MeshTNC is a tool for turning consumer grade LoRa radios into KISS TNC compatib | Hacker News Reader

MeshTNC is a tool for turning consumer grade LoRa radios into KISS TNC compatib

github.com

33 pointstodsacerdoti19 comments

I haven't done any AX.25 (or KISS) for over 30 years. Is it still a thing, or has the packet radio community moved on to something better? Back when I began, there were no turn-key solutions, so you often needed to modify your radio to get something on the air. This was especially true for 9600 baud FSK setups.

For a while, there was a community of stations creating an infrastructure similar to the dial-up BBS world, including message forwarding (UUCP).

There were even a few Internet gateways for a while. (I ran one of the two that were reachable from my corner of LA.) I imagine they're everywhere today.

(Getting married and raising children can quash a lot of hobbies.)

I also wonder about the state of packet radio.

When I had time, I had no money for equipment. Now that I have money and knowledge, no time…

I am just getting into the packet scene in the Boston area.

APRS aside, as far as I've found, there are about a half dozen Winlink nodes in the area and one BBS. And one lovely node in Cambridge (KZ2X-1 [1]) which provides connectivity to a bevy of ancient (though virtualized) OSes.

I don't know how much AMPRnet activity there is. There are only 7 allocations in the area (mine included). I'd love to be able to e.g. log in to my home network from a few radio hops away but I don't think there's any infrastructure in place for that (such as Mobile IP).

[1] https://kz2x.radio/posts/complex/

Here in Europe there's nothing left except APRS which I find boring and incongruous with current developments (I try to avoid sharing my location, not to do it actively). What I miss are the long chats.

What we needed was a faster packet but it stopped at 9600 which is no longer useful for anything these days.

Thank you for the report. I've never had any interest in APRS either. Seems like an Air-tag would be more reliable and less wasteful of spectrum.

I once worked a European station from my station in Los Angeles. It was Halloween sometime around 1990, and I was on 10m (HF) packet. I had used a digipeater in Washington D.C. to have a QSO with G0BOO. It was memorable mostly because of the remote station's callsign and the fact that it was Halloween.

Yeah packet was so nice in the 90s. I have really good memories of it. At times we could even hop across the world using satellites. This sounds a bit bland now but it was in a time where people didn't have a little slab in their pocket connected to the whole world. The "I connected via satellite to somewhere in the US" was pretty much science fiction. I'm pretty sure all of those pacsats re defunct now. And the ISS just routes APRS.

I used to have long chats all night and often during the day. Sometimes even with UI packets (broadcast only) because the connection was too poor. This was an amazing thing for me because at that time I had to pay per minute or BBS chatboards. As a young computer enthusiast this was really annoying because I didn't have much money as a schoolkid and my parents had to pay the bill (and hated me tying up the phone line). Packet was completely free, I could hang out on it all night. It was slow, but most modems were about the same (9600 was still a common modem speed). Despite me being on 1200 baud at the time (9600 was pretty new and expensive).

Of course packet was extra slow because it had to deal with medium contention and interference, but still, it was comparable. It was fine for chatting and sending messages etc.

But what happened was, modem speeds increased. Internet happened. Dedicated lines appeared which invalidated the huge benefit of free access. And APRS attracted HAMs that were not interested in computers. There were many of them and few of us and our numbers were declining as all us interested in computers were mostly online by now. For me APRS was boring for the reasons I mentioned. I don't care about seeing people on a map if I can't chat with them.

I don't know why AX.25 was never upgraded to higher speeds. A 25 kHz channel could have supported a lot more than 9600 baud. But well, it is what happened.

It sounds like we've had some similar experiences, but you must be about 10 years younger than me. I got my ham license because I was interested in packet radio. 1200 baud AFSK was all that there was for the my first few years. (Of course there were still some people doing AFSK FM RTTY in the VHF bands then too.) I did both HF and VHF packet and had a lot of fun. Later, when 9600 baud started to become a thing, I modified my Kenwood TM-721A and got on the air with it. I was pretty proud of the mod. I put a new mini-DIN connector on the rear panel, and it looked like it came that way from the factory. The insides were not as clean, because when I drilled the new hole in the back for the mini-DIN, the drill bit went in too far and tore out a bunch of wires. It was "fun" to repair it. I picked up a Tiny-2 TNC that supported 9600 and I went mobile with my TRS-80 Model 102. I left the Kantronics KAM Plus connected to my base stations. (It could support both HF 300 and VHF 1200 at the same time!) My base stations used a Kaypro II (CP/M) for the packet terminal.

