【コラム】 タージ・マハル旅行団と小杉武久 ― 音を解き放つ旅の軌跡

1971年、タージ・マハル旅行団はヨーロッパ公演を目的に出発する。
彼らの旅は、単なるツアーではなく「音の巡礼」であった。
電源もステージもない野外で、風・水・群衆・電波と即興的に共鳴する演奏を展開...

続きはこちらから🔽
monumental-movement.jp/column-…

◤ Divided Soul: The Life of Marvin Gaye (David Ritz)

マーヴィン・ゲイの決定的評伝。著者が長年にわたり本人や家族、関係者に取材し、天才ソウル・シンガーの内面を深く掘り下げる。〈What’s Going On〉〈Let’s Get It On〉など名作誕生の裏側、名声と信仰、性愛と罪、父親との複雑な関係など、マーヴィンの魂の分裂を生々しく描写。音楽史の重要人物としての創造性と、人間としての脆さや苦悩が交錯する、感動と痛みの伝記文学の金字塔。(英語版)

amzn.to/4i1SIsk

【コラム】 Marvin Gaye：愛と痛みの間で ― ソウルの神が見た救済と破滅

1982年、ブリュッセルのホテルの一室。 Roland TR-808の乾いたビートが鳴り響く中、Marvin Gayeは孤独にマイクへ向かっていた。 『Midnight Love』—— それは、亡命と再生、愛と自己破壊の狭間で生まれた祈りのアルバム...

続きはこちらから🔽
monumental-movement.jp/column-…

◤ 音楽する日乗

作曲家・久石譲が日記形式で綴った創作記録集。コンサートの準備や映画音楽の制作、日々の練習や思索など、音楽とともにある生活の細部が静かに描かれる。新曲への試行錯誤や、ミニマル・ミュージックへの探求、若手音楽家との対話などを通して、久石譲の「音楽するとは生きること」という信念が浮かび上がる。派手なエピソードよりも、日々の鍛錬と誠実な姿勢が伝わる一冊で、創作の現場を覗くような臨場感に満ちている。

amzn.to/4nKoETj

【コラム】 久石譲とスタジオジブリ ― 音楽が紡いだ40年の物語

Quincy（クインシー）→ 日本語的発音「クインシー（kuinshi）」をもじり、「久石（ひさいし）」を当てたものとされ、「ジョーンズ（Jones）」の「ジョー」から “Joe” を取って、英語表記「Joe Hisaishi」 ....

続きはこちらから🔽
monumental-movement.jp/column-…

Writing Type-Safe Generics in C

The fun part about a programming language like C is that although the language doesn’t directly support many features including object-oriented programming and generics, there’s nothing that’s keeping you from implementing said features in C. This extends to something like type-safe generics in C, as [Raph] demonstrates in a blog post.

After running through the various ways that generics are also being implemented using methods including basic preprocessor macros and void pointers, the demonstrated method is introduced. While not necessarily a new one, the advantage with this method is that is type-safe. Much like C++ templates, these generics are evaluated at compile time, with the preprocessor handling both the type checking and filling in of the right template snippets.

While somewhat verbose, it can be condensed into a single header file, doesn’t rely on the void type or pointers and can be deduplicated by the linker, preventing bloat. If generics is what you are looking for in your C project, this might be a conceivable solution.

hackaday.com/2025/11/17/writin…

Binary Clock Also Monitors Weather

There are two things most of us want to know on a daily basis—the weather, and what time it is. [Guitarman9119] built a single device that can provide both pieces of information with a pleasingly nerdy aesthetic.

The heart of the build is a Raspberry Pi Pico W, which is proudly displayed on the front panel of the device. It’s responsible for driving the array of LEDs that display the time in hours, minutes, and seconds in binary format. The Pi Pico W uses its wireless connection to query the WorldTime API and an IP geolocation server. This provides the local date and time, and the location is then used to query the OpenWeather service for current weather information. The weather information is thankfully not displayed in binary format, because that would be straining to read. Instead, it’s displayed in human-readable format on a small OLED display.

