11 ms·

NPR: New Packet Radio

227 points by wglb 8y ago

theWheez 8y ago
This is very cool! I had an idea for something like this a few years back, would love to experiment! Any idea where this falls in terms of broadcast regulations? From what I can recall in the US, data must be in the clear and not encrypted on ham frequencies, although I am not a licensed ham so can't say for sure. Love this!
- wglb 8y ago
  You are correct. Additionally, no commerce is allowed, so connecting this to the actual internet would be fraught with risk.
  - goda90 8y ago
    So encryption isn't allowed, but is there any way to authenticate information sent over packet radio? Open communication has plenty of uses, but some of those can be spoiled by bad actors impersonating others. Edit: I found this article on authentication that supposedly is fine, but I'm not licensed so I don't know how valid it is. https://rietta.com/blog/2009/08/17/authentication-without-encryption-for/ https://rietta.com/blog/2009/08/17/authentication-without-en...
    - nominatronic 8y ago
      There's no problem in principle with using public key cryptography to sign messages sent by ham radio. So authentication is possible although it's not common in practice. It's encrypting the content of messages that's not permitted.
      - kawfey 8y ago
        Legally speaking, it's "messages encoded for the purpose of obscuring their meaning" [0] that's illegal, which is fraught with conflicting interpretations and loopholes. [0] https://www.ecfr.gov/cgi-bin/text-idx?SID=f9e42063714bbba9b7bb83ef26e844ca&mc=true&node=pt47.5.97&rgn=div5#se47.5.97_1113 https://www.ecfr.gov/cgi-bin/text-idx?SID=f9e42063714bbba9b7...
    - vonuebelgarten 8y ago
      > So encryption isn't allowed, but is there any way to authenticate information sent over packet radio? Open communication has plenty of uses, but some of those can be spoiled by bad actors impersonating others. You are not allowed to encrypt data on ham bands (this seems to happen Worldwide, it is also true here in Brazil), but there is no restriction on use of cryptographic primitives for authentication. But some countries (France, AFAIK) also require the protocol specs to be openly available, what should not be a problem in this field. Digital signatures are Ok as long as the content of the message is not hidden (i.e. all listeners can receive and decode the information) and the transmission is identified (as usual in ham).
    - nickysielicki 8y ago
      Cryptography is not categorically illegal on the ham bands, and in the case of satellite/remote operation, it's encouraged (possibly required?) that you have some secure mechanism of controlling your radio. In the general case, there is nothing illegal about sending a PGP-signed email over Winlink, for example. But the message itself must be in the clear. It's just about the fear of foreign spies obfuscating their messages and Red Scare fearmongering. Of course, in a world with Tor and $30 TracFones with unlimited data, this concern is totally ridiculous. But it is what it is, for now.
      - jimktrains2 8y ago
        Well, also so that businesses don't monopolize the free ham bands instead of purchasing licenses for bands that allow encryption.
  - theWheez 8y ago
    Agh, damn government ruining all our fun! I kid, I understand the reasoning but I wish there was some longer-range frequency band with similar restrictions to 2.4GHz
    - packet_nerd 8y ago
      Isn't there a 900MHz band with similar restrictions to 2.4GHz? That would be a bit longer range.
  - billh 8y ago
    It's a little bit more relaxed than "no commerce allowed". Communications in which the operator has a pecuniary interest are prohibited[1]. But, personal communication with a business is allowed. Ordering a pizza via phone patch is allowed but soliciting orders for your pizza shop is not.[2] There are other issues that would make connecting to the internet problematic that are prohibited by §97.113. Things like cryptography, obscene or indecent words or language; false or deceptive messages ... 1. https://www.law.cornell.edu/cfr/text/47/97.113 https://www.law.cornell.edu/cfr/text/47/97.113 §97.113(A)(3) 2. http://www.arrl.org/phone-patch-guidelines http://www.arrl.org/phone-patch-guidelines
gonzoflip 8y ago
This is a very cool project! I would be concerned about unintentionally running afoul of FCC regulations since encryption is not allowed on US amateur frequencies, that means using something like SSH or loading an SSL webpage with this modem would be a violation. I would also be very concerned with the OS background processes that may use encryption by default.
- jessemillar 8y ago
  Excellent concerns! Do you have any resources you could link to on relevant FCC regulations? I'm a curious bugger.
  - diggan 8y ago
    I found this resource, hopefully it can point you to the right answer: https://rsaxvc.net/blog/2014/2/1/Encryption_and_Amateur_Radio.html https://rsaxvc.net/blog/2014/2/1/Encryption_and_Amateur_Radi...
  - gonzoflip 8y ago
    https://www.ecfr.gov/cgi-bin/text-idx?SID=d4b3c60d2d60000a147f885bdee88264&mc=true&node=pt47.5.97&rgn=div5#se47.5.97_1309 https://www.ecfr.gov/cgi-bin/text-idx?SID=d4b3c60d2d60000a14... specifically §97.309 (4) (b)
    - gorkish 8y ago
      This part of the regs have nothing to do with encryption. This is the so called "documented protocol" requirement. §97.113(4) is the relevant main rule that would govern encryption. There are various exceptions to this through the rest of the document such as spread spectrum, space stations, and telemetry/radio control all having some special case language.
