First Attempt at an Electronic Hipp Clock

The power of an electromagnet is proportional to ampere-turns rather than voltage, so you get the same effect from hundreds of amps through a single turn at miniscule voltage, or a few milliamps through thousands of turns, driven by tens of volts. The engineer can wind a magnet coil to suit almost any voltage / current combination. Telephone exchanges used to run on 80V, lorries run on 24V, cars on 12V, electronics on 9, 5 or fewer volts. Contactors run on mains voltage. Relays are available for all of them.

My reaction to John mentioning his pendulum twangs is that the impulse is too powerful and/or slightly mistimed.

There are two cases though:

1. If the pendulum is started from stop by the electromagnet it has to be powerful enough to overcome bob weight inertia. As this is considerable when the bob weighs a few kilograms, a powerful magnet is needed.
2. If the clock is started by hand, then the electromagnet only has to be powerful enough to keep the bob swinging once it's going. The amount of energy needed to keep a well-made pendulum going is low because the main losses are stirring air and heating up the suspension spring, A weak electromagnet will do the job, ideally being only just powerful enough because excess energy upsets time of swing.

Timing is also vital. Like pushing a child on a swing, the impulse has to be timed just so. When the electromagnet is mounted directly below the bottom dead centre, it has to be pulsed just before the bob reaches BDC so that the bob accelerates, and then turned off quickly. If the electromagnet is still on when the bob reaches BDC, it becomes a brake, either removing energy from the pendulum with a bump, or stopping it entirely.

Dave

Edited By SillyOldDuffer on 31/01/2023 10:38:33

SK
I had originally planned to use opto interrupters as I had them (Harris H2182 ) but after playing with the pendulum when first made it I found it was very easy to impart some sideways motion when moving it off to one side and releasing it to set it going. As the slots in my sensors are only 3mm wide, and there would be 4 of them spaced over about 2-2 1/2", I felt there was too much chance of the interrupting vane on the pendulum impacting the side of one of the sensors, hence the decision to use Hall devices. There were two other potential problems with with these devices: stray light activating them and they are quasi linear devices were the output doesn't switch rapidly when the beam is interrupted. John Haine has got around these drawbacks by using Sharp optos which have a 10mm wide gap and a built in Schmitt trigger output, which still leaves the possible stray light problem.

As to the temperature effects on the Hall effect devices, the data sheet I have shows that there is a maximum 10 Gauss change in the actuating or release flux between 0 and 25 deg C. Since the magnet I am using (.125" dia x .125" long neodymium ) has a magnetic flux of ~1200 Gauss at a distance of 1/16", the spacing between the magnet and the sensor I'm using, (according to info from the web site linked to by John Haine in another post here ) the difference in the position that the device would be activated for a 10 Gauss difference, I feel would be insignificant.
I don't think it would make any difference anyway as the time keeping is a function of the pendulum period and exactly when the impulse is sent to the clock would make no difference, only that the spacing between seconds might not be equal, but averaged over time would be correct.

Dave

About your comment about the "twang" I think you are correct in saying the the impulse is more powerful than needed. I am going to try and put some resistance in series with the coils and see how far I can lower the voltage and still get sufficient impulse to keep the pendulum going. Or I could also just lower the coils to increase the distance from the coils to the armature on the pendulum rod to decrease the magnetic flux.

I don't think there is a problem with the timing as the coils are de-energized just as the leading edge of the armature reaches the centre line of the coils.

John

Edited By John Purdy on 31/01/2023 22:39:23

Just a data point, my small clock has a 25cm pendulum with a tiny neo magnet on the end swinging above a small air cored coil which has a resistance of about 200 ohms. It's impulsed for about 15ms every other swing or so from an Arduino output through another 220 ohm resistor. You really don't need much energy or force to impulse a reasonably high Q pendulum.

Impulse timing matters if the strength or phase of the impulse changes. If the phase centre of the impulse is at the centre of the swing it doesn't change the pendulum phase.

Hi all

my post about 5v v 50v, I know that current determines electromagnet strength, I was aluding to the fact that a lower voltage would require a lot more current than higher voltages to produce a reasonable amount of electromagnetic force.

Oh! well, back to sleep, somebody wake me up when an interesting post arrives.

.

This is your Wake-Up Call, Howi

It’s about the energy required to impulse a Synchronome pendulum, and interesting at several levels.

.

.

Credit: Frank Hope-Jones

.

MichaelG.

Alas, Fedchenko's clock that impulses one per cycle at the base of the bob, rather shows that F H-J was wide of the mark in his conclusion.

.

but that doesn’t preclude it from being interesting … and specifically, it does address Howi’s earlier comment about the energy required to maintain a pendulum in its swing.

[quote] I would have thought 5v power to the coils would not be enough to move a long pendulum and weight. [/quote]

MichaelG.

Edited By Michael Gilligan on 01/02/2023 13:37:43

Without knowing how the impulse was applied by FHJ, and what he counts as external friction, it's difficult to comment. As JH says, one of the worlds most accurate clocks applied the impulse below the bob

.

I wasn’t actually inviting comment … Simply waking-up Howi by posting something interesting.

The book is widely available.

MichaelG.

.

I think everyone interested in this topic should at least read it,

but If you don’t want a hard copy:

https://clockdoc.org/gs/handler/getmedia.ashx?moid=57393&dt=3&g=1

Edited By Michael Gilligan on 01/02/2023 14:16:37

For clarity … the part of Fig. 82 that is most relevant here is the baseline energy requirement:

.

