Scenes from the last operational Morse-code radio station in North America | Hacker News Reader

Scenes from the last operational Morse-code radio station in North America

theatlantic.com

149 pointsfortran7774 comments

This reminds me of the fantastic Connections Museum in Seattle [0] that maintains historical telephone switching equipment in (mostly) working condition with an all-volunteer staff. When I first visited, I was surprised to see that all the volunteers were about half the age I expected! You can see some of them on the YouTube channel [1].

[0] https://www.telcomhistory.org/connections-museum-seattle/

[1] https://youtube.com/@ConnectionsMuseum

Whenever I see this my joy is tempered by mourning for the loss of the living computer museum to accountants eating at the corpse of Paul Allen.

The living computer museum was a great place to visit, but it is not the accounts eating away. It is his sister.

I’m with you! It was my dream to retire and become one of the old guys in lab coats repairing stuff at Living Computers. Maybe they’ll get that sorted out and I’ll still get a chance.

Not likely. As mentioned in a parallel comment it’s partially that Jody Allen, his sister, doesn’t value the geek culture in the least and partially that the owner, Vulcan, is run by MBA accountants that are finally unshackled by their eccentric nerd funder. Between them they’re seeking to put the nerd passions of Allen to death and focus squarely on playing money managers. There appears to be little respecting of his legacy or wishes in his estate management and a lot of avarice.

This has been an unexpected and enjoyable rabbit hole for me. Thanks for posting.

KPH receiving station is 45 miles north of San Francisco, without bridge traffic you can make it in a little over an hour.

My father-in-law was an engineer in the navy and loved seeing the morse & radio equipment.

But the trip was fun for all the family; the driveway up to the reviving station building is lined with Monterey cypress trees, which have grown into a tunnel. [0] It’s a beautiful scene, my wife and mother-in-law were really taken aback.

[0] https://www.nps.gov/places/point-reyes-cypress-tree-tunnel.h...

The Maritime Radio Historical Society, KPH's website:

https://www.radiomarine.org/

The volunteers fire up the transmitters every Saturday; the schedule is on the website.

I patented a scheme where cell phones could communicate with morse code. The idea was to put a rocker switch on the phone, one side for dit and the other for dah. Then, you could text people without needing to look at the display. You could be texting people at a meeting with your hand in your pocket. The phone would convert the morse to text to transmit to your penpal.

You could also opt to receive morse by having the phone vibrate in morse so you could stealthily receive texts, too!

A related one-handed, eyes-off way to text fast is with the Twiddler chording keyboard, used by the wearable computing people, circa turn of the century:

https://en.wikipedia.org/wiki/Twiddler

Of course that, too, requires special training to use. And harder to integrate into a smartphone design than 2 side buttons. (Maybe turn the back of an iPhone-like form factor into a chording keypad, with physical buttons or some other tactile affordance?)

Brad Rhodes at MIT would have a Twiddler in a pocket, a small Private Eye brand HUD on the corner of his hat brim, and a lunchbox-sized PC on a shoulder strap. He could be typing and looking up information while he was talking with someone or in a meeting, without them necessarily being able to tell.

(Though, if someone did notice use of the wearable computer, they might not say anything anyway, since -- as he eventually realized -- some people who hadn't gotten the demo thought that the box with the cables coming out of it and going into clothing might be a medical device.)

>>used by the wearable computing people, circa turn of the century

One of the things they discovered, is voice command really wasn't suited to a wearable as it was distracting in meetings, susceptible to background noise, made you look like you were talking to yourself, etc.

One of my favorite BoFH style stories was of someone that worked in an office where each PC was equipped with dragon naturally speaking. Apparently, you could walk thru the office yelling "format c:" <pause> "yes"....

On older nokia phones un 2000-2004 (when I had them), I could text as long as I know I am on the typing screen. I knew pressing 2 three times gives you c, one 3 gives you D, # changes case, 0 gives you space. Two zeros gives you dot.

I remember texting a lot while phone is under the table, or simply in pocket.

This (T9) was incidentally a great accessibility feature for folks with limited vision. My mom is mostly blind and was able to text my brother and I without her giant magnifying glass, often faster than we could reply. Today, for better or worse, she largely uses speech to text.

What GP describes is not T9. The thing known as T9 is a predictive algorithm based on a compressed dictionary that generates words from single presses on each key. If you triple-tap 2 with T9 you get "cab", not C.

With normal pre t9, typing cab was a bit of a pain, 222222 if I remember the “end of character” code right.

