Post AcmHAXP7divNCYT200 by stephen@crabsin.space
(DIR) More posts by stephen@crabsin.space
(DIR) Post #AceCKInck0CRlY5KWO by stephen@crabsin.space
2023-12-09T22:49:07.341566Z
0 likes, 1 repeats
For anyone who's into APRS, I've been working on a modern alternative called CATS for a few months: https://cats.radio/It cleans up the bloat in the protocol. Packets are modular, consisting of whiskers, which can be mixed and matched depending on what you want to convey in your packet.Everything from the physical layer up has been completely revamped. We're running raw FSK for a massive 12dB coding gain over AFSK. Modern commercial transceiver chips makes it economical to do everything in FSK. These didn't exist when APRS was first developed.We're also using LDPC for forward error correction, and have boosted the data rate from 1200 baud (APRS) to 9600 baud.I've developed reference implementations of a Raspberry Pi I-Gate and mobile transceiver. Both are fully FOSS, and I'll also be selling units fairly shortly for $24 USD and $49 USD respectively.But that's all just a summary. I highly recommend checking out our site for more information! https://cats.radioHope you like it! #hamradio #amateurradio
(DIR) Post #Acemoyw0FEO9U0vtA0 by kr1st@mastodon.radio
2023-12-10T01:42:37Z
1 likes, 0 repeats
@stephen Excellent work! I really like what you are doing here and look forward to purchasing the hardware to play with it when it becomes available.Many moons ago (early 2000's) I was very involved in APRS and wrote my fair share of software for the platform, including software that ran on Russian handheld gaming computers that could provide a 900MHz link. I abandoned APRS because there was no forward movement, and frankly, became entirely boring because of it.I'm very excited to see this!
(DIR) Post #Acemro11IXNqUFcrmC by VE2UWY@mastodon.radio
2023-12-10T02:53:20Z
0 likes, 0 repeats
@stephen From semi-following along w/#APRS, a lot of the "bloat" was likely to do w/there being no public road-map, it all being in Bob's head. I have no wish to denigrate WB4APR but having the protocol largely conceived by & hacked on by one guy (even a very smart one) probably isn't the best way forward. Plus, as U said, 30 years ago, U just couldn't do some things in the hobby at reasonable cost.Please don't limit yourself to 70 cm, either - the protocol needs to work on any band. #hamr
(DIR) Post #Acemroufxbu3GrfJWi by stephen@crabsin.space
2023-12-10T05:37:38.632220Z
0 likes, 0 repeats
@VE2UWY The protocol itself works fine on any band. The standard frequency just happens to be set on 70cm but there's no reason you couldn't run it on any other frequency.
(DIR) Post #AcemsWWlnqvOXSoqNE by stephen@crabsin.space
2023-12-10T05:38:40.462927Z
0 likes, 0 repeats
@kr1st That's awesome! If you're interested in purchasing a board feel free to email me at cats@scd31.com. I know it says they're not for sale yet but I do have a small quantity of them ready to go out.
(DIR) Post #AcemtTXAQ1IbXmSUgS by julie@merida.hair
2023-12-10T00:34:27.260Z
1 likes, 0 repeats
@stephen@crabsin.space This is fantastic! I'm looking forward to playing with it. APRS is ... well, old.
(DIR) Post #Acenn3WyHihB74rzns by W1CDN@mastodon.radio
2023-12-10T00:51:45Z
0 likes, 0 repeats
@stephen I love APRS, but excited to see where this goes.Am I right that this all uses custom hardware, not existing transceivers?Also, how to interface with transceiver? KISS over Bluetooth possible?
(DIR) Post #Acenn4fs27PMey2cJU by stephen@crabsin.space
2023-12-10T05:48:53.607792Z
0 likes, 0 repeats
@W1CDN Yup that's correct. All the hardware is currently based around rf4463s, which are about $8 a piece. There's no reason you couldn't use another transceiver though, as long as you write the software for it - all it really needs is FSK.