There’s something about the way this is built—the discrete LEDs, that weird blue color that seemed to disappear by 1984—that gives this a wonderfully old school charm. You could imagine it turning up in a college lab full of old blinkenlights gear. Video after the break.

youtube.com/embed/hAAY1Tiw1yg?…

hackaday.com/2025/11/17/binary…

Building A Smart Speaker Outside The Corporate Cloud

If you’re not worried about corporate surveillance bots scraping your shopping list and manipulating you through marketing, you can buy any number of off-the-shelf smart speakers for your home. Alternatively, you can roll your own like [arpy8] did, and keep your life a little more private.

The build is based around an ESP32 microcontroller. It connects to the ‘net via its inbuilt Wi-Fi connection, and listens out for your voice with an INMP441 omnidirectional microphone module. The audio data is trucked off to a backend server running a Whisper speech-to-text model. The text is then passed to Google’s Gemini 2.5 Flash large language model. The response generated is passed to the Piper Neural Voice text-to-speech engine, sent back to the ESP32, and spat out via the device’s DAC output and a speaker attached to an LM386 amplifier. Basically, anything you could ask Gemini, you can do with this device.

By virtue of using a commercial large language model, it’s not perfectly private by any means. Still, it’s at least a little farther removed than using a smart speaker that’s directly logged in to your Amazon/Google/Hulu/Beanstikk account. Files are on Github for those eager to dive into the code. We’ve seen some other fun builds along these lines before, too. Video after the break.

youtube.com/embed/_BZOLwxb_aU?…

hackaday.com/2025/11/17/buildi…

Meet The Shape That Cannot Pass Through Itself

Can a shape pass through itself? That is to say, if one had two identical solids, would it be possible to orient one such that a hole could be cut through it, allowing the other to pass through without breaking the first into separate pieces? It turns out that the answer is yes, at least for certain shapes. Recently, two friends, [Sergey Yurkevich] and [Jakob Steininger], found the first shape proven not to have this property.

A 3D-printed representation of a cube passing through itself [image: Wikipedia]Back in the late 1600s, Prince Rupert of the Rhine proved it was possible to accomplish this feat with two identical cubes. One can tilt a cube just so, and the other cube can fit through a tunnel bored through it. A representation is shown here.

Later, researchers showed this was also true of more complex shapes. This ability to pass unbroken through a copy of oneself became known as Rupert’s Property. Sometimes it’s an amazingly tight fit, but it seems to always work.

In fact, it was so difficult to find candidates for exceptions to this that it was generally understood and accepted by mathematicians that every convex polyhedron (that is, every shape with flat sides and no holes, protrusions, or indentations) would have Rupert’s property. Until one was found that did not.
Noperthedron pencil holder
The first shape proven not to be able to pass through itself — known as the Noperthedron — is a vaguely ball-like shape, with a flat top and bottom. A fan has already added a cavity to create a 3D-printable pencil holder version of the noperthedron (shown here) if you want your own.

There are other promising candidate objects (they are rare) that may also lack Rupert’s property, but so far, this is the only proven one.

Shapes with unusual properties are interesting, and we love how tactile and visual they are. Consider Penrose tiles, a tile set that can cover any size of area without repeating. For decades, the minimum number of tile shapes needed to accomplish this was two. Recently, though, the number has dropped to one thanks to a shape known as “the hat.”

hackaday.com/2025/11/17/meet-t…

Keebin’ with Kristina: the One with the Cipher-Capable Typewriter

I must confess that my mouth froze in an O when I saw [Jeff]’s Typeframe PX-88 Portable Computing System, and I continue to stare in slack-jawed wonder as I find the words to share it with you. Let me give it a shot.