When I first started out with packet, everybody was doing unproto AX.25 through a digipeater on Mt. Wilson. (I think it was on 145.09 MHz.) I met a lot of people and made some new friends. It became less fun when more and more people were getting on the air, because of the limited bandwidth available. To help with the congestion, people began putting up "nodes", which you would "connect" to instead of operating unproto. I remember when APRS started, and it was also gobbling up a lot of precious bandwidth.

As the first common standard for AX.25 used AFSK (on a 16 KHz FM channel with 5KHz deviation), we were pretty much stuck at 1200 baud. To make the move to 9600, you had to switch to pure FM and ditch the AFSK layer. Pure FM is far less efficient than all of the modern modulation and coding techniques available today (in the DSP era). Once a bunch of people begin to use a particular frequency with a particular type of modulation, it's pretty hard to get them all to change, and you cannot really have more than one type of modulation on an FM channel, because the users of each type will see the opposing type as harmful interference.

I'm vaguely aware of a bunch of far more efficient coding techniques that some people are using today, LoRa is a good example of one, and you don't even need a ham license to use it. (See top level post in this thread.)

It’s still a thing, although I think VARA is more popular right now. IMHO the weak signal modes are getting more use because many newer hobbyists don’t have the money or space for high powered setups.

There are even people building BBS like functionality onto JS8Call.

Nice! I've been looking for such a thing lately. KISSLoraTNC [1] is another implementation: The radio control interfaces differ though.

[1] https://github.com/kc1awv/KISSLoRaTNC

Fuck. Yeah. Bud. I have a userland AX.25 stack I wrote as my pandemic project, and pairing that with MeshTNC will be a dream come true for me. My stack has a few utilities for bridging KISS interfaces together, and exposing APRS-IS as a KISS interface itself, so I can definitely see a lot of cool things coming from this. (If I were feeling extra spicy, I could bind, say, /bin/login to a specific SSID and get a shell into my home machine from across town!)

ax.25 tcp/ip is doable and Battle-tested :)

why reinvent tre whell?

Curious, which part is reinventing the wheel? If you're referring to the /bin/login hack, imagine instead, it's BBS software. (I had a lot of fun accessing my home computer from across town with a Psion 5mx and a Kenwood TH-D7! No practical value, just a good challenge.) As for the APRS-IS bridge, that would be quite useful for bridging relevant APRS-IS traffic within a given radius to LoRA, and vice-versa.

Also, I'm quite aware of IP over AX.25: I generally prefer to avoid that extra overhead for my intended applications. The stack I wrote exposes a BSD sockets-like interface to make it easy to make lots of software able to participate in the AX.25 fun, and even compiles on PowerPC Macintoshes for what it's worth.

Hah! I was literally about to open up my nascent userland AX.25 stack. Is yours open-source, or would you mind sharing? (My e-mail is in my profile.) I want to get something running on an ESP32-S3. My goal is to turn a Cardputer into a companion TNC console for my Kenwood TH-D74.

It is technically! I sent you an email. It's very POSIX-y, unfortunately, but if you were to make some sort of shim for file descriptors (big ask heh) you would be able to use it just fine.

Been quite a while since I've seen that word use outside of an LLM context.

I haven't encountered this term outside of stable diffusion and could probably do with a primer of what this is about.

LoRa is short for Long Range. It's basically about providing a physical layer for networks over radio using spread spectrum modulation. It's got long range, even for fairly low-power devices, but also low bandwidth. Think IoT stuff. LoRaWAN is a layer above in the network stack (MAC) and runs on top of LoRa.

TNC is "terminal node controller". They're kinda sorta like modems for radios. KISS TNC is a particular protocol for communication between a radio and a TNC.

In the LLM space, it's LoRA, not LoRa. Not great!

There are two competing hobbyist projects (Meshcore, Meshtastic), but they're so chaotic and random I don't think they have much ahead of them.

LoRaWAN will remain but this is different.

Reply on news.ycombinator.com