Help wanted designing a source of 0.5Hz pulses, please

[Mike Phelan]

Originally Posted by Hermit6345

This problem sounds familiar!

I have a similar problem with a partially built clock, which seems to work in a similar way to Nicks.

That looks like a modified Synchronome-Shortt system (I've made one of them as well!) without the count wheel. One of these was used as the Greenwich standard time signal until replaced by a quartz oscillator and lots of valves in the 1930s.
Is the solenoid on the left just to drop the gravity arm?

A 30-second impulse should be fine - any more frequent and there will be too much interference with the pendulum.

[Oldtimer74]

As it was originally home built have we any proof that it ever did work properly? I am not familiar with the term Hipp Toggle but it sounds to me as if it is like the old masterclock/slave system where every 1/2 minute an impulse is given to a solenoid to keep the thing going as well as an impulse to run the slave clocks. These things had a long pendulum (1sec?) which had sufficient momentum to keep going for 1/2 minute between impulses. A shorter lighter pendulum might not be suitable. I would agree that it would be worth trying a resistor in series to reduce the power of the solenoid. Does the solenoid actually touch the pendulum when it gets energised? That certainly would be undesirable.

Pat G3IKR

[Hermit6345]

Originally Posted by Mike Phelan

One of these was used as the Greenwich standard time signal until replaced by a quartz oscillator and lots of valves in the 1930s.
Is the solenoid on the left just to drop the gravity arm?

A 30-second impulse should be fine - any more frequent and there will be too much interference with the pendulum.

Mike, yes, that's quite right. The LH solenoid drops the gravity arm and the other one resets it after it's delivered it's energy to the pendulum. The wheel behind the clock face has 120 teeth on it, so that one operation of the associated solenoid and pawl mechanism will move it one tooth and thus the minute hand by 30 seconds.
I will have to get down to finishing it off!

Ian.

[Mike Phelan]

Originally Posted by Oldtimer74

As it was originally home built have we any proof that it ever did work properly?

I'd say it didn't as per my posts regarding the counting.

I am not familiar with the term Hipp Toggle but it sounds to me as if it is like the old masterclock/slave system where every 1/2 minute an impulse is given to a solenoid to keep the thing going as well as an impulse to run the slave clocks.

No - it receives an impulse when the arc of the pendulum has decreased. See post #18.
There is a commercial Hipp system in the Gents "waiting train" movement used in turret clocks (churches and other public buildings etc) that are subject to rain, wind and snow. The clock is set to run fast and is stopped every 30 seconds (on its own time, which might be 28-29 seconds, and a pulse from the master clock restarts it at precisely 30 seconds. Here is one.

These things had a long pendulum (1sec?) which had sufficient momentum to keep going for 1/2 minute between impulses. A shorter lighter pendulum might not be suitable.

Correct - they were always seconds pendulums.

I would agree that it would be worth trying a resistor in series to reduce the power of the solenoid. Does the solenoid actually touch the pendulum when it gets energised? That certainly would be undesirable.

Pat G3IKR

Very undesirable indeed! I don't think that adding a resistor would achieve anything as you could reduce the voltage.

[Nickthedentist]

Goodness, thanks everyone for all the replies!

I think GJ summarises my initial feelings very well in post no. 20. In other words, as I’m no clockmaker, an electronic “bodge” seemed the way to go, especially as the original design is unproven. The pendulum on its own is accurate, so it would probably stay in phase with an accurate oscillator, though I would be “pushing time into the pendulum” as Top Anode nicely puts it.

Mike said:

That is why the timekeeping is no good; it should not miss a tooth or count two teeth. The pawl should count one tooth on all occasions. There are two possible reasons for this, neither of them electronic.

· The Hipp toggle is only tripping when the pendulum arc is too small.
· The design or adjustment of how the pawl operates is wrong.

Absolutely! Easier said than done though, for someone like me. There are literally dozens of potential adjustment points on this movement, far more than with a commercial design, and of course, many are interdependent.