[Jeff] tells us that he designed Typeframe for his spouse to use as a writer deck. That’s good spousing, if you ask me. Amazingly, this is [Jeff]’s first project of this type and scope, and somehow it’s an elegant, yet easy build that’s quite well documented to boot. Whatever Typeframe’s design may borrow, it seems to give back in spades.

Image by [Jeff] via Typeframe.netUse Typeframe for what you will — cyberdeck, writer deck, travel PC — this baby can handle whatever you throw at it. And of course, it’s open source from front to back.

This Raspberry Pi 4B-based productivity machine has all sorts of neat features. The touch screen flips upward at an angle, so you don’t have to hunch over it or carry a mouse around. Want to sit back a bit while you work? The aesthetically spot-on keyboard is detachable. Yeah.

If that’s not enough to get you interested, Typeframe is designed for simple construction with minimal soldering, and the sliding panels make maintenance a breeze.

A little more about that keyboard — this is Keebin’, after all. It’s an MK Point 65, which boasts hot-swap sockets under those DSA Dolch keycaps. See? Minimal soldering. In fact, the only things you have to solder to make the Typeframe your own are the power switch and the status light. Incredible.

Decktility Goes Where You Go

Need something even more portable than the Typeframe? Something that might even fit in your cargo pocket? Decktility could be your answer. You’re gonna have to use your thumbs to type on this one, but that’s the price you pay for ultra portability.

Image by [Ken Van Hoeylandt] via Byte WelderBuilt back in 2023, [Ken Van Hoeylandt] combined the Raspberry Pi CM4 with the BigTreeTech Raspberry Pad 5, and “spent many hours developing a case” that fits perfectly. The idea was to push the limits of a Pi-based handheld without it melting down.

The tricky part about using your thumbs is that the experience can be somewhat lacking. But the Bluetooth keyboard [Ken] used actually looks pretty good, and I say this based on this close-up from GitHub. The keys look responsive and the caps aren’t too shallow.

One of the design challenges centered around the batteries — 18650s, because as [Ken] says, flat Li-Pos get spicy when they short-circuit. Both the 18650s and the screen are somewhat heavy, so everything is balanced with the batteries acting somewhat as handles.

This is a great project, and it seems as though [Ken] learned a lot in the process. Be sure to check out the build log for all the gory details.

The Centerfold: Party Like It’s Nineteen Nine-T9

Yeah, I know, not your conventional centerfold. The cuteness factor of this 12-key macropad got me, as did the terrazzo party background. But there’s slightly more to [Joe Scotto]’s latest build than meets the eye.

Rather than using predictive typing, the ScottoT9 uses QMK tap dance, wherein a single key can do ‘3, 5, or 100 different things’ depending on how many times you hit it in quick succession. [Joe] reports that it’s surprisingly easy get up to speed on the thing.

This RP2040 Pro Micro-based build is completely open source, including those lovely keycaps. But if you want them to look this good, you may want to just buy some.

Do you rock a sweet set of peripherals on a screamin’ desk pad? Send me a picture along with your handle and all the gory details, and you could be featured here!

Historical Clackers: the Diskret Code Typewriter

The Diskret (discreet) code typewriter from 1899 is a stunning German beauty of an index typewriter, far removed from the neon-accented see-through plastic label-maker from the late 80s that I often use.
Image via The Antikey Chop
Yet they are of the same spirit: choose a letter with some pointing device, imprint it with some lever action, and repeat.

Stark aesthetic differences aside, the Diskret does far more than my newfangled index typewriter. Designed with two concentric rings, the inner ring can be spun to allow the user to type in code.

For normal typewritering, the typist chose a letter from the outer ring of characters with a lever, which turned the typewheel mounted behind it. Then they would press down to print, and the carriage would advance by one character width to get ready for the next impression.

To use the Diskret as a cipher machine, the typist would spin the inner ring to create a false index, generating a coded letter. The recipient would have to have a Diskret in the same position in order to read the message. Yes, it’s basic, but it was one of the first code typewriters.

