Hacker News new | past | comments | ask | show | jobs | submit login
The Paris Pneumatic Clock Network (2018) (douglas-self.com)
142 points by Luc 62 days ago | hide | past | web | favorite | 33 comments

The Swiss railways have a slightly similar system of synchronization once per minute, with an interesting nuance:

"The station clocks in Switzerland are synchronized by receiving an electrical impulse from a central master clock at each full minute, advancing the minute hand by one minute. The second hand is driven by an electrical motor independent of the master clock. It requires only about 58.5 seconds to circle the face, then the hand pauses briefly at the top of the clock. It starts a new rotation as soon as it receives the next minute impulse from the master clock.[4] This movement is emulated in some of the licensed timepieces made by Mondaine."


In the United States there were a number of private time services one could subscribe to.

One of the largest was operated by Western Union. A telegraph line would be used to send a synchronized signal a short period before the top of the hour, which energized a solenoid in the clock. The minute hand would be pulled to the :00 position and the clock would restart once the signal was released. Commercial telegraph message traffic would be offline until this signal was completed.


There were also thousands of local "master clock" installations in schools and other facilities. Lots of kids from the 70s and 80s probably remember the IBM school clock:


There's also the time error correction that's done to ensure the accuracy of clocks that get their timing signal from the AC electric supply.


I'd never considered the school clocks would be a centralized system, but it makes sense. Interesting

NB the school clocks were only "locally" centralized...

Or even with "rolling updates", https://thedailywtf.com/articles/SyncingSunk

The whole site for this is great. I really like the multitude of variations they had in early steam/IC engines[0]; it reminds me a lot of the expirementing we have in the software ecosystem today



I too think of these analogies buy my favorite subject is the early power industry.

In the early days you had a multitude of frequencies and voltages. Higher frequencies were desired for flicker free lighting but low frequency was desirable as you could directly run commutated DC motors on low frequency AC without brush arcing. This was important as AC was displacing DC in distribution. Frequency selection was a compromise between motor reuse/longevity and lighting flicker. In the end Westinghouse settled on 60Hz and GE moved from 40Hz to 50Hz, then finally to 60Hz. German AEG which was a licenser of Edison Electric/GE patents stuck with 50Hz and the rest is history. The competition between AEG and Westinghouse is also why part of Japan is 50Hz, the other, 60Hz. Little known fact: California was 50Hz until 1948 as GE was more popular on the west coast.

And to add to your original thought: If you watch Star Trek, everything is ran by LCARS and everything seamlessly integrates. I wonder how and where the industry will settle. Right now it feels like we're partly moving into a post OS ecosystem where the applications have been abstracted from the underlying system. I feel that it's a stop gap between something more akin to LCARS or real life distributed systems like Erlang or Plan 9.

Southern Ontario was still running on 25hz and only finished the transition in the late 50s.


A steel mill still kept a hydroelectric station running in Niagara Falls at 25hz until 2009:


Trains in Norway (and, presumably, also elsewhere) run on 16 2/3Hz AC.

This is a historical artifact - back in the day, electrical motors didn't fare too well if fed with 50hz, so a lower frequency was desirable.

In airplanes, on the other hand, AC supply is typically 400Hz - as higher frequency equates a smaller, lighter transformer for a given throughput.

16 2/3 Hz or 16.7 Hz as the newer standardization says is used by all countries that used AC early. 50 Hz wasn't feasible for railways before the 1960s or so.

16 2/3 Hz is at least in use in Germany, Switzerland, Austria, Sweden, and Norway. In Germany development started already in the 1920s and regular usage in the 1930s.

France, Finland, and Denmark for example were later with AC and they chose 50 Hz.

Re: 400hz: and smaller magnets for the generators/motors.

Just like how a higher frequency antennas can be smaller for the same amount of gain.

And for high voltages, systems are moving back to DC again, because of lower transmission losses.

And modern semiconductors.

The original Pacific DC Intertie used enormous mercury vapor tubes.

There's a high voltage DC link between the North and South islands of NZ, it used mercury arc valves until the 1990s. https://en.wikipedia.org/wiki/HVDC_Inter-Island

If you enjoy "unusual" engines from that site - these seem very scary:


Imagine a steam engine - in which the working fluid come from boiling gasoline...

It was almost certainly related to the "télégraphe pneumatique" messaging system, that allowed to send mail and small objects extremely rapidly across Paris in air-filled tubes: http://www.cix.co.uk/~mhayhurst/jdhayhurst/pneumatic/book1.h...

They had the compressed air generation available already! And this service was available until 1984 (the last line was closed in 2004!).

More details on the French Wikipedia: https://fr.wikipedia.org/wiki/Tube_pneumatique

> And this service was available until 1984 (the last line was closed in 2004!).