It was more accurate than typifying on an iPhone with predictive text though.

I stand corrected on the name. My mom used what they described.

Yes and every middle- and high-school kid with a "dumb" cell phone could text without looking, holding the phone under the desk or in a coat pocket and using T9 code.

You...patented this?

Were you trying to ensure no one ever did this?

I patented a few things so I could be officially called an "inventor" ! I know, I know, the sin of pride.

Here's a list:

    Patent 7831208 Wireless mobile phone having encoded data entry facilities
    Patent 7812993 Lossy method for compressing images and video
    Patent 7711748 Method and apparatus for simplified access to online services
    Patent 7028033 Method and apparatus for simplified access to online services
    Patent 6897977 Lossy method for compressing pictures and video
    Patent 6850782 Wireless device with vibrational communication capabilities
    Patent 6657647 Controlling the order in which content is displayed in a browser
    Patent 6418323 Wireless mobile phone with Morse code and related capabilities

I could have patented many of D's innovative features, but chose not to. Many have been incorporated into other languages without attribution, but although that's a bit irritating I did sign up for that.

You could consider joining the Open Innovation Network. Had you patented the D features and licensed them freely as a member of the OIN, in return, you'd effectively get immunity from patent claims against you from any companies who use the D language!

[1]: https://openinventionnetwork.com/

What prototypes did you build?

The startup I was with did some mockups.

You can do that even now! Gboard has had Morse as an option for years.

But almost always with these things it's not iambic/semi-automatic, which it should be to be somewhat useful.

I have been looking for morse code keyboards for smartphones and I have only found one that was good, and it was an early iPhone thing that hasn't worked on modern iPhones for years.

I call prior art - documentary link below:

https://youtu.be/oye9AmOdsZc

Watch from around 7:53 for around 1 minute.

Cool. But the boss could subscribe too. Is it encrypted? Are there standards for morse code encryption?

The morse app would just convert it to/from text, and then the text would be sent by whatever app you would be keyboarding input to.

There’s a vocal group who want to bring back the abacus. Sometimes you should just let go. .

I always found it disappointing nobody tried making a phone with 5 programmable buttons on each side. It would make an excellent chording keyboard setup if the phone was narrow enough to grasp.

Direct link to photographer's site with photos: https://www.annhermesphoto.com/radio-squirrels

Tangentially related, Hyundai has a nice little tribute to Morse code in their Ioniq 5 and 6.

On the steering wheel, where they traditionally placed the stylized "H" Hyundai logo, there are just four dots:

• • • •

Which is Morse code for the letter "H"!

It makes me happy to know these guys are out there. I like that they've got a thing they care about and are dedicated to.

Sounds like the ship this station communicates with, the S.S. Jeremiah O'Brien, is a restored Liberty ship used for day cruises. I'd love to know about the relationship between the (presumably) WWII naval history nerds who restored that ship and the Morse code nerds who operate this station.

The S.S. Red Oak Victory, a Victory ship built in Richmond, California, is now a museum ship berthed near where it was built. Its radio room, operated by an amateur radio club, is open on Sundays.

https://www.qsl.net/redoakarc/

Does anyone know how the telegraph lines were architected?

Did you have segments of shared stations, where a message sent from one station was echoed to all the connected stations, with one responsible for forwarding messages across segments?

Just curious what the path of a message from, say, New York to Chicago was.

For big operations such as Western Union, it was store and forward. Links were point to point links with messages hand-copied to paper at relay stations. Paper messages were sorted by hand, and re-sent by outgoing operators. Large buildings with large staffs were required.[1] KPH was such a relay station, where messages went from land lines to radio and back.

Low-traffic lines and railroads tended to have many stations on one wire. Similarly, ships could talk to each other by radio. To get into the interconnected system, you had to go through a gateway to a commercial network, and there would be charges incurred at that point.

[1] https://www.wondersofworldengineering.com/telegraph_office.h...

Some messages were sent through relays, or the lines could be tapped, I wonder if any telegraph operators encrypted their messages using a shared key.

Telegraphic ciphers and codes were common, both to save money through compression, and as it was well known that many people would be reading and re-transmitting the message as it worked its way through the system.

There were various standardized codebooks published for when the message was not secret and compression was the only consideration:

https://en.wikipedia.org/wiki/Commercial_code_(communication...

Frequent telegraph users with an additional need for secrecy, such as investment bankers, developed similar proprietary in-house codebooks.