Most of this beautiful machine’s metal components were engraved with Jugendstil, or German Art Nouveau embellishments. It came mounted to a thick wood base, and featured a top with handle. A Diskret cost 75 marks, and, at the time, one mark equaled about $4. So, not exactly cheap, but it’s two typewriters in one.

Finally, Will Gen Alpha Cancel Keyboards Altogether?

No, of course not. But a study by the London School of Economics seems to think that the office drones of 2030 or so will skirt future first-draft keyboard use with AI voice typing tools, like everyday, all the time. And yeah, but they’ll have to use a keyboard to edit. Or, get this — they’ll just send voice notes to their bosses all the time. You know, good old unsearchable audio files that will be a pain to sift through later. Sure.
Gen Alpha stares out the window, arms akimbo, leaving a lengthy voice note for their Gen-X boss who just wants to read messages, and quickly. Image via Jabra
Did I mention that the study was done in collusion with Jabra, an audio-video technology company? They make highfalutin’ headsets and such. Ahem.

Now, I have a little secret — Y.T. has used powerful (and not-so-powerful) dictation tools here and there, especially after the surgery, and I could swear I’ve heard others talk-typing around the Hackaday Dungeon, although I can’t be sure. But I really like the act of typing, especially now that it doesn’t cause me intense, weird pain.

If I may cherry-pick a quote from the Fortune article, I choose this one: “speaking replaces typing because it matches how we think: fast, iterative, conversational.” I say all of that depends on your level of written wit, something that gets refined throughout the course of one’s life by, you know, writing or typing, and then reading, editing, removing, replacing, and so forth until it appears flowy and conversational.

And no, I didn’t speak-type a bit of this column. I missed typing on my Kinesis Advantage too much. See you next week-ish.

Got a hot tip that has like, anything to do with keyboards? Help me out by sending in a link or two. Don’t want all the Hackaday scribes to see it? Feel free to email me directly.

hackaday.com/2025/11/17/keebin…

Watching Radioactive Decay with a Homemade Spinthariscope

Among the many science toys that have fallen out of fashion since we started getting nervous around things like mercury, chlorinated hydrocarbons, and radiation is the spinthariscope, which let people watch the flashes of light on a phosphor screen as a radioactive material decayed behind it. In fact, they hardly expose their viewers to any radiation, which makes [stoppi]’s homemade spinthariscope much safer than it might first seem.

[Stoppi] built the spinthariscope out of the eyepiece of a telescope, a silver-doped zinc sulfide phosphor screen, and the americium-241 capsule from a smoke detector. A bit of epoxy holds the phosphor screen in the lens’s focal plane, and the americium capsule is mounted on a light filter and screwed onto the eyepiece. Since americium is mainly an alpha emitter, almost all of the radiation is contained within the device.

After sitting in a dark room for a few minutes to let one’s eyes adjust, it’s possible to see small flashes of light as alpha particles hit the phosphor screen. The flashes were too faint for a smartphone camera to pick up, so [stoppi] mounted it in a light-tight metal box with a photomultiplier and viewed the signal on an oscilloscope, which revealed many small pulses.

While a spinthariscope is a bit easier to set up, they’re considerably less common among amateurs than are cloud chambers, another way to view radioactive decay. For scientific instruments, though, this project’s scintillator-and-photomultiplier approach is the standard, from tiny gamma ray spectrometers to giant neutrino detectors.

youtube.com/embed/ifEmv3BeM1I?…

hackaday.com/2025/11/17/watchi…

Asus rilascia aggiornamento firmware in emergenza per router DSL vulnerabili

Asus ha rilasciato un aggiornamento firmware di emergenza per diversi modelli di router DSL. La patch corregge una vulnerabilità critica che consente agli aggressori di assumere il controllo completo dei dispositivi senza autenticazione.