(DIR) Post #AcfSgYGNiiVaWP7xPk by loic_fejoz@mastodon.gougere.fr
2023-12-10T09:42:41Z
1 likes, 0 repeats
@stephen I do not see SMS message whisker. Is that planned? Same question for weather related message?
(DIR) Post #AcfShA4JA3Qhlxlqs4 by stephen@crabsin.space
2023-12-10T13:27:14.241200Z
0 likes, 0 repeats
@loic_fejoz An SMS whisker seems too specific to bloat the protocol with. You could just use the normal comment whisker and (for example) use a specific destination whisker to tell your SMS gateway you want to relay your message onto the phone network. In fact, you could even put a phone number in the destination, since it's just UTF-8.A weather whisker is planned, but it's been on the back burner a bit while I figure out exactly what it should encode.
(DIR) Post #Acg8rDL63Fbwt56EdM by chfour@wetdry.world
2023-12-10T19:43:52Z
1 likes, 0 repeats
@stephen the quansheng has that funny beken transceiver ic in it that can also do 2fsk i feel like it should be possible to do something with that, it's stupid popular so i think it would help the protocol get more popular
(DIR) Post #Acg8wSv9XyYuf1T7SK by VE2UWY@mastodon.radio
2023-12-10T19:42:50Z
1 likes, 0 repeats
@stephen The 70-cm band is so underused outside the biggest cities … Use it or lose it. Our club started a siren test net a few years ago to help the city w/its outdoor warning sirens & we specifically chose the club's 70-cm repeater vs 2 m as it was idle 99.9% of the time. So good on you for trying to promote the band!#hamr
(DIR) Post #Acg9kQuvGr9gvRBgCu by w8emv@mastodon.radio
2023-12-09T23:47:24Z
0 likes, 0 repeats
@stephen Nice! Looking forward to seeing this in the field.I noticed that the CATS protocol uses a 5-byte encoded Unix timestamp, with an epoch of January 1, 1970. I *think* this gets you past the year 2038 problem (which it would be an issue if you only had 4 bytes) but you might want to comment in the standard on the date range supported and the conversion to/from a 64 bit (8 byte) timestamp.
(DIR) Post #Acg9kSHeBqvXACpLGa by chfour@wetdry.world
2023-12-10T19:35:32Z
1 likes, 0 repeats
@w8emv @stephen a 5 byte timestamp in seconds should last up to around year 34865 if i'm calculating it correctly (2^(5*8)/60/60/24/365)
(DIR) Post #AcgHjAzRnijEs9DPsW by M0YNG@mastodon.radio
2023-12-10T22:57:13Z
1 likes, 0 repeats
@stephen this looks excellent and I am super pleased to see the blackhole option on the mobile node. Is there a possibility of having ambiguous positions? Not just "error"? Specifically I'd like to run a base station but not give away the exact location of my front door...
(DIR) Post #AcgHjOmqVMy1f7QYBE by stephen@crabsin.space
2023-12-10T22:58:59.826189Z
0 likes, 0 repeats
@M0YNG Hmm, that seems like functionally it would be the same as error wouldn't it? In either case it's some ambiguity in the given values. Or did you want to encode that you're doing it deliberately and it's not a limitation of the GPS hardware?
(DIR) Post #AcgJD9vOQV6mD4BCpk by M0YNG@mastodon.radio
2023-12-10T23:01:54Z
0 likes, 0 repeats
@stephen exactly! My GPS might be perfectly located, or the location be set manually, BUT I don't want you to know specifically where I am.
(DIR) Post #AcgJDAl9K4VanaOXVQ by stephen@crabsin.space
2023-12-10T23:15:40.179580Z
0 likes, 0 repeats
@M0YNG So assuming you're talking about the functionality of the transceiver firmware and not a protocol-level change this is actually a pretty difficult problem to solve.How do you introduce the error? If you do it randomly, then with enough samples you'll be easy to triangulate. If you statically offset it then that's relatively easy to figure out too ("Oh, they're following the path of this road, just 500 meters north. They're clearly just offsetting their position.")If you're just doing it fixed then it's pretty easy as you can just fudge the coordinates yourself. The firmware would have to be updated to add the ability to specify deliberate error though.