You raise an interesting question, but note that this article is about KPH, a radio telegraph station, not land-line telegraphy.

I don't know much about your actual question though.... I know that there were "repeaters" in the lines, basically a relay with a local battery, on longer lines. I know railroads had their own network of telegraph circuits between stations, with poles along the track right-of-way. I know that the first trans-Atlantic telegraph wires was what led to research into basic transmission line theory, since the first cables only allowed pitifully slow signalling and the reason why was not well understood.

If you look at old photographs of rail road right-of-way, you see 6 or 8 wires or more, so they obviously were running multiple telegraph circuits along the line.

In passing, I will note the trivia that land-line Morse (the original) and international radio-telegraph Morse is different in a few characters, because some of the character "clacks" become difficult to distinguish as "beeps" under noisy conditions.

You have successfully nerd-sniped me today, well played.

If you're interested in this topic, here's a book to read:

https://en.m.wikipedia.org/wiki/The_Victorian_Internet

It's been twenty+ years since I last owned a copy, but it was one of my favorite books for a while. The bulk of the material is about pointing out amazing parallels between behaviors on the early internet and century-previous telegraph / wireless networks, but it also discusses physical and network protocols - not, perhaps, in huge depth (I don't remember exactly), but everything mentioned in sibling comments was familiar to me.

Don't forget that morse is still used by amateur radio operators worldwide.

I am curious if its use is dropping off. I heard knowing morse is no longer a requirement to get a license.

It is no longer a license requirement.

It is still used for casual contacts by people that enjoy it. It is used for various things that fall under the banner of "radiosport" -- for my part I enjoy Morse-code-only radio contests. It is used for identification beacons for repeaters or propagation beacons. And for radio experimenters, a Morse code transmitter makes a great first project since it is so simple and is foundational to everything that comes after it. Pretty much any transmitter that you build is just extra stuff on top of a Morse code transmitter in one form or another.

Morse code has a distinct signal-to-noise ratio advantage over any other simple modulation scheme, since the receiving filter can be so narrow, and 100% of the power is going into the carrier any time you are key-down. This is why Morse code is good for propagation beacons and also for other experiments. There are, of course, modern error-correcting codes that beat it, but then both the transmission and reception become much more complex. (See FT8 for a currently popular weak-signal mode).

I view Morse code radio contests much like a sail boat racing. Nobody moves containers of goods from China to LA on sail boats. But lots of people enjoy a good boat race on a sunny afternoon, and enjoy keeping the old marlin-spike seamanship skills alive. Likewise, ftp is a much smarter way to move a 3GB file from London to SF than using Morse code, but that doesn't mean that Morse code isn't good humor during a weekend radio contest.

It’s still very popular… tuning around CW portion of 40 meters still yields plenty of stations. There’s a fair number of enthusiasts in the Long Island CW Club [0] and CWOPS [1]—bith very welcoming groups committed to teaching CW. It’s fun too—a bit like learning an instrument tbh. Learning it gives you access to very cheap, small, lightweight QRP (low-power) transcievers with great performance worldwide. It’s also a useful skill when band conditions aren’t as favorable.

[0] https://longislandcwclub.org/ [1] https://cwops.org/

Nice, I hope people keep at it. And the low power radios sounds interesting.

One area Morse code is popular is using limited radios. Morse code works well with tiny (phone sized), low power radios. There is challenge called SOTA where hike to top of mountain and make contacts. It is much easier to take small radio and Morse paddle than the bigger radios needed for voice. There are low-power data modes, but then have to carry computer along and deal with software.

In absolute terms, probably, but some are declaring a renaissance associated with the increasing popularity in portable operation. Summits On The Air (SOTA) in particular favors lightweight transmitters. CW is such an efficient mode, in terms of speed, reach and circuit design that it is worth learning Morse for that single use case.

Personally, once I gained some proficiency with CW I realized that it is an enjoyable pursuit in and of itself, so now while not climbing mountains I stay on the low parts of the band and exchange postcards with octogenarians and I couldn't be happier.

It's still done in the UK, however you don't have to memorise anything. You're allowed to take a Morse code table into your examination, so most people don't consider 'learning' Morse code to be a requirement.

I got my license during the pandemic and it was not a requirement.

It is no longer a requirement, but there is a significant amount of amateur spectrum where CW (morse) is the preferred or only permitted modulation.

Morse code lives on in the form of Station Identification [1]. By law, radio stations are required to periodically transmit identifying information. The announcer on a commercial radio station will typically say it, but there are lots of Land Mobile Radio stations out there that don't have a person in the loop.