La vulnerabilità, identificata come CVE-2025-59367, riguarda i router DSL-AC51, DSL-N16 e DSL-AC750. Il problema consente ad aggressori remoti di accedere a dispositivi non protetti accessibili tramite Internet. L’attacco non richiede alcuna preparazione o interazione da parte dell’utente: è sufficiente conoscere l’indirizzo IP del router vulnerabile.

“È stata scoperta una vulnerabilità di bypass dell’autenticazione in alcuni router DSL che potrebbe consentire ad aggressori remoti di ottenere un accesso non autorizzato al dispositivo”, avvertono gli sviluppatori Asus .

L’azienda consiglia vivamente ai proprietari dei dispositivi vulnerabili di installare immediatamente gli aggiornamenti del firmware alla versione 1.1.2.3_1010.

Se per qualche motivo non è possibile aggiornare immediatamente il dispositivo, il produttore offre misure di sicurezza temporanee.

Innanzitutto, disabilita tutti i servizi accessibili da Internet: accesso remoto tramite WAN, port forwarding, DDNS, server VPN, DMZ, trigger di porta e FTP.

ASUS ricorda inoltre agli utenti di utilizzare password complesse per il pannello di amministrazione del router e per le reti Wi-Fi, di controllare regolarmente la disponibilità di aggiornamenti del firmware e di evitare di utilizzare le stesse credenziali di accesso per servizi diversi.

L'articolo Asus rilascia aggiornamento firmware in emergenza per router DSL vulnerabili proviene da Red Hot Cyber.

L’Energia Nucleare tutta in mano alle AI! E se poi scoppia un incidente?

Nell’ambito tecnologico, si sta radicando l’idea che lo sviluppo futuro dei sistemi generativi dipenda inesorabilmente dall’ampliamento della capacità nucleare, spingendo i protagonisti del settore a individuare strategie per velocizzare la costruzione di nuovi reattori.

A questo proposito, sono state recentemente integrate le reti neurali nella predisposizione della documentazione richiesta per l’avvio di nuovi progetti. Tuttavia, gli istituti di ricerca mettono in guardia dalle possibili ripercussioni negative associate a questo metodo.

Secondo un rapporto di AI Now, un’iniziativa di Microsoft e della società energetica americana Westinghouse mira a utilizzare modelli per accelerare il rilascio delle licenze per gli impianti nucleari. Il rapporto sottolinea che gli algoritmi non possono sostituire il processo in più fasi progettato per ridurre al minimo i rischi e prevenire errori che portano a incidenti gravi.

AI Now sottolinea che il rilascio delle licenze non è un semplice insieme di moduli, ma un processo complesso che richiede analisi, discussione delle soluzioni progettuali e giustificazione dell’affidabilità del futuro impianto. Gli autori del rapporto sottolineano che il tentativo di ridurre questo processo alla sola generazione di documenti potrebbe comportare il completamento di alcune fasi di controllo come una formalità.

Microsoft propone di addestrare il modello utilizzando materiali archiviati da agenzie di regolamentazione e dati sui futuri siti degli impianti. L’algoritmo genererà documenti di progettazione, che saranno poi esaminati dai dipendenti dell’azienda. L’Idaho National Laboratory sta già utilizzando un approccio simile, sperando di utilizzare una rete neurale per accelerare la preparazione dei materiali per le richieste di autorizzazione. Lloyd’s Register segue un approccio simile e Westinghouse sta promuovendo il proprio sistema chiamato Bertha, promettendo di ridurre i tempi di preparazione della documentazione da mesi a minuti.