(DIR) Post #AcgKZpWmWmsbZ32gM4 by M0YNG@mastodon.radio
2023-12-10T23:21:32Z
0 likes, 0 repeats
@stephen if it's any help, APRS does it by just dropping the number of digits in the location, giving a larger possible "square" (because every coordinate is still a square of possible locations, just an increasingly small square!)This is officially a "thing" in the protocol but it's not well known or supported.https://blog.aprs.fi/2011/01/position-ambiguity-support.htmlIt would be nice to say that a station is in this 3km by 3km square, rather than this 3m by 3m square. Especially if it's at my home.
(DIR) Post #AcgKZs4X4H47SIqfEu by stephen@crabsin.space
2023-12-10T23:30:56.781474Z
0 likes, 0 repeats
@M0YNG Fair, essentially locking yourself to the GPS grid is a pretty easy fix. Definitely may be worth implementing on the mobile transceiver!
(DIR) Post #AcgLFr0zO9t8L6D5bU by M0YNG@mastodon.radio
2023-12-10T23:35:13Z
0 likes, 0 repeats
@stephen I'm not that fussed about mobile, although an "ambiguous switch" might be nice for general "I'm out and about talk to me" stuff and a "precise switch" for "I am specifically here, on this hill, come talk to me because it's an interesting hill"I don't want my home igate to be accurate to the door, just knowing there is one in this town/neighbourhood should be enough?And I do want one at home if I can make a little handheld messaging client device / add that to the #beepy.
(DIR) Post #AcgLFrl4cokee1ltR2 by stephen@crabsin.space
2023-12-10T23:38:33.291012Z
0 likes, 0 repeats
@M0YNG A beepy with a CATS transceiver? Haha, we're on the same wavelength. I have one most of the way done: https://gitlab.scd31.com/cats/jumpySwapped out the screen for a much cheaper one that does full colour, and took the RPi so it all runs on a microcontroller.For my home I-gate I just didn't give it the precisely correct coordinates so it's off by a few hundred meters. But I could also add an `error` option so that you can say exactly how much error to report in the packet.
(DIR) Post #AcgNKpcMD8rXMVBpyK by M0YNG@mastodon.radio
2023-12-10T23:49:47Z
0 likes, 0 repeats
@stephen exciting! I'd like something I could "plug on" the beepy's broken out gpio but I've no idea how easy that would be.But to the point, the protocol is new so changes are expected, and I think position ambiguity as a "first class" feature would help adoption and encourage a wide variety of people who legitimately don't want to give away exactly where they live assurance that this is a protocol that cares about them, like the blackhole on mobile does.
(DIR) Post #AcgNKqlFxXZiuOMSTw by stephen@crabsin.space
2023-12-11T00:01:51.863544Z
0 likes, 0 repeats
@M0YNG I'm still not really clear on what the advantage is over just using the existing "error" field. It does seem like error in the clearest sense. Is it really worth bloating the protocol over?
(DIR) Post #AcgODRHWOS67oPU0ae by M0YNG@mastodon.radio
2023-12-11T00:08:25Z
0 likes, 0 repeats
@stephen I think clarity of intent. If error can be explicitly known to be either "inaccurate" due to GPS _OR_ "ambiguous" on purpose then maybe it's enough? If you think documentation can be sufficient and the implementation of both uses me mandatory?
(DIR) Post #AcgODSNaJOXfDVKMgC by stephen@crabsin.space
2023-12-11T00:11:43.735107Z
0 likes, 0 repeats
@M0YNG That's what error is supposed to be; I can probably make it clearer in the docs. I mean error only in the mathematical sense of deviation from the actual value, regardless of if intentional or not
(DIR) Post #AchrpFj4teJso2iPwG by vampirdaddy@chaos.social
2023-12-11T16:25:00Z
0 likes, 0 repeats
@stephen What are the actual licenses for the suggested hard- and software? I could not find any LICENSE file (or according content) in the git repository.I will have a look into the protocols later this week.