      - gonzoflip 8y ago
        Well, I stand corrected. I had admittedly just googled it myself. I had known the ban on encryption was pretty accepted in the amateur radio community. Thanks for the correction.
chooseaname 8y ago
In my utopia, everyone would have access to something like this and the internet would be free from the chains of corporate greed. There's a lot to overcome (having everything be completely decentralized), but a person can dream...
- reaperducer 8y ago
  50 years ago the notion of everyone having their own computer (or several!) would have been laughable. Perhaps in the next 50 years people will have their own personal satellites in orbit and we will achieve your goal.
  - penagwin 8y ago
    Getting a satellite into space is ~50,000$ for a 1U cube sat when I looked around. You might be able to get NASA or somebody to send it up for free if it's educational and they have room. So it's getting cheaper for sure! While that's certainly not the cheapest, it's now within reach of small startups/passionate makers which is exciting!
    - 420codebro 8y ago
      $50K - wow that is really cheap to put something into orbit.
    - convolvatron 8y ago
      curious - is a cubesat enough of a footprint to gather enough solar energy to act as an RF relay until the orbit decays?
    - monocasa 8y ago
      More likely in 50-100 years is manufacturing capacity in space gets to the point that building a 1u satellite up there is way cheaper than sending from a deep gravity well like Earth's. I imagine at some point, rockets will only be used for sending people and other things that can't be transmitted or manufacturered.
bloomer 8y ago
Useful description from the documentation about what is new about this as amateur packet radio has a long history. The NPR protocol is designed to transfer IP data over radio links, in a bidirectional way (single frequency). This protocol is in the middle between old packet radio (AX.25) and HSMM-Hamnet with Wifi equipment. This protocol is designed by a HAM for HAMs. The project is 100% open source : specification, software, PCB. This solution is complementary to HSMM-Hamnet (which uses Wifi equipments), on lower frequencies 70cm). Radio links on 70cm band is much more robust to obstacles. The data-rate available is also much smaller, but yet useable. We can achieve several hundreds of kbps. The protocol is designed for “point to multipoint” topology, with 1 central relay (called Master) and several clients around.
henrikeh 8y ago
I’m so happy to see something like this! Communication engineering is a field which impacts everyday life all over the world and more people should have a chance to learn about it. Open hardware and software is an excellent resource.
tyingq 8y ago
Very cool. Might consider that having the terms "NPR" and "Radio" in the same phrase will confuse some people. Most will think of https://www.npr.org/ https://www.npr.org/
- ericcholis 8y ago
  Immediately the first thing that came to mind, I thought that NPR had created some new form of distribution.
- joncp 8y ago
  I bet they'll be hearing from NPR's lawyers
  - 24gttghh 8y ago
    Just call it GNPR?
- ComputerGuru 8y ago
  I think it was an intentional play on words.
  - tyingq 8y ago
    Looks like it may have originated in France, so it's plausible they've never heard of the other NPR.
- gcb0 8y ago
  author seems to be in france. even people in california barely knows what npr is...
nickysielicki 8y ago
I wonder if IP is really the right decision. AX25 isn't the bottleneck with existing packet radio, and because it's made for hams, it makes it easy to be sure that you're really within your license requirements. ie: identification. It's also difficult to accidentally send encrypted messages over AX25, whereas on IP you have to really watch Wireshark and make sure you're not doing something wrong. It's also not as chatty, which is a real concern when you have a mesh network and someone is hearing you better than you're hearing them --- it could get really messy without coordination. That being said, this is a very exciting project. I'd love to see packet radio make some progress since the 80s (outside of just repurposing 802.11 gear). We need the ARRL to stop wasting their time on a watered down HOA bill that they've been working on for like a decade, and start talking to the FCC about modernizing regulations for the 21st century. Two things that come to mind are that the spread spectrum/transmit bandwith restrictions are outdated, and encryption requirements need to be formally relaxed so that common digital communication like SSH or HTTPS can be supported to an extent -- I'm not saying I want the ham bands filled with encrypted traffic, but I'm fine with the small community of packet radio enthusiasts being able to remote into a computer from afar.
- Aloha 8y ago
  AX.25 was never designed to create very large networks with, it also doesn't integrate well into a modern world. Hams have been using IP technology for a very long time too, this isn't new.
  - nickysielicki 8y ago
    Where do you live? VHF packet radio is extremely active in Wisconsin, and none of it is IP based. Outside of the guys in the northwest doing 802.11 with external amplifiers (hamwan), who else is using IP?
- finnthehuman 8y ago
  I’m ok with cryptographic authentication of unencrypted messages (eg https with null encryption). But as soon as encrypted payloads happen, everyone will be doing way more than the “ssh where no other radio transport is available” use case.