Gli esperti di AI Now considerano questo scenario una minaccia per la sicurezza nucleare. Sottolineano che anche una piccola imprecisione nella versione del software o nelle specifiche delle apparecchiature può portare a conclusioni errate sul comportamento del sistema e innescare una catena di errori. Citano come esempio l’incidente diThree Mile Island del 1979, in cui una combinazione di malfunzionamenti e interpretazioni errate delle letture portò a una fusione parziale del nocciolo. Gli autori del rapporto ritengono che i modelli siano soggetti a errori sottili, quindi il rischio che uno scenario simile si ripeta quando parte del processo di autorizzazione viene automatizzato aumenta.

Oltre ai rischi tecnici, viene sollevata anche la questione della riservatezza nucleare . Secondo AI Now, le richieste delle aziende tecnologiche di accesso a set di dati estesi dimostrano di fatto un interesse per informazioni solitamente non divulgate al pubblico. Queste informazioni potrebbero facilitare la creazione di risorse a duplice uso, pertanto l’accesso ad esse è strettamente regolamentato. Gli autori del rapporto sottolineano che il trasferimento di tali dati ai servizi cloud crea ulteriori minacce.

Il contesto politico è particolarmente preoccupante. L’amministrazione statunitense sta promuovendo una riforma normativa, cercando di ridurre i tempi di revisione dei progetti. I rappresentanti della Nuclear Regulatory Commission hanno già avvertito il Congresso che le decisioni e le riduzioni del personale potrebbero compromettere le capacità di supervisione della sicurezza. Allo stesso tempo, la Casa Bianca sta collegando lo sviluppo delle infrastrutture nucleari alla necessità di fonti energetiche ad alta densità da parte dell’IA , considerando la costruzione accelerata di reattori un imperativo per la sicurezza nazionale.

L’industria nucleare, tuttavia, ha una prospettiva diversa. Alcuni analisti ritengono che l’uso attento di algoritmi per sistematizzare grandi volumi di documenti possa migliorare l’efficienza e facilitare il lavoro degli enti regolatori. Tuttavia, anche i sostenitori di tali tecnologie sottolineano che l’automazione non può sostituire la supervisione umana e che affidarsi ciecamente alle conclusioni dei modelli rappresenta una minaccia per il settore, dove gli errori hanno conseguenze irreversibili.

Gli autori del rapporto AI Now concludono con un avvertimento: tentare di sottoporre la progettazione nucleare alla logica di una corsa all’intelligenza artificiale e ridurre il livello di controllo per accelerare il processo potrebbe minare la fiducia nella tecnologia nucleare e aumentare contemporaneamente i rischi di proliferazione dei dati necessari per creare risorse a duplice uso.

L'articolo L’Energia Nucleare tutta in mano alle AI! E se poi scoppia un incidente? proviene da Red Hot Cyber.

Tech in Plain Sight: Pneumatic Tubes

Today, if you can find a pneumatic tube system at all, it is likely at a bank drive-through. A conversation in the Hackaday bunker revealed something a bit surprising. Apparently, in some parts of the United States, these have totally disappeared. In other areas, they are not as prevalent as they once were, but are still hanging in there. If you haven’t seen one, the idea is simple: you put things like money or documents into a capsule, put the capsule in a tube, and push a button. Compressed air shoots the capsule to the other end of the tube, where someone can reverse the process to send you something back.

These used to be a common sight in large offices and department stores that needed to send original documents around, and you still see them in some other odd places, like hospitals or pharmacy drive-throughs, where they may move drugs or lab samples, as well as documents. In Munich, for example, a hospital has a system with 200 stations and 1,300 capsules, also known as carriers. Another medical center in Rotterdam moves 400 carriers an hour through a 16-kilometer network of tubes. However, most systems are much smaller, but they still work on the same principle.

That Blows — Or Sucks?

Air pressure can push a carrier through a tube or suck it through the tube. Depending on the pressure, the carrier can accelerate or decelerate. Large systems like the 12-mile and 23-mile systems at Mayo Clinic, shown in the video below, have inbound pipes, an “exchanger” which is basically a switchboard, and outbound pipes. Computers control the system to move the carriers at about 19 miles per hour. You’ll see in the video that some systems use oval tubes to prevent the tubes from spinning inside the pipes, which is apparently a bad thing to do to blood samples.