(DIR) Post #AchrpGxeIxZMdWXZI0 by stephen@crabsin.space
2023-12-11T17:18:12.892106Z
0 likes, 0 repeats
@vampirdaddy The software is MIT licensed - it's specified in the Cargo.toml file but I really should add a proper license file. Haven't decided on hardware yet but I'll be picking something similarly-permissive
(DIR) Post #AchruaW51fb4jaBD8K by vampirdaddy@chaos.social
2023-12-11T16:42:16Z
0 likes, 0 repeats
@chfour @w8emv @stephen as even 9600bit/s still is *slow* (i.e. below ~1000 bytes/sec), a (auto-negotiated) mode with reduced transfer might be an idea? Transfer of e.g. 18bit timestamp (good for 3 days), 6bit-standard-character encoding? The minimized data transfer would be converted to full standards (64bit epoch, UTF-8) on client/TNC side, but free up 25% at basically no cost (and still would be readable as plaintext).Untested, idea, will look into the protocols later this week.
(DIR) Post #AchrubJM4T0pCPEYwC by stephen@crabsin.space
2023-12-11T17:19:11.481890Z
0 likes, 0 repeats
@vampirdaddy @chfour @w8emv Negotiation is a hard problem to solve. It complicated things massively, and introduces state to what is currently a stateless protocol. I would prefer to stick to a single data rate.
(DIR) Post #AchrucP40jAmaOudTU by vampirdaddy@chaos.social
2023-12-11T17:00:04Z
1 likes, 0 repeats
@chfour @w8emv @stephen Okay, not overly compatible with the current protocol. I will inspect further and then maybe propose something more fact based. But: thanks @stephen for the idea & work!
(DIR) Post #Aci9kxZ0HSssUc6cYy by stephen@crabsin.space
2023-12-11T20:39:09.175330Z
0 likes, 0 repeats
@rozcakj So, CATS definitely has the potential to be significantly lower on power consumption, while beaconing more often. For a balloon I was envisioning beaconing every second (usually the norm is to do every minute on APRS). That said, the current hardware isn't very balloon-oriented (it's big and relatively heavy) but a proper balloon tracker is on my backlog
(DIR) Post #AciA5S6aRAML0XaNEG by vampirdaddy@chaos.social
2023-12-11T18:41:56Z
0 likes, 0 repeats
@stephen (your website(s) seem to be down?)Where are discussions supposed to happen?How fixed are the protocol specs?
(DIR) Post #AciA5SuvQ0cpWf8Zgu by stephen@crabsin.space
2023-12-11T20:42:51.118110Z
0 likes, 0 repeats
@vampirdaddy Sorry, website went down for ~30 minutes as my power went out, UPS died, and I had to switch things over to running on a generatorThe discussions thing I haven't nailed down but I think a Matrix channel (or space) would probably be a good idea.Right now the protocol specs are very loosey-goosey. I'm not afraid to create backwards-incompatible changes, because there are only a handful of transceivers out in the wild so making everyone update isn't too much of a hassle. I want the first iteration of the protocol to be as perfect as possible. As more transceivers are sent out and the protocol matures, I'm going to lock it down and from then on, only backwards compatible changes will be allowed.I'm open to adding additional whiskers to the protocol as needed, but I'm also fairly critical that the new whiskers must be useful. I'm doing what I can to prevent the protocol from bloating too much.
(DIR) Post #Acid5uQq16HfX2d0Qi by stephen@crabsin.space
2023-12-12T02:07:53.759674Z
0 likes, 0 repeats
@rozcakj Unrelated to CATS I'm currently involved in a balloon project where we're streaming real-time video to the ground. I wrote all the software and built the hardware - it's running on the same transceiver chip as CATS. Thought you might find that interesting!With a bare si4463 chip it should be possible to build a CATS beacon that's smaller and lighter than any APRS beacon on the market - and probably cheaper, too!