rsync 8y ago
Forgive my ignorance ... What is the range of HSMM/Hamnet ? Can you have a "Hamnet BBS" and if so, do you need a separate radio for each "line" ? Or can the BBS have a single radio that shares incoming connections on different sub-frequencies of (whatever band you're using) ?
- jrockway 8y ago
  This is on the 70cm band so basically line of sight. The curvature of the Earth is the limiting factor (but you can always bounce it off something reflective in the sky; see meteor scatter, airplane scatter, moonbounce, etc.)
- int_19h 8y ago
  You don't need a separate radio for each "line". It's not like a BBS running directly on your modem - you have multiplexing and packet routing, so a single physical channel can accommodate many logical connections.
zamadatix 8y ago
I think IP is a better fit for the 900mhz range though I'm no expert in the radio space. It's a shame 802.11ah never took off in hardware.
- int_19h 8y ago
  You get higher bitrates at higher frequencies, but you also get more attenuation and hence shorter distances.
johnklos 8y ago
Ummm... Could you pick a name that isn't already taken by something we all already know?
- saul_goodman 8y ago
  ...Because everyone is from Murica and knows our radio snow names.
philpem 8y ago
I love the concept - I've been working on something similar myself using an MMDVM (with custom programming) and a pair of Icom mobile transceivers. My only comments from reading the protocol spec -- and I need to emphasize, this is one British ham's opinion and is worth precisely what you paid for it, i.e. nowt :) - Data whitening is a great addition. Too few ham protocols use this, and as a result it's a pain to find suitable transceivers (other than ham-spec ones) for them on the used market. - Forward error correction scheme looks a bit weak. Nice compared to most ham protocols though (which sometimes just have a CRC). The main thing is, if you get a burst error which straddles two blocks, you lose block N's CRC and block N+1's first few data bytes -- which would be uncorrectable. - It'd be nice to see some narrowband modes, maybe 8Kbps in a 12k5 narrow channel or 16Kbps in a 25kHz channel. With 45kHz deviation 360kbps 4GMSK (Mode 22), you're looking at 270kHz bandwidth. That's pretty darn wide! So wide, in fact, it wouldn't be reasonably usable in the UK if you wanted to adhere to the RSGB Bandplan. I think it should be possible to get some pretty big improvements in error correction by using a better error-correction code (POCSAG's (31,21)BCH code would be a good starting point) and employing bitwise interleaving. The idea is -- you'd add the error correction first, then interleave the bits. This means that when you deinterleave things at the other end, any burst errors will become single-bit errors in multiple codewords. Single-bit errors are much easier to fix than bursts! 8:1 interleaving with POCSAG's error-correction scheme would give a 21-byte packet size inside a 32-byte transmitted block (giving ~67% efficiency or ~33% overhead). It'd be able to correct up to a 16-bit error burst thanks to (31,21)BCH's ability to correct 2 erroneous bits. This is, actually the same method used by the FLEX-TD paging protocol (documented in ARIB standard RCR STD-43) :) 73's, de M0OFX :) EDIT: Forgot to show my working, oops! The GFSK occupied bandwidth formula I used was from here: https://www.silabs.com/community/wireless/proprietary/knowledge-base.entry.html/2015/02/17/calculation_of_theo-S9wI https://www.silabs.com/community/wireless/proprietary/knowle... (bear in mind GMSK is GFSK with the requirement that the modulation be phase-synchronous, i.e. data changes only occur when the IF/RF is at zero phase) I just calculated the OBW of Mode 24 (4GMSK 500ksym/sec). 750kHz! That'd be 30 of the 25kHz Digital Comms channels... yikes. I'd expect complaints from local hams about the QRM.
kerouanton 8y ago
Nice project, but why are all hardware guys designing projects without ANY security? Page 14, the guide explains the connection to the modem is possible with Telnet with no password... That's a common issue in IoT and embedded systems, it seems, so I'm wondering if there is a reason why hardware developers just fail at security ?
- microcolonel 8y ago
  This protocol is designed for HAM, and for the most part you can not make encrypted transmissions over amateur radio.
  - kerouanton 8y ago
    Maybe (I'm also a HAM licensee) but making a device available on a IP network without any authentication and weak protocols such as telnet doesn't justify this.
    - microcolonel 8y ago
      I don't think there would be any issue with signing your datagrams, you just can't obscure the content of the communication or its purpose. So yea on authenticity and access control, nay on transmitting ciphertext.
- bjt2n3904 8y ago
  To chime in, the hardware build looks like a microcontroller connected to an SPI based ethernet chip. SSH is likely not an option here.
- F4HDK 8y ago
  I'm F4HDK, the creator of these NPR modem. The unprotected telnet access is only for local management, especially when you connect only 1 PC to one NPR Client modem. If you need security for accessing configuration/management features of the modem, then you can deactivate the telnet server inside the modem, and put a R-Pi dedicated for that feature, which will be plugged locally to the modem via USB. Then you access the R-Pi via SSH. This is explained in the "advanced user guide". But of course trafic will be kept unecrypted, due to amateur-radio regulations. (telnet is only for management-configuration).

