Sending IP Over Morse, Because Why Not?

Are you a ham radio operator? Have you ever felt the need to send IP over Morse? If you answered yes (or no) and don’t mind a few manual steps between converting network packets to and from and Morse code, [Folkert van Heusden]’s IP Over Morse project has you covered.

To send data, a network packet is first split into 5-bit words. Then those 32 different values are mapped to Morse characters (A-Z, 0-5, and a ‘/’ for termination), and the result is turned into an audio file ready to be sent over the airwaves, because no one is insane enough to want to do it by hand. To receive, the process is reversed. The GitHub repository for the project hosts the custom bits that [Folkert] created, should anyone wish to give it a shot.

The process of turning binary data into a completely different format sounds a bit like UUencoding, and is certainly an unconventional use of Morse code. Luckily, learning Morse code is easier than it’s ever been and it’s just waiting to be worked into novel projects, because why not?

29 thoughts on “Sending IP Over Morse, Because Why Not?

  1. Why not send IP over CW?

    Does the FCC require CW be morse? I don’t think so.

    Then the speed will be baseband frequency dependent. Could we transmit 50% of the baseband frequency?

    1. Actually, the FCC defines CW to be Morse. See FCC Part 97.3(c)(1). I’m not a fan of that definition, but they didn’t ask me.

      But the FCC gives ham radio operators wide latitude to use unconventional modulation schemes, and to invent their own modulation, as long as it’s not encoded for the purpose of obscuring meaning. So as long as you properly identified your transmissions, and took care to avoid interfering with other operators, you could use on/off keying of a CW transmitter in whatever scheme you desired for sending digital data.

      However, there are many other modulation schemes that are better suited to being received by machine. Hams have been handling IP traffic via packet radio for a long time.

      I believe this project is intentionally using Morse, in no small part because it’s an old technology. Samuel Morse was born in the 18th century, and did his telegraph work in the 19th. To have the code bearing his name used in the 21st century for carrying IP traffic may not be the most practical idea I’ve heard this week, but I can appreciate it, nonetheless.

      1. There is potentially another benefit to using CW/OOK for things like this, which is that it’s cheap. Yeah, a lot of old ham data radio stuff from the 80s and 90s uses AFSK, which is cheap, but AFSK also sucks these days, largely because the hardware is limited to FM. For any real data throughput, you’d be using GMSK or QPSK or some other more modern modulation, which would cost more in terms of either having an SDR or at minimum a SSB transceiver available to you. But a CW radio? I could build one with things I have in my junk box right now (minus, I think, a mixer IC — and even that’s not strictly necessary), and for only a few dollars more I could add proper amplification and better filtering to it. So for getting people into RF homebrew, doing notvel things with cheap CW gear is fantastic.

        There’s another project I’m interested in by KL4YFD called DCW, and it’s another cool use of ancient radio tech.

  2. Sounds a lot like uuencoding. I wonder if anybody ever wrote an ascii to morse converter.. It does not sound hard. Next thing will be using pbmplus to do jpeg -> pbm -> ascii -> morse and sending pictures.

    1. There have been many, many ASCII to Morse code converters. More than a few ham radio operators send high-speed Morse over the air by typing at a keyboard (though plenty more stick with old-fashioned telegraph keys, or slightly more modern electronic keyers that help a bit in automatically generating dits and dahs).

      As an example of an automatic ASCII to Morse translator, see . It takes the current CNN news feed from twitter and automatically converts it to an audio feed in Morse code. Various speeds are available. It can be used as an aid to those wishing to learn to receive Morse code, or a semi-practical way to get the news for those who already understand the code. Or it can be appreciated just as a curiosity.

    2. ASCII to morse is easy.

      But even before the home computer age there were morse to ASCII (and probably Baudot) converters. Not as common as morse keboards, but they existed. Certainly commercial/military circles, but in Ham Radio magazine in 1970 or 71 some ham described his morse to whatever (I can’t remember if he used an ASCII or Baudot printer) converter. Either just using transistors and diodes, or small scale logic.

      1. Seems you can decode anything symbol for symbol with enough diodes…. and enough voltage… I was shooting for HAD FOTW at age 10 when I tried to decode a custom keyboard to a 5 bit input port with a huge diode array, and didn’t think “Duh, voltage drop” so of course, a logic high didn’t make it through runs of 7 diodes.

        1. If anyone is interested, the CW to printer project was in the Nov 1971 issue of Ham Radio magazine, on page 8. it’s at that site.

          I think the real trick is determining what’s a dot and what’s a dash.

  3. TCP / IP : Transmission Control Protocol / Internet Protocol. Saying “Sending IP” is like saying “sending script” or “sending italics”. I can pick up the telephone and send Morse code over the telephone, as long as I can say “dot” and “dash”.

    1. Hi, it actually IS ip over morse: an ip packet is obtained from a virtual network interface and then encoded and transmitted as morse. So that way you can do udp, tcp, etc over ip over morse!

    1. No it is not. The actual IP packet is transmitted: an ip packet is obtained from a virtual network interface and then encoded and transmitted as morse. So that way you can do udp, tcp, etc over ip over morse!

  4. and how is your 3 way handshake establishing over morse?
    This must be something even 1000x times worse than trying to use DNS tunnel.
    Also don’t you think that you pollute the spectrum for other operators who try to send legitimate morse code with your junk?

    Impressive project tho.

    1. It isn’t a thing, except for hobby use. Ships at sea were likely one of the last holdouts, and they went to a more automated system at least two decades back.

      But it wasn’t obsoleted by the arrival of voice, which I assume you mean. CW could get through when voice couldn’t, and a CW transmitter was simpler and more efficient. Even after Teletype came along, CW saw plenty of use.

        1. Yes. Though it’s sent slow and only three or four characters and always repeating, so even if pilots need to decode it, it’s probably easy. I have no idea if there are decoders in airplanes to make it automatic. Perhaps they just keep those beacons in operation as backup.

    2. 1) Because dot dot dot dash dash dash dot dot dot can be transmitted over almost any sound or visual link, and is universally recognised. Cheap lipo LED torches from China have it as a mode.
      This carries well in difficult situations such as mountain rescue, and can be improvised with almost any equipment in an emergency. It’s never the first choice, but if your phone / radio has packed up, it’s potentially lifesaving.

      2) because sometime it’s useful for stuff like tapping walls in POW camps etc.

      But yeah, as a radio transmitting encoding for normal use, outdated, which is why it’s only used by hams and in unusual circumstances.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.