bps and zero FEC).
Isn't there some weird regulation about "non-decipherable" communication holding back usage of Ham bands?
Not really, as long as you publish a spec and it isn't encrypted(except for some really rare satellite control use cases) it's fair game.
The really annoying one is they put symbol rate into law instead of bandwidth so we'll never see above 9.6kpbs in the VHF bands(there was a petition a while back to repeal it but from what I recall a bunch of HF folks though it was going to cause chaos across the bands and I don't think it went anywhere).
Wasn't able to reply when I first saw this, I think because it was too new a comment - posted a long-winded reply as a sibling though, but generally it's not a matter of regulation, just a matter of interested and skilled parties willing to donate their time.
> I would recommend a SPOT or inReach as a your primary safety device with this as a (neat) backup. All they need is a line of sight to the sky which is much easier to do than either LoRA, VHF or HF.
I have quite a bit of experience with both, and wouldn't recommend either as a primary safety device. PLBs are far better.
SPOT is based on Globalstar, which leaves one without coverage in many areas. I also found it to be incredibly flaky even in areas of allegedly great coverage with a totally open sky.
inReach is vastly more reliable, and with global coverage, so if I was buying one for a (secondary) safety that would be the easy winner of the two. It also has the added benefit of bringing a convenience that PLBs don't, as one can grab weather updates and easily text others for information.
That said, it's not difficult to get into a situation where terrain makes quick satellite acquisition very difficult, and one could wait hours for a message to get out. PLBs are much more reliable in adverse situations (N-S slot canyons), have higher power output than satellite messengers, and are generally built to higher standards. They also don't require a monthly fee.
A legitimate PLB would give one more paths to a swift rescue, especially if SAR aircraft are already up and looking for you.
Are there PLBs that let you actually text now? I’ve been told that many have been abused and are not treated as urgent any more, so a device like an in reach where you can actually communicate your situation is preferred. It also means rescue could arrive with the supplies needed for a particular injury, which makes a huge difference in some evacs.
> I’ve been told that many have been abused and are not treated as urgent any more
There is an old type of PLB and a new type. That was a problem for the old type, in principle. In practice, the search coordination centres got pretty good at triage. Going down Route 66 at 70 mph, unlikely to be a real emergency; halfway up Mt McKinley, scramble the chopper. Accidental activations are far more common than abusive ones.
The new type broadcasts its serial number, and everyone is required to register an emergency contact person when they buy their PLB. If the beacon is activated, the contact person gets a call from search and rescue, asking what the beacon user is up to.
The ACR ResQLink View is usually a good default choice. I've read good things about McMurdo, but everyone I know who has need for a PLB carries an ACR, either the older 400 or the newer View, which I would recommend over the 400.
But just as an inReach isn't a replacement for a PLB, neither is a PLB a replacement for an inReach. It's great to be able to communicate in emergency (and in many non-emergency) situations.
I think that the mesh aspect of this would be a substantial improvement over VHF/HF radios. I'm sure I'm wrong, but I don't know of anyone that is running a mesh VHF or HF system in the mountains. I always carry my HAM when I'm out, just in case anyone else is, but obviously that's truly local communication and has the downsides that you mention.
With a strategically placed mesh network and some solar panels, you could potentially overcome a lot of the downsides of radio because you could target high areas, such as mountain peaks, to help bridge those obstructions.
I definitely don't disagree that the #1 safety device is a GPS/sat phone, but those are prohibitively expensive and most people don't have them. If we could get some solar powered radios along ridges, and a decent percentage of people carrying them, I think there could be a very comprehensive mesh network that would work for most needs. Bearspray is already pretty ubiquitous, I don't think it'd be that hard to get people to bring a $30 mesh radio with them as a routine thing.
Take a look at the Ars article, it's really hard to get coverage in moderately flat areas over a reasonable distance(1m+). You also quickly start running into the Hidden Node[1] problem or the N^2 aspect of channel congestion(which is why LoRA has such low throughput).
Cell companies spend billions on this and even then you'll still have gaps outside of the metro areas, and they're using much more advanced modulations like CDMA and FDM/TDM with a central coordinator.
I've talked with a fair number of people who maintain commercial repeaters in mountain areas and those can be incredibly difficult(combination of hard access in winter, large temperature swings and theft of solar panels/batteries) so which it might be possible, my guess is the level of investment just doesn't pencil out for the usage.
From some quick Googling it looks like LoRa accounts for the hidden node problem by using varying timing? Sure, it increases the amount of time required to send a message but as long as the device continues to attempt until it gets an ACK it seems like this could work? Of course, it also eats up the available airtime for the mesh but that's the trade off right?
I'm not a radio hardware expert and I'm not fully up on LoRa or the code of this project at all so I likely have no idea what I'm talking about. Was just wondering if someone else does?
No problem, it's a really neat domain and there's a fascinating intersection of physics + math(you can reconstruct freq mixers/modulation with a basic trig foundation) and practical application that makes radio a really cool space.