227 points by wglb 8y ago

This is very cool! I had an idea for something like this a few years back, would love to experiment! Any idea where this falls in terms of broadcast regulations? From what I can recall in the US, data must be in the clear and not encrypted on ham frequencies, although I am not a licensed ham so can't say for sure. Love this!

wglb 8y ago
You are correct. Additionally, no commerce is allowed, so connecting this to the actual internet would be fraught with risk.
- goda90 8y ago
  So encryption isn't allowed, but is there any way to authenticate information sent over packet radio? Open communication has plenty of uses, but some of those can be spoiled by bad actors impersonating others. Edit: I found this article on authentication that supposedly is fine, but I'm not licensed so I don't know how valid it is. https://rietta.com/blog/2009/08/17/authentication-without-encryption-for/ https://rietta.com/blog/2009/08/17/authentication-without-en...
  - nominatronic 8y ago
    There's no problem in principle with using public key cryptography to sign messages sent by ham radio. So authentication is possible although it's not common in practice. It's encrypting the content of messages that's not permitted.
    - kawfey 8y ago
      Legally speaking, it's "messages encoded for the purpose of obscuring their meaning" [0] that's illegal, which is fraught with conflicting interpretations and loopholes. [0] https://www.ecfr.gov/cgi-bin/text-idx?SID=f9e42063714bbba9b7bb83ef26e844ca&mc=true&node=pt47.5.97&rgn=div5#se47.5.97_1113 https://www.ecfr.gov/cgi-bin/text-idx?SID=f9e42063714bbba9b7...
  - vonuebelgarten 8y ago
    > So encryption isn't allowed, but is there any way to authenticate information sent over packet radio? Open communication has plenty of uses, but some of those can be spoiled by bad actors impersonating others. You are not allowed to encrypt data on ham bands (this seems to happen Worldwide, it is also true here in Brazil), but there is no restriction on use of cryptographic primitives for authentication. But some countries (France, AFAIK) also require the protocol specs to be openly available, what should not be a problem in this field. Digital signatures are Ok as long as the content of the message is not hidden (i.e. all listeners can receive and decode the information) and the transmission is identified (as usual in ham).
  - nickysielicki 8y ago
    Cryptography is not categorically illegal on the ham bands, and in the case of satellite/remote operation, it's encouraged (possibly required?) that you have some secure mechanism of controlling your radio. In the general case, there is nothing illegal about sending a PGP-signed email over Winlink, for example. But the message itself must be in the clear. It's just about the fear of foreign spies obfuscating their messages and Red Scare fearmongering. Of course, in a world with Tor and $30 TracFones with unlimited data, this concern is totally ridiculous. But it is what it is, for now.
    - jimktrains2 8y ago
      Well, also so that businesses don't monopolize the free ham bands instead of purchasing licenses for bands that allow encryption.
- theWheez 8y ago
  Agh, damn government ruining all our fun! I kid, I understand the reasoning but I wish there was some longer-range frequency band with similar restrictions to 2.4GHz
  - packet_nerd 8y ago
    Isn't there a 900MHz band with similar restrictions to 2.4GHz? That would be a bit longer range.
- billh 8y ago
  It's a little bit more relaxed than "no commerce allowed". Communications in which the operator has a pecuniary interest are prohibited[1]. But, personal communication with a business is allowed. Ordering a pizza via phone patch is allowed but soliciting orders for your pizza shop is not.[2] There are other issues that would make connecting to the internet problematic that are prohibited by §97.113. Things like cryptography, obscene or indecent words or language; false or deceptive messages ... 1. https://www.law.cornell.edu/cfr/text/47/97.113 https://www.law.cornell.edu/cfr/text/47/97.113 §97.113(A)(3) 2. http://www.arrl.org/phone-patch-guidelines http://www.arrl.org/phone-patch-guidelines