.

… Which is independent of the points on which John and Duncan have commented.

.

MichaelG.

Thanks for pointing out that book. The introductory comment about quartz clocks should be accompanied by a sober-sounding narrator intoning "little did they know ..."

Unfortunately, it's written in that overly wordy and flowery fashion that was common in earlier times, which makes it a slog for modern readers.

So that figure 82. What are the points being made? Seems like:

(1) That the amount of energy needed purely for keeping the pendulum in motion is the same whether spread over 30 impulses or applied in one larger impulse? Seems logical.

(2) That more energy is wasted if 30 small impulses are applied vs. 1 larger one? Also seems logical in a mechanical system, but likely irrelevant in magnetically-impulsed systems.

(3) That more energy is wasted if the impulse is applied at the bottom vs. near the top of the pendulum? I have more trouble accepting this, but I could imagine that additional friction could be encountered at the bottom because a larger throw of the impulsing mechanism could be necessitated due to the larger amplitude near the bottom. But again, this supposed additional wasted energy is likely irrelevant in magnetically-impulsed systems.

As for the ideal position for impulsing, I'll stick with the center of percussion, which would be near the bottom in a pendulum with a heavy bob.

One other point made is that the mechanical impulse was designed to be applied with nearly zero force at its onset, growing to some maximum before release. I think this could be pertinent to magnetically applied forces as well, and I had imagined using an RC network or similar to avoid an abrupt application of force.

Edited By S K on 01/02/2023 16:05:22

There's a certain amount of post-hoc rationalisation here. The shape of the "impulse slope" of the Synchronome pallet is essentially a circular arc or 0.75" radius. Apparently that's what they were made with but when F H-J was presenting a paper at the IEE, William Shortt who was in the audience worked out what the shape should be to apply a "raised cosine" impulse that started gradually, built up to a maximum, and then rolled off again to zero, but that shape isn't the circular arc. With the circular arc I think the idea is that the roller drops on to the pallet just before the start of the slope but that depends on the clock being very precisely set up. There's a time-lapse video on-line (or used to be, the page seems to have been removed from Brian Mumford's site) that shows this is an idealisation. Actually, except for considerations of mechanical "ringing", the shape of the impulse is virtually irrelevant for any reasonably high-Q pendulum, all that matters is the position of its phase centre. Nasty spiky forces just amount to higher order harmonics that the pendulum filters out.

.

I dug it out mainly because I remembered that FH-J had quantified the energy required to keep a Synchronome pendulum swinging … and such information is relatively scarce.

Given Howi’s remark about Voltage, I thought it worth posting here.

We all know [or can easily find] the design details for a Synchronome pendulum, so it seems like a convenient benchmark for anyone wishing to start designing a clock.

MichaelG.

I've just been experimenting with my clock and it is very easy to induce the dreaded "twang" by varying the timing of the impulse and the position of the coil. If you are getting a "twang" then essentially some of the impulse energy is being dissipated in unwanted vibration (and corresponding disturbance of the pendulum motion), so it is well worth figuring out why this is happening - I suspect it will actually prove to be a timing issue in your case. If the timing is wrong, then the magnet will initially accellerate the pendulum, but if you don't switch it off soon enough (i.e., before the swing reaches the point of max attraction) it then acts as a brake for the last part of the "impulse" and causes the pendulum to twang.

My impulse coil activates with the pendulum near enough at BDC, but I have offset the electromagnet sufficiently to the right of BDC so that the impulse is guaranteed to have finished before the armature attached to the pendulum reaches/passes it.

I can induce the "twang" very easily by shifting the coils left too far and/or by making the pulse length too long. With my current setup the pendulum swings for about a minute between impulses and the impulses are completely silent.

The two sensors that determine the pulse length are set approx 13mm apart and the coils are offset approx 35mm from BDC.

.

Not sure if this will be of any interest, but it’s a plot that I did for someone back in 2018, so I might as well chuck it in the pot here:

.

.

.

.

MichaelG.

I've just moved the coils down so the gap between the armature on the end of the pendulum rod and the coils is 3/8" instead of the 1/16" previously but kept the same lateral distance from BDC as before. The impulses are now at 32 second intervals vice the 3 1/2 minutes previously and the "twang" is gone. Which seems to indicate that the magnetic flux was much stronger than necessary at the closer distance.

John

It's great to see that enginners are still experimenting with the idea of making a "free pendulum" clock using modern electronics. My two clocks are still going well after about 6 years and I find the best way to live with them is to ignore the day to day variations in seconds. I decided that unless perfect correction for temp, pressure, humidity and even perhaps gravity (tides !) is incorporated once a well built clock is regulated as best as is possible it will average out to be a good timekeeper. A friend who has kept records of the going of a Synchronome clock by computer comparison with NTC reckons he can see the changes through the seasons of the year. It seems to me pressure variations are the greatest issue.
On the question of Post Office clocks, although the exchanges are (were?) powered by 50v batteries only the contact making circuits were connected directly from that voltage (50v relays and series loop clock circuits). The circuit for driving the Hipp controlled drive magnet were electrically separate in the clocks and were fed from 50v via a high value resistance. The drive EMs of a PO clock are about 9ohms and will work happily from 4.5 volts.