Purpose of pendulum suspension springs

[karlmansson]

Hello!

I was asked the other day about the purpose of the suspension springs often used in pendulum Clocks and I realized that I hadn't really though about it.

What is their purpose?
My initial thought was that it might be way to counteract cicular error. As a spring stores exponentially more energy the more it is flexed a higher amplitude would result in a greater return from the spring. In terms of escapement error it might help to simulate the effect of an epicycloidal pendulum arc by accellerating the pendulum when it Changes direction.

Or maybe it's to shield the movement from the impulse delivered to the pendulum rod?

Am I thinking along the right lines?

Best regards

Karl

 

[shutterbug]

I suspect that it has minimal usefulness, but it can affect the timing. Some clocks (cuckoo's) use a trapeze arrangement without a spring, and some French clocks use a silk loop (like a trapeze) without a spring. I've never tried it, but I would guess that most clocks would run with a similar arrangement. The spring is just easier.

 

[karlmansson]

Why not just hook it over the pendulum post then? Friction?

 

[gmorse]

Hi Karl,

That's effectively what the knife-edge suspension on a verge is, and the problems with those are friction and the associated wear, which over time alters the geometry of the escapement.

Regards,

Graham

 

[karlmansson]

I see!

So no effect on circular error? Could it be made to have an effect on circular error?

Best

Karl

 

[MartinM]

Beyond the ability of a too-heavy spring to stop a clock, I've always just thought it was simply a piece of metal with enough mass to keep the pendulum suspended and that it is a spring because of their ability to resist fatigue. And, there is the added benefit of removing wear from the equation, as is seen in other suspension methods.

 

[Tinker Dwight]

I see!

So no effect on circular error? Could it be made to have an effect on circular error?

Best

Karl

The effect is so small as to not count for much.
If gravity were not there, it would be the only
restoring force and be like a balance wheel.
I'm afraid its primary purpose is suspension.
On older tall clocks they used longer springs in
attempts to reduce the spring effect on timing
but most modern clocks just use a short spring
that works fine.
Tinker Dwight

 

[john e]

Hello!

I was asked the other day about the purpose of the suspension springs often used in pendulum Clocks and I realized that I hadn't really though about it.

What is their purpose?
My initial thought was that it might be way to counteract cicular error. As a spring stores exponentially more energy the more it is flexed a higher amplitude would result in a greater return from the spring. In terms of escapement error it might help to simulate the effect of an epicycloidal pendulum arc by accellerating the pendulum when it Changes direction.

Or maybe it's to shield the movement from the impulse delivered to the pendulum rod?

Am I thinking along the right lines?

Best regards

Karl

The suspension spring can indeed be a compensation for circular error, in the exact way you have mentioned.

It is also possible to modify the spring at either the top of the spring or the bottom in order to alter the compensation. This is detailed at length in Goodrich's "The Modern Clock", reprinted 1984, pages 83-84, details this at length.

I do not recall if he also mentioned that when the movement is small, like a degree or two or three, the bending of the spring is essentially lossless, meaning that it returns "all" the flexure energy back to the pendulum. So cheap, easy, and little hit on energy loss. Tis why they are still so popular.

John

 

[eskmill]

Fedchenko's suspension spring is indeed a solution for circular error correction. It forces the pendulum into a parabolic arc. However, Fedchenko's clock has a very short arc.

 

[eskmill]

More...thoughts (my editing time for my reply #9 expired)

Karl's thinking is in the right direction. The mechanics of pendulum suspension is complex and many theorists have looked for a more perfect pendulum. The solution finally arrived as a mathematically designed suspension spring invented by Fedchenko during the late 1940's or a little later. Fedchenko's suspension spring does force the pendulum into a parabolic arc.


However, Fedchenko's clock has a very short arc. Fedchenko's precision pendulums are supported within a vacuum chamber much in the same way as were the Shortt Synchronome clocks.


The ideas quoted in Goodrich's "The Modern Clock" using cycloidal "cheeks" did force the pendulum motion into a paraboilc motion but there are many other errors within the machanics that restore the lost forces that keep the simple pendulum in motion far outweigh the effect of the circular error.


A simple example of circular error correction can be observed in the motion of the jointed wire pendulum of the Comtoise or Morbier clocks. The recoil of the Morbier simple escapement imposes a force that causes the bob to move more-or-less in a parabola.