(DIR) Post #AcidDmF0M0mOReqRjk by vampirdaddy@chaos.social
2023-12-11T23:23:22Z
0 likes, 0 repeats
@stephen Did some sampling - a "reduced" character set would reduce e.g. a 500 character plain morseable text with 25 number-switches by... * baudot CITTY-2 (5bit plus 2x 5bit-switchchar per number) = 500chars + 50 switches = 344 byte => 32% * baudot 6bit (i.e. 1 bit for charset switch) = 375 byte => 25% * ASCII 7bit = 438 byte => 12% * UTF-8 8bit = 500 byte => 0% For short QSOs 6bit (many switches) is better. Are 25% worth the conversion hassle? Maybe @ max 1200bit/s (for HF)?
(DIR) Post #AcidDn0VVOmEoz4NmK by stephen@crabsin.space
2023-12-12T02:09:18.064901Z
0 likes, 0 repeats
@vampirdaddy I don't think this is something that CATS is for. One of the awesome features of CATS is that it supports full UTF-8, whereas APRS supports just plain ASCII (though in some situations it may accept full binary as well)
(DIR) Post #AcjrlUyscsAZHBEjA0 by IU5FHO@mastodon.radio
2023-12-12T12:32:34Z
1 likes, 0 repeats
@stephen Wow, I went through the protocol and this looks really nice!Does the standalone trx support something like KISS over bluetooth or similar for remote control?I'm really looking forward to be able to play with the boards once they are available! 🤩 PS: Is there a hashtag already for CATS? Would #CATSradio work? Cheers! :radio_tower:
(DIR) Post #Acjs0eN7zaVEkInqoy by stephen@crabsin.space
2023-12-12T16:29:43.424284Z
0 likes, 0 repeats
@IU5FHO The mobile transceiver doesn't have a bluetooth module or anything. It does have USB-C so it could theoretically support doing CATS over that, but the current firmware doesn't support it. I do like the idea of having some kind of standardized KISS-like protocol for communications between the actual transceiver and a host.I do actually have one mobile transceiver (and a few I-gate boards) available if you want to purchase either or both. Feel free to send me an email at cats@scd31.com!#CATSradio sounds great (:
(DIR) Post #Aclqs5qsk4PPjFzojg by kilroy_was_here@dobbs.town
2023-12-13T14:52:05Z
1 likes, 0 repeats
@stephen A digital ham radio protocol?That's not from when Turbo Pascal was en vogue?Nah, I gotta be hallucinating.
(DIR) Post #AcmHAXP7divNCYT200 by stephen@crabsin.space
2023-12-13T20:21:03.245936Z
0 likes, 0 repeats
@vampirdaddy Just following up on this - I created a matrix space for CATS at #cats:crabsin.space (:
(DIR) Post #AdGqZHWA2BCalgJSYi by Aloysius@3dp.chat
2023-12-28T11:29:14Z
1 likes, 0 repeats
Hey @stephen, very interested in where this goes in the future (thinking as a Amateur Rocketry payload). I didn't see it on the website, so I assume the answer is no, but have you looked at a GNURadio implementation?
(DIR) Post #AdGqcWFnhaXvL9CjSq by stephen@crabsin.space
2023-12-28T14:19:37.156669Z
0 likes, 0 repeats
@Aloysius I have no plans for one but I know one of my friends is working on a CATS implementation in GNURadio!
(DIR) Post #AdGqjB8IcywXrTs4bw by stephen@crabsin.space
2023-12-28T14:20:50.156602Z
0 likes, 0 repeats
@Aloysius Oh and I should mention I think amateur rocketry is a great idea. With such short packet times you could easily have a CATS transmitter that beacons its own location several times per second (though perhaps not on the official frequency of 430.500MHz)