The pawl is vertical on this movement, and I found that it was dropping too far on some occasions. Adjusting a thumbscrew very slightly corrected this, and stopped two teeth being counted.

The Hipp toggle’s tripping is another matter though. The agate(?) block is a totally different shape to any I’ve seen in other HT clocks, with barely any well-defined V-notch. Maybe it’s very worn, but I suspect it’s always been like that. It’s operation is rather hit-and-miss, not like the precise action seen here (look carefully at the “dangly” bit on the pendulum at 0’ 08”): http://tiny.cc/moU7c (from the same site that Mike linked to - fascinating). The contacts have to be a mere 0.3mm apart when open to get it to operate reliably.

As to the supply voltage / current through the coil, I’ve had best results from about 12V which results in a nudge every 30s or so. I’m using a tiny SMPS wall-wart rescued from a scrapped scanner, though a lead-acid battery would be more authentic.

Anyway, after a lot more trial and error, I’m pleased to report that I seem to have got it working as-designed. It’s now run for 24h and kept time well, though I hope to be able to achieve greater accuracy in the long term (c.f. my parents’ very crude weight-driven longcase clock which manages better than 10s a week).

Setting-up is VERY critical indeed though, so I hope it stays working. If not, then maybe a discreet (and not necessarily discrete) electronic device with an opto sensor and counter to give the pendulum a nudge every 30s would be a better bet than my original LF oscillator plan.

What do people think of this circuit (fig. 9)? http://sound.westhost.com/clocks/free-pendulum.html

Next project will be the electromagnetic chimes – a bit of a nightmare as something just doesn’t seem right, and the previous owner felt they’d never been commissioned properly.

Here are some pics. Excuse the temporary PVC wiring and all the mangled screw heads (not me!).

1. The entire clock. I’ve removed the the door’s top panel and ebony bezel for repair. The lower door opens to reveal space for the battery. The case is beautifully made of inch-thick mahogany and all the glass bevelled J

2. The Bain-style pendulum bob and magnet that passes through its bobbin, at maximum leftwards excurstion.

3. The movement (minus some of the chime-related wheels). Note that the two contacts on the lower edge are part of the chiming mechanism, and nothing to do with timekeeping.

4. The Hipp toggle

Thanks to everyone who contributed, and to the mods for allowing it to appear here.

Nick.

[kalee20]

Just reading this thread, although I'm a bit late to contribute, it seems that there are 2 problems:

1) The Hipp toggle contacts may not be above suspicion;

2) The electromagnet gives a too powerful 'kick', making the pendulum swing excessively and advance the pawl two teeth.

Having an independent 30 second timer to energise the electromagnet won't work, because it won't be synchronised to give the pendulum a push at the right part of its swing (unless you use a carefully adjusted precision timer, in which case why bother with a pendulum at all?)

The way forward, seems to me to use the trip contacts to control base drive to a transistor, which then drives the pendulum. This solves the first problem, because the contacts will switch only a few tens of milliamps (enough to keep the contacts clean, but not enough for minute sparks. A small power transistor, a diode to catch the solenoid back-EMF, and a couple of resistors are all that is needed.

The second problem can be solved, as others have said, by tweaking the operating voltage to the electromagnet (or, winding the electromagnet appropriately for the supply). Alternatively, can you increase the mass of the pendulum bob, so that the impulse has less effect on velocity? This will lengthen the period marginally, (ideal pendula have a period independent of mass but practical ones have a rod with non-zero mass and this affacts the mass distribution). A heavy bob will increase the Q of the pendulum, improving stability.

[Nickthedentist]

Thanks for your words of wisdom, Kalee - never too late!

A relay is a recognised mod. for clocks like this, if only to prolong contact life. I may well do this (or what you suggest), and will at the very least, fit a spark quencher.

I used a variable voltage PSU to work out what current was best to drive the pendulum (home made many years ago around a LM317 regulator). 12V seemed like the best compromise, and doesn't cause the coil to heat much if the contacts stick closed (not impossible)

Nick.