Odd that it fell out of favor, sending things around only became more popular with all the internet shopping!

What's old can be new again! :-) My own great-grandmother was user of this pneumatic mail network. Perhaps it was displaced by the Minitel -- another remarkable invention that fell by the wayside.

Hamburg built a system with a pipe diameter of 45cm in the 1960ies but the system was scrapped soon after. It would be a godsend nowadays. See Großrohrpost on this page https://de.m.wikipedia.org/wiki/Rohrpost_in_Hamburg (sadly no English translation exists)

> The pipes ran through the sewers of the city, and the tunnels of the Metro and the RER. (...) Operation began in 1880...

The Paris Metro was inaugurated in 1900, and the RER in 1977, so this is unlikely. (It would be possible to run pipes through tunnels before the inauguration, but works on the Metro started no earlier than 1898).



According to TFA, the pneumatic clock system was in operation until 1927, so there might have been plenty of time to make use of the Metro's tunnels at least a little bit - perhaps to extend service to new buildings.

Maybe, but not at the time of implementation, and certainly never with the RER.

I liked the "explode all clocks" lever, which puts 73 PSI of pressure into the lines designed for just 10 PSI.

> Note that there is a provision for connecting the high pressure air directly to the low-pressure reservoir, by means of valve f. Why you would want to do this I do not know, because as mentioned above I would have thought it would have caused all the clocks to explode.

A military precaution? Any valuable system has to have a self-destruct button, to prevent it from falling into enemy's hands.

I wonder if any equipment from the system survives today... a quick Internet search doesn't turn up anything but the old drawings and photos shown here. Like a lot of other obsolete technology, it's one thing to read about it, but I'd be very interested in seeing a (modern) photo.

amazing. I had no idea it was so complex. Very informative. Thanks a lot

I'm just sort of amazed that people felt it was worth that much apparent expense and infrastructure to have synchronized public clocks. One central clock tower with a bell would tell people what time it is.

Peter Galison's book "Einstein's Clocks, Poincaré's Maps" is based on his personal speculation that it was the late 19th C preoccupation with simultaneity that inspired the theory of relativity. Whether that's true or not, he discusses the systems in the book.

"When I came back to the United States I started poking around old Swiss, British, German, and American patents and industrial records, and it turns out that there was an enormous industry in coordinated clocks in the late 19th century. Suddenly the famous metaphor with which Einstein begins his 1905 paper began to look not so peculiar. Einstein asks us to interrogate what we mean by simultaneity. He says, imagine a train comes into a station where you are standing. If the hour hand of your watch just touches 7:00 as the train pulls in front of your nose, then you would say that the train’s arrival and your watch showing 7:00 were simultaneous. But what does it mean to say that your clock ticks 7:00 at just the moment that a train arrives at a distant station? Einstein goes on to develop a technique for saying what it would mean to coordinate clocks, and explains that this is what simultaneity is. This quasi-operational definition of simultaneity becomes the foundation of his theory and leads to his startling conclusions that simultaneity depends on frame of reference, that therefore length measurements are different in different frames of reference, and to all of the other famous and amazing results of relativity theory. Suddenly I could see that Einstein’s seemingly abstract metaphor about trains and stations was actually both entirely metaphorical and yet altogether literal. Far from being the only person worried about the meaning of simultaneity—a lighthouse keeper in splendid isolation--there was a vast industry of people worrying about what it meant to say that a train was arriving at a distant train station. And they were determining simultaneity by sending electrical signals down telegraph lines to distant stations in ways very much like the way Einstein was describing in that fateful paper."

If you were a sailor, accurate time was pretty important. Lots of maritime communities used time balls to sync up the clocks on ships.


(This is also where the NYC Times Square ball drop on New Year's Eve comes from)

Technically, you had a farther and faster reach with an optical system than you did with a church bell. But at this scale the error difference wasn't that large.

Given the noise level in the city, a bell would only cover a few blocks.

Also, this solves a different problem: having accurate clocks. Unless you want to station an attentant under each clock to reset it when the bell rings. The assumption is that there's a lot of drift in the clocks of that time so frequent re-syncs are necessary.

Mechanical clocks need regular (probably daily) winding, and drift by quite a bit per day - as much as a few minutes.

The hordes of people you'd need to employ to maintain such a system might well have been expensive enough to make an installation like this the more economical option. And, given that pocket watches were a very expensive luxury item at the time, having lots of publicly visible clocks might have been a public service that could be justified on grounds such as enabling smoother commerce.

It's true the whole site is full of gems!

I wonder why Elon hasn't tried to go prop with high speed locos! http://www.douglas-self.com/MUSEUM/LOCOLOCO/proprail/proprai...

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact