Oddly Powered Clocks

Gallery opened: Oct 2004

Updated: 28 Mar 2017

And yet more on the hydrogen clock
This gallery of the museum is dedicated to clocks with unusual motive power.
Back to Home PageBack to The Museum


Left: Drebbel's barometric clock: 1610

Cornelis Drebbel- the same remarkable man who is supposed to have rowed a boat underwater up the Thames- built in 1610 a clock telling the time, date, and moon phases. It appears to have been powered by changes in air pressure.

The clock mechanism is contained in the golden casing A in the centre. Around this is a circular tube containing water that is clearly at different levels on each side. The top of the tube appears to have connected to a piston or bellow inside the gold casing which wound a spring.

The mode of operation is somewhat speculative, but some very interesting experiments have been done by H R SantaColoma.


Left: Cox's barometric clock

In the 1760s the well-known clockmaker Mr James Cox developed a clock which was were wound up by changes in barometric pressure. The work was done in collaboration John Joseph Merlin, with whom Cox also worked on developing automata. Two large glass vessels containing no less than 68 kilograms (150 pounds) of mercury worked as a massive barometer; they were connected together by an ingenious system of cords and pulleys so that the pulleys would rotate back and forth as the atmospheric pressure and so the glass vessels, rose and fell. A rack-and-pinion mechanism converted this to unidirectional motion so that winding of the mainspring occurred on both rising and falling pressures, and there was a safety-device to prevent overwinding. Cox claimed that his design was a true perpetual motion machine, which of course it was not.

The clock is shown here without mercury in the vessels. This is probably a safety-measure rather than an economy measure; if my calculations are correct filling the clock to get it working again would cost something like £600.

Cox was a well-known clockmaker. He showed his self-winding clock in a private museum along with other fine clocks. When he died in 1788, a Mr Thomas Weeks bought the clock for his museum. It stayed in his museum until his death in 1833. It was not included in the sale catalogue of his effects in 1834, and remained lost until 1898 when it was exhibited at the Clerkenwell Institute. After a period on loan to the Laing Gallery in Newcastle, it was auctioned, and finally acquired by the V & A Museum in 1961.

Cox's clock has a Wikipedia page.

Cox's clock is still in the Victoria & Albert Museum in London, but I do not know if it is on display; one of these days I mean to go and find out.


You might think that a hydrogen-powered clock would function by using the gas to run a small internal-combustion engine that would rewind a spring or raise a weight, but a little thought shows that this would be a complicated (and noisy) bit of machinery. Pasquale Andervalt had other ideas...

Left: The hydrogen clock of Pasquale Andervalt

The hydrogen clock was made by Pasquale Andervalt in Italy in about 1835. There seems to be some evidence that several were built, but this is the only survivor. Very possibly all the others blew up.

The red glass jar contained sulphuric acid. Zinc pellets were stored in the brass spiral above the clock. When a pellet was dropped into the acid, hydrogen was evolved and this pushed up a small piston that raised the driving weight. The clock has a pin-pallet escapement, and the ornate pendulum can be seen in front of the red jar.

The big wheel above the dial has the driving cord running over it, and this wheel was pushed upwards by the piston and cylinder below it, raising the driving weight at bottom left. The smaller weight near the top of the red cylinder at the right simply kept the driving cord taut.

The hydrogen was then presumably released to the atmosphere when a port in the cylinder was uncovered. The clock dropped a pellet automatically when it was running down, and given the large number of pellets that could be stored in the brass spiral, it would presumably run attended for a very long time. Eventually it would of course be necessary to replenish the pellets and replace the exhausted acid. I imagine the Unique Selling Proposition was something like "almost perpetual motion".

Thus the inflammable nature of hydrogen was not exploited at all. It all seems a bit hazardous- you have a big glass jar of sulphuric acid, and clouds of hydrogen wafting about. If the pellet-dropping mechanism malfunctioned, and dropped all the pellets into the acid at once, there would seem to be a good chance that the clock would explode, sending glass fragments in all directions; followed by a second explosion when the released hydrogen encountered the nearest naked light. One hopes there was some sort of safety-valve.

Carbon dioxide might have been safer as a working fluid- marble pellets could have been dropped into hydrochloric acid. However, in the event of a multi-pellet incident, there might be issues with suffocation. The ideal gas would appear to be nitrogen, but I am not aware of any way to generate it by dropping a solid into a liquid. The standard laboratory method of preparing nitrogen is to heat a mixture of ammonium chloride and sodium nitrite dissolved in water; this evolves ammonium nitrite which is unstable and breaks up into nitrogen and water. Laboratory textbooks warn you to avoid explosions when doing this. (Though they are less clear on how to avoid them) The process does not sound too practical for a clock.