[GeorgSc]

Hello Nickthedentist,
it is possible to operate a clock in this way, the pendulums of secondary clocks in ATO systems were actuated by pulses from the master clock.
There were Brillies which had a second winding on the coil, used to synchronize
the clock using impulses from a clock higher in hierarchy.
The general strategy in clock control of that kind is to have
the pendulum slightly slower than the pulse frequency.

But as a collector of electric clocks I would say: please try to
restore the original mechanism!
Would You mind posting a picture here? I am really curious
about a hombrew Hipp toggle.
Did You try to contact some collectors of electric clocks?
There are a lot around in UK!
Georg
PS
Such clocks need a reliable source of electricity.
Often they were energized from the batterys of a
telefon exchange.
For private use 1.5 Volts from a air depolarized cell
was popular.
ATOs, Brillie and Bulle clocks worked with 1,5 Volts.
Maybe that would solve some problems with Your clock?

[Nickthedentist]

For the benefit of Georg and anyone else who's interested...

I've just bought a fascinating book on electric clocks, which seems to be based on articles in the Percival Marshall publication "The Model Engineer". Not sure of the date, but seems to be 1930s. Amazingly, one chapter gives details about building the very clock I have here. Incredible!

It appears that the original design had a latching mechanism on the contacts (the spring steel bit in fig. 140 between E and I), which is missing from my example. When the toggle caused the contacts to close, they were meant to stay closed until a projection on the pendulum (P in fig. 135) released them. This seems to be a mechanical pulse-lengthening system.

The text says:

The successful features of this clock are that is has a contact lever which, when depressed into electrical contact, is locked in postition until the pendulum unlocks it towards the end of its swing or vibration to the right. This period of time between the locking and unlocking is the duration of the electrical contact, and can be varied at will.

The advantages of this propelment... is that the solenoid, when energised, is taking a long, gentle glide over the permanent magnet, entirely doing away with the sudden kick or shock which the electro-magnet and armature gave the pendulum. Again, in our permanent magnet, we have already a strong field waiting to be utilised; therefore, we are not expending battery power to set up that energy, and only the faintest of currents is necessary in the solenoid wire to give the pendulum the firm but gentle pull necessary to maintain its motion. The most important advantage, however, is that the current passing is so small that it has no burning effect on the contacts, and this latter is the secret of its sucess.

I find I need 12V and very minimal contact separation to get this clock to run at all. The text says it should run off two cells, i.e. much less.

I would consider building a latching mechanism as intended, but my skills and resources are very limited

Maybe an electronic pulse lengthener (555 and relay triggered by the original contacts) would be worth a go?

Nick.

[Top Anode]

I never cease to be amazed at the brilliance of past inventions. I had a gut feeling that there would be something very cleaver about the design since it lasted for 'over a hundred years'.

To have little contact erosion, and not jar the pendulum with a short pulse, is pure genius.

[Mike Phelan]

Nick
I'd be very inclined to reinstate the missing parts - have you drawings and precise dimensions? PM me if you want to take that route - I might be able to help.

A 555 as a monostable should work fine, though. It's a pity that the design of the coil and armature needed a longer pulse, as most Hipp designs did not need that complication.

TA
It was Matthaus Hipp's notched block, rather than the pulse lengthening, that was designed in the late 19th C. Using a longer pulse has no real advantage, and there should be no contact erosion as you would have used sliding contacts, or in more recent times, a diode.