gonzoflip 8y ago

This is a very cool project! I would be concerned about unintentionally running afoul of FCC regulations since encryption is not allowed on US amateur frequencies, that means using something like SSH or loading an SSL webpage with this modem would be a violation. I would also be very concerned with the OS background processes that may use encryption by default.

jessemillar 8y ago
Excellent concerns! Do you have any resources you could link to on relevant FCC regulations? I'm a curious bugger.
- diggan 8y ago
  I found this resource, hopefully it can point you to the right answer: https://rsaxvc.net/blog/2014/2/1/Encryption_and_Amateur_Radio.html https://rsaxvc.net/blog/2014/2/1/Encryption_and_Amateur_Radi...
- gonzoflip 8y ago
  https://www.ecfr.gov/cgi-bin/text-idx?SID=d4b3c60d2d60000a147f885bdee88264&mc=true&node=pt47.5.97&rgn=div5#se47.5.97_1309 https://www.ecfr.gov/cgi-bin/text-idx?SID=d4b3c60d2d60000a14... specifically §97.309 (4) (b)
  - gorkish 8y ago
    This part of the regs have nothing to do with encryption. This is the so called "documented protocol" requirement. §97.113(4) is the relevant main rule that would govern encryption. There are various exceptions to this through the rest of the document such as spread spectrum, space stations, and telemetry/radio control all having some special case language.
    - gonzoflip 8y ago
      Well, I stand corrected. I had admittedly just googled it myself. I had known the ban on encryption was pretty accepted in the amateur radio community. Thanks for the correction.

chooseaname 8y ago

In my utopia, everyone would have access to something like this and the internet would be free from the chains of corporate greed. There's a lot to overcome (having everything be completely decentralized), but a person can dream...

reaperducer 8y ago
50 years ago the notion of everyone having their own computer (or several!) would have been laughable. Perhaps in the next 50 years people will have their own personal satellites in orbit and we will achieve your goal.
- penagwin 8y ago
  Getting a satellite into space is ~50,000$ for a 1U cube sat when I looked around. You might be able to get NASA or somebody to send it up for free if it's educational and they have room. So it's getting cheaper for sure! While that's certainly not the cheapest, it's now within reach of small startups/passionate makers which is exciting!
  - 420codebro 8y ago
    $50K - wow that is really cheap to put something into orbit.
  - convolvatron 8y ago
    curious - is a cubesat enough of a footprint to gather enough solar energy to act as an RF relay until the orbit decays?
  - monocasa 8y ago
    More likely in 50-100 years is manufacturing capacity in space gets to the point that building a 1u satellite up there is way cheaper than sending from a deep gravity well like Earth's. I imagine at some point, rockets will only be used for sending people and other things that can't be transmitted or manufacturered.