Author's photograph

On further thought, what about oxygen as a working fluid? That should be safe enough, unless the concentration in the air becomes high enough to make ordinary materials highly flammable; that seems highly unlikely. So, can we make it by dropping pellets of something into a liquid? One method that suggests itself is dropping pellets of sodium peroxide into ordinary water. This will react "violently or explosively" evolving oxygen and caustic soda, so better make the pellets fairly small. Actually sodium peroxide reacts violently with all sorts of things, and I'm not sure we are heading in the direction of greater safety.
Oxygen can be evolved by dropping manganese dioxide into 6% hydrogen peroxide, but this is a catalytic decomposition and I imagine all of the hydrogen peroxide might decompose at once, bringing back the possibility of an explosion.

Other gases which can be made by dropping solids into liquids are chlorine and sulphur dioxide, but I think the drawbacks there are obvious.

At this point I ran out of innocuous gases, and appealed for anyone with more chemical knowledge than me to suggest a working fluid for this clock that would neither blow you up nor poison you. I received this reply from my correspondent Pigeon:

"In response to your request for comments on possible alternative working fluids:

As a non-serious suggestion: acetylene, from calcium carbide dropped into water. The exhaust is not vented, but stored in a reservoir, and used to feed a little flame so you can still see what time it is at night. (I am presuming that the working pressure does not exceed 1 bar gauge, of course.)

Being serious, though, I would definitely plump for carbon dioxide: unreactive, non-toxic and dead easy to make.

As a suffocation hazard it can be ignored. It is only dangerous in that regard if there is so much of it that it displaces enough oxygen-containing air that you can't profitably breathe the result, which would require a heck of a lot of it. The amount of substance contained in the clock's curly reservoir tube could not possibly produce that much CO2 unless both you and the clock were trapped in a compartment so small that you'd soon breathe all the oxygen in it anyway.

As far as suffocation hazards are concerned, nitrogen is actually worse. Both act purely by displacing breathable air, but CO2 will provide at least some warning that this has happened, because the body determines how hard it needs to breathe principally by detecting excess of CO2, not shortage of oxygen. So excess of CO2 in the atmosphere will cause breathlessness and clue you in that something is wrong (although in practice the effect is small and it still helps to be on the alert for it).

Nitrogen, on the other hand, as one might expect since we breathe 78% of it all the time, produces no warning at all; you just fall over. There was an incident in the King's Cross area when the Victoria line of the Underground was being constructed, due to ingress into the works of air that had percolated down through the clay, losing its oxygen on the way to reactions with reducing minerals and ending up as more or less pure nitrogen. As long as construction was actually in progress there was enough ventilation that this didn't matter, but when operations were paused for the weekend and ventilation ceased, the tunnels filled up with nitrogen. Result was that the first blokes through the door on Monday morning just plain conked out more or less instantly. Fortunately someone following on was sufficiently with it to realise what had happened, so they were able to rescue them in time and without the rescuers suffering the same fate.

And of course CO2 can be produced by the reaction of chemicals so commonplace and harmless that they can be obtained from standard kitchen stock. Vinegar and bicarb will do it. Or vinegar and chalk, for something that conveniently comes as lumps rather than powder.

(Just make sure it's real chalk: blackboard chalk doesn't work, because it isn't chalk, it's gypsum - calcium sulphate. This was a source of disappointment to me at school, when I pinched blackboard chalk out of the classroom and put it into the vinegar pot at dinner, expecting lots of CO2 and a frothy squirty mess... only to find that nothing happened at all. Which continued to puzzle me right up until the internet came along and told me what blackboard chalk really was.)"

At this stage I questioned. I pointed out that there certainly is such a thing as carbon dioxide poisoning, also known as hypercapnia, though a 10% concentration is likely to be needed to kill. However 7% brings on mental confusion. Pigeon replied:

"...Assuming for the sake of argument that a Total Pellet Release Incident generates 100 litres of CO2, to achieve a 10% concentration would require that both you and the clock were shut in a cabinet so small as to leave only 1 cubic metre of unoccupied volume, which would not be a situation you'd willingly get yourself into in the first place. To generate 100 litres of CO2 at STP requires about 400g of calcium carbonate; guesstimating from your photo I'd reckon the capacity of the coiled tube, in terms of pellets of chalk, is of that order, so I still think we're safe."

I am convinced. Carbon dioxide it is.

The clock is currently on display in the Science Museum, London. The clock was presented to the Museum of the Worshipful Company of Clockmakers by William Wing in 1874. The only William Wing known to Google was an entomologist who died in 1855, so the donor is currently mysterious.