In general, carriers going up will move via compressed air. Downward motion is usually via suction. If the carrier has to go in a horizontal direction, it could be either. An air diverter works with the blower to provide the correct pressures.

youtube.com/embed/-izeXQpRd_k?…

History

This seems a bit retro, but maybe like something from the 1950s. Turns out, it is much older than that. The basic system was the idea of William Murdoch in 1799. Crude pipelines carried telegram messages to nearby buildings. It is interesting, too, that Hero understood that air could move things as early as the first century.

In 1810, George Medhurst had plans for a pneumatic tube system. He posited that at 40 PSI — just a bit more than double normal sea-level air pressure — air would move at about 1,600 km/h. He felt that even propelling a load, it could attain a speed of 160 km/h. He died in 1827, though, with no actual model built.

In 1853, Josiah Latimer Clark installed a 200-meter system between the London Stock Exchange and the telegraph office. The telegraph operator would sell stock price data to subscribers — another thing that you’d think was more modern but isn’t.

Within a few years, the arrangement was common around other stock exchanges. By 1870, improvements enabled faster operation and the simultaneous transit of multiple carriers. London alone had 34 kilometers of tube by 1880. In Aberdeen, a tube system even carried fish from the market to the post office.

There were improvements, of course. Some systems used rings that could dial in a destination address, mechanically selecting a path through the exchange, which you can see one in the Mayo Clinic video. But even today, the systems work essentially the way they did in the 1800s.

Famous Systems

Several cities had pneumatic mail service. Paris ran a 467 km system until 1984. Prague’s 60 km network was in operation until 2002. Berlin’s system covered 400 km in 1940. The US had its share, too. NASA’s mission control center used tubes to send printouts from the lower floors up to the mission control room floor. The CIA Headquarters had a system running until 1989.

In 1920 Berlin, you could use the system as the equivalent of text messaging if you saw someone who caught your eye at one local bar. You could even send them a token of your affection, all via tube.
Mail by tube in 1863 (public domain; Illustrated London News)
In 1812, there was some consideration of moving people using this kind of system, and there were short-lived attempts in Ireland, London, and Paris, among other places, in the mid-1800s. In general, this is known as an “atmospheric railroad.”

As a stunt, in 1865, the London Pneumatic Despatch Company sent the Duke of Buckingham and some others on a five-minute trip through a pneumatic tube. The system was made to carry parcels at 60 km/h using a 6.4-meter fan run by a steam engine. The capsules, in this case, looked somewhat like an automobile. There are no reports of how the Duke and his companions enjoyed the trip.
A controller for the Prague mail system that operated until 2002 (public domain).
A 550-meter demonstration pneumatic train showed up at the Crystal Palace in 1864. Designed by Thomas Webster Rammell. It only operated for two months. A 6.7-meter fan blew air one way for the outbound trip and sucked it back for the return.

Don’t think the United States wasn’t in on all this, too. New York may be famous for its subway system, but its early predecessor was, in fact, pneumatic, as you can see in the video below.

youtube.com/embed/9g9dikUrbmk?…
Image from 1867 of the atmospheric train at Saint Germain (public domain).
Many of these atmospheric trains didn’t put the passengers in the capsule, but used the capsule to move a railcar. The Paris St. Germain system, which opened in 1837, used this idea.

Modern Times

Of course, where you once would send documents via tube, you’d now send a PDF file. Today, you mainly see tubes where it is important for an actual item to arrive quickly somewhere: an original document, cash, or medical samples. ThyssenKrupp uses a tube system to send toasty 900 °C steel samples from a furnace to a laboratory. Can’t do that over Ethernet.