 

[Tinker Dwight]

My statement wasn't to say that one couldn't use the suspension spring
for compensation, it was that as typically used, it was there to suspend
the pendulum and little else.
As for circular error, I've always thought that a magnetic compensation
could be made that would be better than other mechanical methods.
Tinker Dwight

 

[MartinM]

I wonder why more research and/or development was not done on attempting to create a precision regulator using a torsion pendulum.
In that system, the effects of gravity are almost nil.
Compensation for temperature variance could be easily done with a device that changes the effective length of the spring and would need to accommodate for change in the length of the suspension as well as the change in the spring's esasticity.
Something would need to be done about balancing the input power across the wind, as well. (Weight or fusee, for example).

 

[john e]

The ideas quoted in Goodrich's "The Modern Clock" using cycloidal "cheeks" did force the pendulum motion into a paraboilc motion but there are many other errors within the machanics that restore the lost forces that keep the simple pendulum in motion far outweigh the effect of the circular error.

I didn't bother going over his verbage on cycloidal cheeks, as that wasn't part of the question. I was only speaking about how he considered the suspension spring as part of the isochronal adjustment.

As for circular error, I've always thought that a magnetic compensation
could be made that would be better than other mechanical methods.
Tinker Dwight

Do you mean eddy current dissipation as it travels too far? If so, the eddy losses will scale with the gap dimension, and the magnetic gap would extend it's effect at least two or three gaps distance from the edge of the magnet assembly. If that's not what you speak of, I'd love th hear your idea.

I wonder why more research and/or development was not done on attempting to create a precision regulator using a torsion pendulum.
In that system, the effects of gravity are almost nil.
Compensation for temperature variance could be easily done with a device that changes the effective length of the spring and would need to accommodate for change in the length of the suspension as well as the change in the spring's esasticity.
Something would need to be done about balancing the input power across the wind, as well. (Weight or fusee, for example).

Barometric pressure is an issue. Torsion spring material is a very big one, I think they're invar now? And I recall some major issues with the alloy. Granted, the rotational mass is consistent, but so is gravity.

Maybe it was the materials problem that gave the simple pendulum the advantage.

John

 

[MartinM]

...Barometric pressure is an issue. Torsion spring material is a very big one, I think they're invar now? And I recall some major issues with the alloy. Granted, the rotational mass is consistent, but so is gravity.

Maybe it was the materials problem that gave the simple pendulum the advantage.

John

Not Invar, but similar in certain respects. There were several different bronze or nickel-steel-chromium alloys used, Horolovar's "Ni-Span C" being the most successful.
Barometric pressure could be mostly defeated by placing the clock in a cloche with the air mostly evacuated (Like pretty much all of the high-accuracy clocks.).

 

[Tinker Dwight]

Hi John
Many single transistor balance wheel clocks us Eddy currents
to limit the balance wheel travel. It doesn't seem to have bad effects
on the rate but that isn't what I meant.
I just meant a simple compensator with pairs of magnets on either
side. As the pendulum swings, the magnets repel, increasing
the restoring force. If the surfaces of the magnets were shaped
right, the addition force could balance needed addition restoring force.
Tinker Dwight

 

[john e]

Hi John
Many single transistor balance wheel clocks us Eddy currents
to limit the balance wheel travel. It doesn't seem to have bad effects
on the rate but that isn't what I meant.
I just meant a simple compensator with pairs of magnets on either
side. As the pendulum swings, the magnets repel, increasing
the restoring force. If the surfaces of the magnets were shaped
right, the addition force could balance needed addition restoring force.
Tinker Dwight

Sounds pretty neat. Making the magnets accurate enough in strength would be an issue. The assumption is that given a profile, the setup could have some adjustments to finalize the setup to maximize isochronism.

I don't know how lossy the system would be though. Could be a jump from the frying pan....

John

 

[john e]

Not Invar, but similar in certain respects.

I was going by the Atmos spring, which I believe is Elinvar. From what I recall, the spring has to have a controlled expansion coefficient as well as remain linear and elastic through temperatures, as well as good fatigue properties. (not asking much, no?)

There were several different bronze or nickel-steel-chromium alloys used, Horolovar's "Ni-Span C" being the most successful.
Barometric pressure could be mostly defeated by placing the clock in a cloche with the air mostly evacuated (Like pretty much all of the high-accuracy clocks.).

Thanks. Although, I believe an Atmos would not not work well in a controlled vacuum.