Left: The hydrogen clock of Pasquale Andervalt

This picture is an improvement on the one here before. It shows at bottom centre, just beyond the end of the pipe, the critical valve that drops zinc pellets into the acid. (Unfortunately it is hard to get sharp pictures through glass) There is a round port into the jar, closed with a disc which is presumably pressed up against its seating when there is pressure in the jar.

Further back up the pipe there is some sort of geared mechanism for releasing pellets one at a time, but its details are not clear. This mechanism is triggered by the lever (seen just above the spoked wheel) when the carriage holding the big wheel descends.

The spoked wheel is one of the four guiding wheels for the up and down movement of the carriage.

Author's photograph

According to Popular Mechanics for April 1914: (p552) "The variations of pressure in the water mains are utilised by a French inventor for the operation of a self-winding clock." And that's all they wrote.

Presumably there was a spring-loaded piston that moved as the water pressure varied with the daily demand cycle; that should provide plenty of power to wind up a clock. Whether it would work with modern water supplies, which one imagines would have good pressure regulation, is another matter.

Google has nothing on this.

The Beverly Clock is displayed in the foyer of the Department of Physics at the University of Otago, Dunedin, New Zealand. It is powered by changes in air pressure, and more importantly temperature, acting on a 1 cubic-foot box of air which presses on a diaphragm and raises the clock weights, presumably by some sort of ratchet mechanism. The clock was built by Arthur Beverly in 1864. The clock has, like the Atmos described below, a torsional pendulum with a very slow period that requires very little power to keep it working; torsional pendulums are used in so-called "400-day" clocks. The Beverly Clock occasionally stops if the ambient temperature has not fluctuated enough.

The images in this section were very kindly provided by John Howell.

Left: The back of a Puja clock made by the German firm of Jauch and Schmid.

The notion here is to power a clock reliably when faced with mains electricity of uncertain voltage and frequency. It is perhaps significant that the patent for the principle (No. 714893) was granted in Germany in 1940.

At the lower left, shielded by a translucent housing, is a carbon rod resistance that heats the coloured alcohol in the glass vessel just above it. This causes some of the alcohol to vapourise, the pressure pushing the liquid up the connecting pipe to the vessel at top right. As the latter gets heavier the wheel bearing the four vessels experiences a torque that rewinds a remontoire* spring driving a conventional gear train and escapement. This clock has a pendulum-controlled escapement, but models with balance wheel escapements also existed.

The firm of Jauch and Schmid was registered in 1930

*A remontoire, from the French 'remonter' (to rewind) is a spring or gravity reserve of power that can be configured to give a near-constant driving torque because it is rewound at frequent intervals from another power source- usually this was a mainspring, whose own torque would slowly decrease as it unwound. The idea was that rewinding a spring or lifting a weight at relatively frequent intervals isolated the escapement from the variable torque of the mainspring.

Left: Advertising material for the Puja clock movement.

This version has a balance-wheel escapement attached instead of a pendulum.

To save you the trouble of grappling with a German-English dictionary, here are the translations of the salient words in the advert above; "wechselström" means "alternating current", "gleichström" means "direct current", "thermo-aufzug mit glaskolben-laufrad" translates as "thermo-lifter with glass bulb impeller", and "gehwerk" as "movement".

Left: Another Puja clock movement.

It occurs to me that this arrangement must be very inefficient. It looks as though much of the heat would escape without doing anything useful.

I also wonder if they were prone to catching fire.


Anyone interested in oddly-powered clocks will have heard of the Atmos clock, which appears to be mostly powered by changes in temperature, and not, as its name might suggest, solely by changes in atmospheric pressure. A flexible metal capsule is filled with an inert gas and a little ethyl chloride, which vapourises as the temperature rises, causing the bellows to expand, and vice versa. A chain transfers this movement to wind the mainspring. A torsional pendulums with a long period is used to minimise the power required.

Left: An Atmos clock

A temperature variation of only one degree in the range between 15 and 30 degrees Celsius, or a pressure variation of 3 mmHg, is said to be sufficient for two days' operation. I don't know if it is the case, but that seems to imply that the clock would stop working if the temperature fell below 15 degrees and stayed there, with the ethyl chloride remaining liquid. This is of course very possible in Winter. I also wonder how it copes with thermostatically-controlled central heating.

The torsional pendulum makes only two oscillations per minute, which is 1/60th the rate of the standard seconds pendulum in a conventional clock. Because of the very slow movement of the gear train, no oil is used; it is claimed no measureable wear occurs.

Find out more:



Back to Home PageBack to The Museum EntranceTop of this page