I also think that there's still untapped potential for secondary services and the like so it wasn't my intention to dissuade anyone form picking one up and giving it a try.
I would be surprised if it was cheaper than something like a Spot or Garmin inReach when you factor in the cost of building out and maintaining such a network.
I've carried SPOT/inReach/[Iridum/Globalstar phones] extensively in remote areas, and while they are nice to have for convenient communication, I experienced enough issues with each that I would never rely on them in a life-threatening situation.
Careful you don’t expect too much here. At long range the bandwidth is very very low and latencies are very very high. Meshtastic will never be a replacement for a sat phone. Think of it as more like a two directional pager. Short text messages with latencies possibly as high as 40 or 50 seconds.
Lots of people disappear from the forums when they realise they won’t be able to something like use this to stream YouTube to their off grid cabin...
Right, I was more thinking a mesh network solar operated to let people report their location and status across rough terrain. I may try making a few of these over the year for next hunting season, and place them in a mile or so radius from basecamp for the duration of my trip, anyways. That way we could split up but still maintain contact, get help if we shoot an elk, etc. We're definitely outside direct walkie-talkie range, and I'm in the only one in the group with a HAM, so this could be pretty useful.
The project (mostly) assumes one device per person. Each user carries a tBeam (or one of the other available boards) and Bluetooth connects it to an android phone. The devices mesh with each other, and the phone is used as the interface, with a map view and a messaging view.
Meshtastic automatically sends each nodes location (at a fairly low default update rate, but you can configure that).
There’s a “repeater node” functionality, so a node at base camp or on the highest local terrain can improve the connect ability of the mesh. There’s also a group of people on the forum experimenting with solar power solutions for these nodes, so there’s a fair bit of reported experimental data on solar panel size vs battery capacity vs device runtime (which is all trickier that almost everybody assumes before they’ve tried to do that in real life. “Specs” for solar panels, battery charging controllers, batteries, and even Meshtastic device power consumptions all have way way bigger error bars that most people expects...)
I haven't looks deeply into Meshtastic's repeater mode, but it runs on the same hardware as regular nodes. (None of my current personal experimental use cases call for repeaters.)
What I do recall is it can "repeat" encrypted "channels" by just replaying the encrypted packets without having any keys to read them (I think there's some very minor metadata in the unencrypted part of the packets, perhaps just a hop count?)
Also note that "channels" in Meshtastic are not frequencies like in typical UHF terminology - but pre shared encryption keys. I _vaguely_ recall maybe they get used to seed freq hopping algorithms (but I might be misremembering that).
I think the big difference between "repeater nodes" and regular nodes is that the repeater settings shut down the GPS because its unnecessary power draw (on the assumption you know where you installed a repeater node), and participate in the mesh in a "hidden" way, so they don't show up as members you could communicate with.
The project is really interesting and I think I am going to try it.
It looks like it's text only? would it be possible to add a very filtered, low-bandwidth audio stream since the phone can do all the calculations?
I have no idea on the actual bandwidth of these devices, wikipedia says LoRa is 0.3-27 Kb/s and Opus should work decently at ~10Kb/s I think. So maybe?
Although audio would probably be half duplex and only point-to-point instead of mesh, but it still could be interesting
Codec2 is part of it but you need a lot more pieces on top(framing, modulation, etc). LoRA doesn't really have the bandwidth for Codec2(for a few reasons, mostly power and the way LoRA handles channel congestion).
M17[1] has been making really big strides in this space and really excited to start seeing a proper open hardware/open source digital radio stack.
Another thing to consider would be latency due to retransmissions. Something like voice notes would probably be better than a live call send it could be re-transmitted as many times as necessary and broken into small chunks.
Meshtastic has been talking about merging with Disaster.Radio for eons, because the two seem to have a lot of conceptual overlap. I can see uses for both feature sets and would love if they were just one thing.
disaster.radio seems kind of dead. No commits to the GitHub in 4 months. The contributor who responded to the latest issue says they aren't actively working on it. I think it would be great if they merged as well. Do you have links to any conversations about this?
we talked about it in the early but there were some fundamental technological limitations that made it seem like not a good fit. At this point I think meshtastic is more feature complete and has lots of areas not serviced by disaster radio (ie. python API, ports to many different types of lora boards, extremely good battery life).
I think the only feature we miss wrt disaster radio is dual radio support and full DSR.
Patiently waiting for my T-Beam to arrive so I can try this out.
This is an open source alternative to the popular GoTenna devices and if any HNers can point me to other options in this space I'd love to try them all out.
I believe so, yes. But in all seriousness a line has to be drawn somewhere about open source'ness.
For example, there's a few ways to remediate that. My preferred is a reference Gnuradio flow that, with an SDR, allows for communication with their device network. Secondarily, open firmware would be good... but with open firmware still results in closed silicon (and not everyone has access to a fab!).
Closed source isn't always enough to run me away. Closed and closable interoperability does.
Yeah. Things get a little fuzzy this close to the hardware.