For an unattended station, by far the cheapest and easiest thing to do is program the base station to periodically transmit the station ID as a burst of Morse code. It's not that many years since I wrote code to transmit Morse, that code being embedded in radio base stations that are still being sold.

[1] https://en.wikipedia.org/wiki/Station_identification#Identif...

Same is true for aviation nav aids: VOR, ILS, … stations transmit their name in Morse code for identification.

Same is true for maritime racon navigation marks which transmit a morse code letter as identification in response to the reception of a radar pulse...the morse code is then visible on the radar display.

We had to learn morse code for Hong Kong's "grade 1 pleasure vessel certificate of competency" -> a credential required to operate pleasure vessels greater than 15 meters in length.

It's pretty easy to pick up and there are a few phone mobile phone apps that make learning morse code fun.

Much more efficient than the "tap code" prisoners re-invent.

POW's who knew more code still resorted to a new "tap code" because dot, dash, pause can't quickly be encoded with simple tap and pause.

This is my annoyance when people talk about knocking in morse code. Though I suppose WalterBright in a sibling comment solved that handily.

Tap and scrape. Or taptap for dit, tap tap for dah.

I realize why you think this might work but understand people went with a new system for multiple reasons.

First the goal is for the guards not to hear what your doing or just to hear seemingly random tapping. Scrape etc means your producing two different types and levels of noise making patterns more obvious and it's harder to control who can hear.

Also what matters is time not the number of taps. One common encoding was 1 to 5 taps then pause followed by 1 to 5 taps and another pause for each letter thus averaging 6 taps and two short pauses per letter. Morse code generally means 4 dots or dashes per letter so now you have 4 to 8 taps ie ~6 taps, but far more pauses in alternate encodings.

I'm not sure the guards would be so stupid as to not know what the tapping means. It may be hard to tell where the tapping is coming from, though.

> averaging 6 taps and two short pauses per letter

This would be true if the letters are used with equal frequency. But they are not. Morse encoding uses short sequences for more common letters and longer ones for less common letters. For example, Morse E is dot, and T is dash.

https://en.wikipedia.org/wiki/Morse_code#/media/File:Interna...

For tap code, E is 5-1 and T is 4-4.

https://www.inverse.com/article/11684-the-flash-and-tap-code...

The advantage to tap code is it is easier to learn. Which, of course, also makes it easier for the screws to learn it and "tap" into those communications.

Moving which letters are the most frequency would help tap encoding but the much worse Morse encoding more than accounts for this.

With tap encoding you need 1 unit for tap, 1 unit for pause between sections and 2 units for pauses between letters. Thus A is 1 + 1 + 1 + 2 = 5 units 2 letters take 6 units, 2 letters take 7 ... etc and only one Z takes sits at 5 + 1 + 5 + 2 = 13. (The exception is C and K share the same encoding at 1 + 1 + 3 + 2 = 7 units.) Average is ~9 units before considering letter frequency.

For Morse the best you can do is tap and 1, 2, or 3 units of time for pauses. Let's assume tap takes 1 units. Pauses between dot's and dash are the most common pause so that's 1 unit. The pause between taps on a dash the next most common and uses 2 units, and the pause between letters is 3 units -1 because you don't need the short pause after a dot or dash.

In effect a tap is 2 units (tap pause) dash is 5 units (tap, pause, pause, tap, pause) on it's own and every letter has a penalty of 2.

So E is now easily beats tap encoding at 2 + 2 = 4, T is (5 + 2) = 7, but things get much worse, The reasonably common C is 5 + 2 + 5 + 2 + 2 = 16. With the worst case Q sitting in at an abysmal 5 + 5 + 2 + 5 + 2 = 19. And that's if you can keep all the taps and pauses strait, I would be shocked if most people could actually pull off Q most of the time.

Feel free to do the time calculation on this sentence for each encoring, Morse code ends up far worse.

PS: Did some editing in the first 16 minutes after posting.

> In effect a tap is 2 units (tap pause) dash is 5 units (tap, pause, pause, tap, pause) on it's own and every letter has a penalty of 2.

See my other post in this thread about tap for dit and taptap for dah. This shortens the time. A dash would be 3 units - tap tap pause, not 5.

> I would be shocked if most people could actually pull off Q most of the time

An experience Morse operator doesn't attempt to decode Morse. His brain is trained to shortcut it and the letters are formed automatically, just like a skilled piano player is not actually aware of his finger movements or the individual notes.