John

 

[shutterbug]

I recall a torsion clock being made that would be millennial, promising accuracy withing a second in 1000 years or so. It's underground, mostly, and if memory serves is about a mile deep. I'll try to find it and report back

Edit: Even more impressive, but not as deep as I remembered . Here's a link.

 

[Tinker Dwight]

I recall a torsion clock being made that would be millennial, promising accuracy withing a second in 1000 years or so. It's underground, mostly, and if memory serves is about a mile deep. I'll try to find it and report back

Edit: Even more impressive, but not as deep as I remembered . Here's a link.

I believe the clock was self correcting. My understanding is that the sun's passing
would cause the clock to speed up or slow down.
Not exactly a clock that ran at a constant rate.
Tinker Dwight

 

[novicetimekeeper]

Returning to older clocks, in particular longcase, it was said earlier that the long suspension feathers were to overcome the effect of any spring in the feather, have I got that right?

Any idea how they came up with the design? Was it in Huygens original design?

The verge pendulum doesn't work like that at all does it, isn't it directly attached to the verge? How did the long crutch and the two blocks on a springy steel come about. Were the original steel bits even springy?

(I have a wall clock with a bit of copper strip as a feather and that seems to work fine though clearly not original)

 

[Tinker Dwight]

I don't know why they use the long feathers. I'd always assumed it was
to minimize the spring force on the pendulum.
The mass of the pendulum caused it to bend, mostly near the top
end in more of a cantilever fashion. I suspect that the main reason was
that making .004 inch thick spring was not that easy then.
The verge escapements ran with limited mass pendulums because
the pivot bearing had to support the weight. This made the iregularities
of the drive force more significant in effecting the rate of the clock.
It is all about the evolution of clocks. Using the separate crutch and
independently supported pendulum was a significant improvement.
Tinker Dwight

 

[novicetimekeeper]

I don't know why they use the long feathers. I'd always assumed it was
to minimize the spring force on the pendulum.
The mass of the pendulum caused it to bend, mostly near the top
end in more of a cantilever fashion. I suspect that the main reason was
that making .004 inch thick spring was not that easy then.
The verge escapements ran with limited mass pendulums because
the pivot bearing had to support the weight. This made the iregularities
of the drive force more significant in effecting the rate of the clock.
It is all about the evolution of clocks. Using the separate crutch and
independently supported pendulum was a significant improvement.
Tinker Dwight

Did that come with Huygens, do you know, or was that a later development. I see an awful lot of 18th century clocks but I never get to see any really early longcase. I've looked at verge bracket clock movements, just not sure how the developments took place. I think I need another book!

 

[Tinker Dwight]

There is a drawing showing a Huygens design with verge escapement,
crutch and suspension spring.
When he first did this is not in my knowledge.
Tinker Dwight

 

[Berry Greene]

What a fascinating thread. Wow!
Rgds BjG

 

[novicetimekeeper]

Now some time later I think the feather (Spring suspension) does go a way to compensate for circular error as the pendulum moves up and down in the crutch fork/slot so must be altering length.

The silk suspension you see on French clocks was used with cheeks that were there to compensate for circular error. How successful all this was can be seen in the accuracy of the clocks, as they can easily manage really good timekeeping on the most basic movements. One of my 30 hours was regularly just a minute out after a month.

 

[Berry Greene]

Thank you novicetimekeeper in Dorset! Interesting stuff.
A lot of clever people have applied their minds to all these things long before they were as well equipped as we are now. What intuition they displayed. What patience they possessed. What endurance too. We grab a rich slice of history when we open up these old clocks don't we? Thank you for your contribution to this excellent Thread.
Sincerely, Berry G Chichester Sussex

 

[Berry Greene]

I wonder why more research and/or development was not done on attempting to create a precision regulator using a torsion pendulum.
In that system, the effects of gravity are almost nil.
Compensation for temperature variance could be easily done with a device that changes the effective length of the spring and would need to accommodate for change in the length of the suspension as well as the change in the spring's esasticity.
Something would need to be done about balancing the input power across the wind, as well. (Weight or fusee, for example).