bloomer 8y ago

Useful description from the documentation about what is new about this as amateur packet radio has a long history. The NPR protocol is designed to transfer IP data over radio links, in a bidirectional way (single frequency). This protocol is in the middle between old packet radio (AX.25) and HSMM-Hamnet with Wifi equipment. This protocol is designed by a HAM for HAMs. The project is 100% open source : specification, software, PCB. This solution is complementary to HSMM-Hamnet (which uses Wifi equipments), on lower frequencies 70cm). Radio links on 70cm band is much more robust to obstacles. The data-rate available is also much smaller, but yet useable. We can achieve several hundreds of kbps. The protocol is designed for “point to multipoint” topology, with 1 central relay (called Master) and several clients around.

henrikeh 8y ago

I’m so happy to see something like this! Communication engineering is a field which impacts everyday life all over the world and more people should have a chance to learn about it. Open hardware and software is an excellent resource.

tyingq 8y ago

Very cool. Might consider that having the terms "NPR" and "Radio" in the same phrase will confuse some people. Most will think of https://www.npr.org/ https://www.npr.org/

ericcholis 8y ago
Immediately the first thing that came to mind, I thought that NPR had created some new form of distribution.
joncp 8y ago
I bet they'll be hearing from NPR's lawyers
- 24gttghh 8y ago
  Just call it GNPR?
ComputerGuru 8y ago
I think it was an intentional play on words.
- tyingq 8y ago
  Looks like it may have originated in France, so it's plausible they've never heard of the other NPR.
gcb0 8y ago
author seems to be in france. even people in california barely knows what npr is...

nickysielicki 8y ago

I wonder if IP is really the right decision. AX25 isn't the bottleneck with existing packet radio, and because it's made for hams, it makes it easy to be sure that you're really within your license requirements. ie: identification. It's also difficult to accidentally send encrypted messages over AX25, whereas on IP you have to really watch Wireshark and make sure you're not doing something wrong. It's also not as chatty, which is a real concern when you have a mesh network and someone is hearing you better than you're hearing them --- it could get really messy without coordination. That being said, this is a very exciting project. I'd love to see packet radio make some progress since the 80s (outside of just repurposing 802.11 gear). We need the ARRL to stop wasting their time on a watered down HOA bill that they've been working on for like a decade, and start talking to the FCC about modernizing regulations for the 21st century. Two things that come to mind are that the spread spectrum/transmit bandwith restrictions are outdated, and encryption requirements need to be formally relaxed so that common digital communication like SSH or HTTPS can be supported to an extent -- I'm not saying I want the ham bands filled with encrypted traffic, but I'm fine with the small community of packet radio enthusiasts being able to remote into a computer from afar.

Aloha 8y ago
AX.25 was never designed to create very large networks with, it also doesn't integrate well into a modern world. Hams have been using IP technology for a very long time too, this isn't new.
- nickysielicki 8y ago
  Where do you live? VHF packet radio is extremely active in Wisconsin, and none of it is IP based. Outside of the guys in the northwest doing 802.11 with external amplifiers (hamwan), who else is using IP?
finnthehuman 8y ago
I’m ok with cryptographic authentication of unencrypted messages (eg https with null encryption). But as soon as encrypted payloads happen, everyone will be doing way more than the “ssh where no other radio transport is available” use case.

rsync 8y ago

Forgive my ignorance ... What is the range of HSMM/Hamnet ? Can you have a "Hamnet BBS" and if so, do you need a separate radio for each "line" ? Or can the BBS have a single radio that shares incoming connections on different sub-frequencies of (whatever band you're using) ?

jrockway 8y ago
This is on the 70cm band so basically line of sight. The curvature of the Earth is the limiting factor (but you can always bounce it off something reflective in the sky; see meteor scatter, airplane scatter, moonbounce, etc.)
int_19h 8y ago
You don't need a separate radio for each "line". It's not like a BBS running directly on your modem - you have multiplexing and packet routing, so a single physical channel can accommodate many logical connections.

zamadatix 8y ago

I think IP is a better fit for the 900mhz range though I'm no expert in the radio space. It's a shame 802.11ah never took off in hardware.