I suggest tapping with a pencil to compare rather than trying a mathematical approach. Try SOS for example. I don't see a tap taking the same time as a pause. You can do a double tap in the same time period as a tap-pause easily enough. The tap itself doesn't take time.

Morse operators don't need to think about things because the sequences are so short and there's a rythim that keeps letters distinct. Abstractly dot dot dot with even spacing could be EEE or S but not EI or IE. Listing to longer messages makes the pause pattern clear, and you can even learn who's sending the message based on how they encode it.

However, with SOS anyone can just dot dot dot, dash dash dash, dot dot dot longer pause on repeat and the message is clearly SOS not EEETTTEEE. That's why SOS was chosen it allows out of band communication of information. Even better triangulation adds location information allowing effectively anyone to say HELP I AM AT LAT X LONG Y.

For a slightly less trivial sequence like HAMBURGERANDACOKE you very much need 3 distinct types of pauses to make things understandable. Further you really need to try communicating to someone else not just congratulating yourself after doing something simpler than average Morse code.

> you really need to try communicating to someone else not just congratulating yourself after doing something simpler than average Morse code.

I didn't advocate something simpler than Morse, I wrote that Morse was more efficient than Tap. Nor did I "congratulate" myself. Your reply also seems answer points I did not make.

SOS is also easy to remember without knowing full Morse.

> SOS is also easy to remember without knowing full Morse.

It's not simply a question of memory. SOS and Morse are meaningfully different protocols.

But I don't think I conveyed that bit particularly well. Try alternating between IAB, SOS, and IJS all are are dot dot dot dash dash dash dot dot dot. However, IAB and IJS is uses 3 different pauses not just 2. And that's before your tap morse code adds yet another type of pause.

P.S. After experimenting a bit, tap for dit and taptap for dash seems to work fine.

For fun, get a pencil and tap SOS on your desk (or rattle your jewelry):

    tap tap tap taptap taptap taptap tap tap tap

If you don't have a pencil, bang your cell phone instead!

Hey, that gives me another inspiration. Don't need a rocker switch to enter Morse in your phone. Phones have an accelerometer, just bang the phone!

also a microphone

if you like this, you might like 'magic boxes and secret knocks' https://www.mail-archive.com/kragen-tol@canonical.org/msg000...

I didn't think about the mike. Nice.

thanks! i had the advantage (?) of writing that before cellphones had accelerometers so i had to think up something else. haven't implemented it yet, tho

Nice to see reliable terrestrial radio services chugging along with orbital services becoming less assured.

I was recently pleased to see eLORAN is gaining momentum to fill the void left by LORAN-C's shutdown in 2010. 8 meter resolution through radio triangulation is pretty impressive.

The Jeremiah O'Brien, while docked in San Francisco, actually is in running order.

Once in a while there are cruises around the Bay.

It was at the 50th anniversary memorial for D-Day, steaming off the Normandy shore.

So I've tried to get aboard for a tour more than once, but from what I gather it's not really something you just casually do. When I called them (standing at the locked door) they said they just run tours whenever possible, which may or may not happen on any day. I'm assuming you need to know someone involved with the museum to get aboard.

Ignoring the Morse Code radio station, in the context of those banks of instruments and controls... each individual readout can be done cheaper and more reliably in software, but doing everything in software means one or two LCD screens.

Will it ever be wall-sized banks of screens covered in eInk displays as standard industrial controls? Will it skip straight past that to Augmented Reality?

through-hole components and physical controls are standard, durable, and maintainable in a way that microelectronics will never be

i see small touchscreens on devices where maintenance doesn't matter

but if you need a complex bespoke interface that's durable and lasts for years, you will use a physical control surface made of off-the-shelf parts

Had an interesting conversation with someone who used to run a radio shack and now works for one of the few remaining 'screwdriver shops' (PC shop).

If I understood it, her theory was basically that radio shack wouldn't come back because the size of components nowadays is too small.

And its funny you mention touchscreen. Because I bought a mechanical switch from her to do a DIY Mod on a 20 year old car. It was pretty simple and worked well. But this car also had a touch screen that was extremely anti-DIY friendly. These screens tend to get 'burnt' by the sun after 20+ years and stop working correctly. The typical fix for this screen is complete replacement or extremely complex repair via mail. Parts won't be available forever and the whole thing is closed source and incredibly hard to reverse engineer.

Why do those pictures look like they were created with DALL-E?

Reply on news.ycombinator.com