The “open source” project is a (platform.io) project that runs on the ESP32 (or a few other similar enough) microcontroller. It controls the LoRa (totally closed) firmware on the LoRa components (and GPS as well).
Even your Gnuradio setup has similar black box firmware underneath that it relies on (unless you’ve got to extraordinary lengths to ensure the machine it’s running in has no propitiatory firmware for things like disk/ssd controllers, USB controllers, battery management systems, etc. )
My current node is line of sight to the U.S. mountains south of Victoria, so if that's your location and you have line of sight, you can probably connect to me right now!
Me and my friends would pay for this. We fly paramotors and almost every weekend we fly to places with no cell connection.
Plain old radio is terrible (there's always someone transmitting all the time without realizing, blocking communication for everyone else) and the ability to check each other's GPS position would be awesome to know if the whole party is still flying together.
Try PilotAware (pilotaware.com). It's a Raspberry Pi with SDR comms. A web app gives you a radar type display. It'll integrate with moving map software for navigation if you want (eg. I use SkyDemon). I'm not sure about availability outside the UK though.
FANET [1] is specifically for paragliding. It is also LoRa based and open source, but so far only Skytraxx and Naviter really seem to be using it in their products. I think it's exactly what you're looking for, though.
I would love this for my dog, she runs in the woods and I like to know where she is. I've tried a Lidar collar but the range is too low. The only other option is a cell phone in a case (which is used by search and rescue) but it's to expensive since you need a cell phone plan, and it's and too bulky to attach to a moving animal.
There is a market for this, I'll help prototype it , someone start a company to track dogs and children (out playing, biking, skiing) :)
That's similar to what I had, you also need a gps receiver, so your investing close to 650$, not great for when your dog smacks it on a rock or tree. The problem is in a mountainous area you need line of sight, so it became practically useless, I would constantly loose the signal. Combined with the battery life in cold weather... The only solution was a satellite connection or creating some sort of mesh coverage like the OP link suggests.
ps. Most dog trackers also do not work outside usa.
Check out https://getfindster.com/. I've using it for 3+ years.
It has 2 parts: GPS sender and receiver in your pocket.
Works where there is no cell connection.
It has some "magical" wireless connection over 2 km. I've can confirm 1.5km
I was hoping for something like this but unfortunately LoRa seems pretty meh if you don't have line of sight between the devices or maybe open enough areas that don't drown the signals. I was hoping that it could go for a couple hundred meters in urban areas (was trying this out in the outer London zones, UK) but the results were pretty bad. Reception was spotty the moment direct LOS was lost and non existent ~10m after that. Not even worth mentioning the results if one of the nodes was inside a building.
If anyone had any success in dense areas I'd really appreciate any links to the hardware/antennas/firmware/software you used to get good results.
Antennas and height make a huge difference. I’ve got a hop from my place to a friend that’s a little over 3km non line of sight (terrain/hill between). Can’t do that with the stock rubber ducky antennas, but with a swr meter tuned home made dipole antenna on a 4m pole on my roof it’s rock solid (using 433MHz, not sure if 815/915 would get that far).
unfortunately that raises the entry level by quite a bit but definitely worth a shot. I need to check about the band, cause most of the LoRA I saw being sold in the UK are 868/815, but a cursory google search says that 433 might be allowed. Will take another look thank you.
Very cool, I just ordered a couple radios to play with. Looks like the app is Android only although an iPhone app is under development. Any recommendations for a cheap android phone/tablet to use mostly for this application?
I’m using old Samsung S3 and S4 phones with Android7 Lineage on them. I would go any older than that, and the Bluetooth works more reliably if you can get Android 8 or newer (none of the Lineage images of 8 will boot on my S3s.)
You don’t need anything that you probably couldn’t find in the back of your or a family members junk drawer...
Moderately well. I can get a couple miles of range punching through trees, far far better than any other unlicensed option. Chirp spread spectrum is pretty cool.
Can you elaborate on specifically what you have that's getting a couple miles? I tried out Meshtastic with a bog standard board exactly as it existed. I carried it down the road and found it got ~500m of reliable range with no obstructions. Has the software improved significantly, or is there different hardware to support getting multiple miles?
Boards are the Heltec boards that are an ESP32/SX127x. One of the boards is using the stock antenna that came with it, the other has a TBS Immortal T I had lying around. Using the 915mhz bands as well.
I use it out in the mountains, with line of sight between hilltops, but pretty forested so it's having to punch through foliage. I did some unscientific testing around my house and did notice a lot less range, I'm suspecting there's a fair amount of noise from other things in the ISM bands in the city.
I was wondering the same - I've seen gotenna stuff sold at REI and other outfitters. Meshtastic makes the claim that the gotenna project is focusing on enteprise deployment and not end user (plus the usual dig that it is closed source). Though the gotenna webpage still has the personal mesh devices for sale.
Meshtastic sounds like an open source alternative, but if you have to built your own device I don't see it growing beyond serious DIY.
Fun fact--there is a completed WiFi standard for this frequency (802.11ah / WiFi HaLow), but as far as I know, there have never been any chipsets made for it.