There have been attempts to send food over tubes and even take away garbage. Some factories use them to move materials, too. So pneumatic tubes aren’t going away, even if they aren’t as common as they once were. In fact, we hear they are even more popular than ever in hospitals, so these aren’t just old systems still in use.

We haven’t seen many DIY pneumatic tube systems that were serious (we won’t count sucking Skittles through a tube with a shop vac). But we do see it in some robot projects. What would you do with a system like this? Even more importantly, are these still common in your area or a rarity? Let us know in the comments.

hackaday.com/2025/11/17/tech-i…

TMS9900-based Home Brew Computer

[Usagi Electric] is known for minicomputers, but in a recent video, he shows off his TMS9900-based homebrew computer. The TMS9900 CPU was an early 16-bit CPU famously used in the old TI-99/4A computer, but as the video points out, it wasn’t put to particularly good use in the TI-99/4A because its RAM was hidden behind an inefficient interface and it didn’t leverage its 16-bit address space.

The plan is for this computer to have 2K words of ROM, 6K words of RAM, and three serial lines: one for the console terminal, another for a second user console terminal, and the third for access to a tape drive.

Note that we have two user terminals: this is a multiuser system! The computer will use the TI series 10 “Insight” data terminal.

In the video, [Usagi Electric] spends a fair bit of time making the rack-mount casing for his computer and its two power supplies. The UART for 300-baud terminal access is currently in breadboard format, but it is set up to transmit and is functional so far! Up next will be support for receiving. The UART he’s using is the TR1602B, and he spends some time reviewing its datasheet in this video.

If you’re interested in the TMS9900, you might like to check out TMS9900 Retro Build and How The TI-99/4A Home Computer Worked.

youtube.com/embed/OUp0VRuMVCA?…

hackaday.com/2025/11/17/tms990…

Reverse Engineering the Miele Diagnostic Interface

The infrared transceiver installed on the washing machine. (Credit: Severin)
Since modern household appliances now have an MCU inside, they often have a diagnostic interface and — sometimes — more. Case in point: Miele washing machines, like the one that [Severin] recently fixed, leading to the firmware becoming unhappy and refusing to work. This fortunately turned out to be recoverable by clearing the MCU’s fault memory, but if you’re unlucky, you will have to recalibrate the machine, which requires very special and proprietary software.

Naturally, this led [Severin] down the path of investigating how exactly the Miele Diagnostic Utility (MDU) and the Program Correction (PC) interface communicate. Interestingly, the PC interface uses an infrared LED/receiver combination that’s often combined with a status LED, as indicated by a ‘PC’ symbol. This interface uses the well-known IrDA standard, but [Severin] still had to track down the serial protocol.

Research started with digging into a spare 2010-era Miele EDPW 206 controller board with the 65C02-like Mitsubishi 740 series of 8-bit MCUs. These feature a mask ROM for the firmware, so no easy firmware dumping. Fortunately, the Miele@Home ‘smart appliance’ feature uses a module that communicates via UART with the MCU, using a very similar protocol, including switching from 2400 to 9600 baud after a handshake. An enterprising German user had a go at reverse-engineering this Miele@Home serial protocol, which proved to be incredibly useful here.

What is annoying is that the PC interface requires a special unlock sequence, which was a pain to figure out. Fortunately, the SYNC pin on the MCU’s pins for (here unused) external memory was active. It provided insight in which code path was being followed, making it much easier to determine the unlock sequence. As it turned out, 11 00 00 02 13 were the magic numbers to send as the first sequence.

After this, [Severin] was able to try out new commands, including 30 which, as it turns out, can be used to dump the mask ROM. This enabled the creation of a DIY transceiver you can tape to a fully assembled washing machine, for testing. As of now, the next target is a Miele G651 I Plus-3 dishwasher, which annoyingly seems to use a different unlock key.

Of course, you can just trash the electronics and roll your own. That happens more often than you might think.

Thanks to [Daniel] for the tip.

hackaday.com/2025/11/17/revers…