[GeorgSc]

Hello Nickthedendist,
thank You very much for the pictures!
Finding a old book like this is a real coincidence indeed!
The mechanism depicted is a little bit strange.
I would think th latch mecanism is the lever L? The
reason of I/E could be some damping by friction, maybe?
What is the wheel A in fig. 135 good for?
Another question: what is the weight of the pendulum bob?
Hipp toggle drive depends on really heavy bobs,
my FAVAG 1/1 has a brass cylinder of about 10 kg!
I understood that Your clock has a moving coil at the
pendulum, which is rather uncommon. Most electric
clocks (with pendulums driven directly) have the coil
below the pendulum and some magnet or armature
on the pendulum.
Regards
Georg
Here's a link to a site on electric clocks.
Michel Viredaz is "The Authority" on electrical
clocks in general, of course very much so with
respect to the Swiss Hipp toggles. (FAVAG)
http://homepage.bluewin.ch/electric-clocks/index.htm

[Nickthedentist]

Thanks for your enthusiasm, T.A.

Originally Posted by Mike Phelan

I'd be very inclined to reinstate the missing parts - have you drawings and precise dimensions? PM me if you want to take that route - I might be able to help.

Hi Mike! Thanks for your comments, and the offer of possible help.

Originally Posted by Mike Phelan

A 555 as a monostable should work fine, though. It's a pity that the design of the coil and armature needed a longer pulse, as most Hipp designs did not need that complication... Using a longer pulse has no real advantage, and there should be no contact erosion as you would have used sliding contacts, or in more recent times, a diode.

That's interesting. I wonder whether a longer pulse is really needed, or whether the author was exaggerating this design's advantages. Would you advocate a diode in preference to a C-R network across the contacts to preserve them? Or would cheating with a relay be better?

Originally Posted by GeorgSc

The mechanism depicted is a little bit strange.
I would think th latch mecanism is the lever L? The
reason of I/E could be some damping by friction, maybe?
What is the wheel A in fig. 135 good for?
Another question: what is the weight of the pendulum bob?

Hello Georg! You're right of course, the latch is L. Not sure about I/E, but I'll look in the book when I get home.

The wheel A on the pendulum rod is what actually operates the mechanism. A lever with a roller on its end rests on the wheel, and is raised as the pendulum passes it.

I'm not sure of the mass of the bob, but it's certainly heavy, at least 5kg I'd say. It seems heavier than the driving weights you'd find on a normal, non-electric longcase clock.

Originally Posted by GeorgSc

I understood that Your clock has a moving coil at the pendulum, which is rather uncommon. Most electric clocks (with pendulums driven directly) have the coil below the pendulum and some magnet or armature on the pendulum.

Yes, I haven't seen made this way round. Do you think this may be why the pulse-lengthening device was incorporated?

Nick.

[GeorgSc]

Hello Nickthedentist,
No, this "reverse" drive ist somewhat more complicated,
one has to construct one or two isolated leads from
"outside" to the moving pendulum, which is the most
sensitive part in the "Bulle Clock", the only commercial
"reverse" I know.
But the force You can excert is not dependent on this.
The "normal" construction was used by ATO, Brillie,
Kundo, FAVAG which makes 4:1 against Bulle Clock.
This "normal" construction is just less complicated.
A factor which maybe was mentioned already in this thread,
is the ease of working of the "clock" mechanism.
When I got my FAVAG, it worked for about 15 seconds
between inpulses, cleaning and oiling the clockwork
improved that to 45 seconds.
Regards
Georg

[Mike Phelan]

Originally Posted by Nickthedentist

That's interesting. I wonder whether a longer pulse is really needed, or whether the author was exaggerating this design's advantages. Would you advocate a diode in preference to a C-R network across the contacts to preserve them? Or would cheating with a relay be better?

Nick
It's not really that critical, but a diode is the simplest idea; many contacts, in clocks and other things, made the contacts slide by a small amount. A relay would be OTT.

I'm not sure of the mass of the bob, but it's certainly heavy, at least 5kg I'd say. It seems heavier than the driving weights you'd find on a normal, non-electric longcase clock.

Not too much of an issue; less interference with the magnetic impulse, though.

Yes, I haven't seen made this way round. Do you think this may be why the pulse-lengthening device was incorporated?
Nick.

I don't see why that was done at all, as the "Hipp Trip" as it's referred to, could be designed in such a way to give any possible pulse length and position.