int_19h 8y ago
You get higher bitrates at higher frequencies, but you also get more attenuation and hence shorter distances.

johnklos 8y ago

Ummm... Could you pick a name that isn't already taken by something we all already know?

saul_goodman 8y ago
...Because everyone is from Murica and knows our radio snow names.

philpem 8y ago

I love the concept - I've been working on something similar myself using an MMDVM (with custom programming) and a pair of Icom mobile transceivers. My only comments from reading the protocol spec -- and I need to emphasize, this is one British ham's opinion and is worth precisely what you paid for it, i.e. nowt :) - Data whitening is a great addition. Too few ham protocols use this, and as a result it's a pain to find suitable transceivers (other than ham-spec ones) for them on the used market. - Forward error correction scheme looks a bit weak. Nice compared to most ham protocols though (which sometimes just have a CRC). The main thing is, if you get a burst error which straddles two blocks, you lose block N's CRC and block N+1's first few data bytes -- which would be uncorrectable. - It'd be nice to see some narrowband modes, maybe 8Kbps in a 12k5 narrow channel or 16Kbps in a 25kHz channel. With 45kHz deviation 360kbps 4GMSK (Mode 22), you're looking at 270kHz bandwidth. That's pretty darn wide! So wide, in fact, it wouldn't be reasonably usable in the UK if you wanted to adhere to the RSGB Bandplan. I think it should be possible to get some pretty big improvements in error correction by using a better error-correction code (POCSAG's (31,21)BCH code would be a good starting point) and employing bitwise interleaving. The idea is -- you'd add the error correction first, then interleave the bits. This means that when you deinterleave things at the other end, any burst errors will become single-bit errors in multiple codewords. Single-bit errors are much easier to fix than bursts! 8:1 interleaving with POCSAG's error-correction scheme would give a 21-byte packet size inside a 32-byte transmitted block (giving ~67% efficiency or ~33% overhead). It'd be able to correct up to a 16-bit error burst thanks to (31,21)BCH's ability to correct 2 erroneous bits. This is, actually the same method used by the FLEX-TD paging protocol (documented in ARIB standard RCR STD-43) :) 73's, de M0OFX :) EDIT: Forgot to show my working, oops! The GFSK occupied bandwidth formula I used was from here: https://www.silabs.com/community/wireless/proprietary/knowledge-base.entry.html/2015/02/17/calculation_of_theo-S9wI https://www.silabs.com/community/wireless/proprietary/knowle... (bear in mind GMSK is GFSK with the requirement that the modulation be phase-synchronous, i.e. data changes only occur when the IF/RF is at zero phase) I just calculated the OBW of Mode 24 (4GMSK 500ksym/sec). 750kHz! That'd be 30 of the 25kHz Digital Comms channels... yikes. I'd expect complaints from local hams about the QRM.

kerouanton 8y ago

Nice project, but why are all hardware guys designing projects without ANY security? Page 14, the guide explains the connection to the modem is possible with Telnet with no password... That's a common issue in IoT and embedded systems, it seems, so I'm wondering if there is a reason why hardware developers just fail at security ?

microcolonel 8y ago
This protocol is designed for HAM, and for the most part you can not make encrypted transmissions over amateur radio.
- kerouanton 8y ago
  Maybe (I'm also a HAM licensee) but making a device available on a IP network without any authentication and weak protocols such as telnet doesn't justify this.
  - microcolonel 8y ago
    I don't think there would be any issue with signing your datagrams, you just can't obscure the content of the communication or its purpose. So yea on authenticity and access control, nay on transmitting ciphertext.
bjt2n3904 8y ago
To chime in, the hardware build looks like a microcontroller connected to an SPI based ethernet chip. SSH is likely not an option here.
F4HDK 8y ago
I'm F4HDK, the creator of these NPR modem. The unprotected telnet access is only for local management, especially when you connect only 1 PC to one NPR Client modem. If you need security for accessing configuration/management features of the modem, then you can deactivate the telnet server inside the modem, and put a R-Pi dedicated for that feature, which will be plugged locally to the modem via USB. Then you access the R-Pi via SSH. This is explained in the "advanced user guide". But of course trafic will be kept unecrypted, due to amateur-radio regulations. (telnet is only for management-configuration).