Interesting thoughts. I want for the torsion too but I would take issue with the idea that gravitational effects are minimal. I don't think they are, as a similar idea has been used to detect earthquakes. More than that - it has been my experience that tiny changes in the weight of the pendulum have quite a large effect on the regulation. Th length of the suspension would be a neater way to fine trim. The equal power is also a big issue even before we get to flutter.
Hey hey I didn't come here for this and we're off topic. So sorry! I came to learn if possible or discuss pendulum suspensions. I'll ask elsewhere. Rgds, BerryG

 

[Berry Greene]

I came here looking for more practical, rather than theoretical, issues on the pendulum suspension - interesting though it is. The stiffness of the spring must be an issue when it comes to the selection of a replacement. Is there a measurement method? Just as important, it seems to me, is the issue of spring bias. In a movement with (maybe) reduced power, it presents as insufficient over-swing. If the spring is too heavy it can even stop the clock. The adjustment allowance of the suspension post clamping screws does allow for a small amount of pallet depth and spring bias adjustment. Some posts are riveted, like the Smiths K6A that I have here, and some do have a screw fixing which allows complete rotation of the post & slot to which the spring is pinned.
Do you have a procedure or method of setting up?
1) Spring selection with regard to strength. The length is reasonably obvious.
2) Neutralising the spring bias to the plumb vertical.
3) Setting the beat with the crutch to verge alignment

If I wasn't struggling here I would not be asking. Maybe I've been lucky in the past.
I have tried different springs and fine tuning and the clock runs and maintains but the pendulum swing is so small. ..... Speak to me please.... Rgds to you all,
BerryG

 

[R. Croswell]

I came here looking for more practical, rather than theoretical, issues on the pendulum suspension - interesting though it is. The stiffness of the spring must be an issue when it comes to the selection of a replacement. Is there a measurement method? Just as important, it seems to me, is the issue of spring bias. In a movement with (maybe) reduced power, it presents as insufficient over-swing. If the spring is too heavy it can even stop the clock. The adjustment allowance of the suspension post clamping screws does allow for a small amount of pallet depth and spring bias adjustment. Some posts are riveted, like the Smiths K6A that I have here, and some do have a screw fixing which allows complete rotation of the post & slot to which the spring is pinned.
Do you have a procedure or method of setting up?
1) Spring selection with regard to strength. The length is reasonably obvious.
2) Neutralising the spring bias to the plumb vertical.
3) Setting the beat with the crutch to verge alignment

If I wasn't struggling here I would not be asking. Maybe I've been lucky in the past.
I have tried different springs and fine tuning and the clock runs and maintains but the pendulum swing is so small. ..... Speak to me please.... Rgds to you all,
BerryG

1) Spring selection with regard to strength. The length is reasonably obvious.
Ideally the suspension spring should be "invisible" to the clock from a mechanical standpoint to allow the pendulum to swing as close as possible to its natural rate influenced only the acceleration due to the earth's gravity. To that end some clocks do not use a suspension spring at all. Many older French clocks suspend the pendulum with a silk thread, some others use a knife edge in a "V" notch, and of course the common cuckoo clocks use a clevis and pin suspension. A suspension spring must be strong enough to suspend the weight of the pendulum, it must be strong enough to maintain the direction of swing, and as thin as possible to have as little effect on the pendulum rate as possible. From a repair or replacement standpoint, one should replace suspension springs with the original size (if known). A thicker (stronger) spring will be less likely to be damaged and perhaps add a bit more stability to the swing at the cost of influencing the natural rate of the pendulum and how fast the clock runs.

2) Neutralising the spring bias to the plumb vertical.
The spring should be flat and straight in line with the attached leader so the entire pendulum suspension system is straight. It should be attached to a post that has a true vertical slot, and when there is a rate adjuster, the chops through which the spring passes should also be true vertical. There should be no need to neutralize bias if the assembly is straight and hung vertically. Never adjust "beat" by bending any part of the pendulum suspension.

3) Setting the beat with the crutch to verge alignment
Simply put, you adjust this alignment so that that with the crutch at the center resting position (on a vertical line directly under the point where the suspension spring is attached) the crutch must be moved the same distance right or left of center to cause the escapement to release an escape wheel tooth. The actual adjustment depends on the design of the clock. Most American mantel clocks the adjustment is made by simply bending the crutch wire. When the crutch is a flat metal strip the adjustment is usually made at the verge where the crutch attached to the verge arbor. Caution, if the locks and drops off the verge pallets are not equal it will be difficult to adjust the clock to sound like it is in beat.

RC

 

[Berry Greene]

Thank you so much for that very full and prompt reply which I am in your debt for. I have ordered some more suspension springs as I think this must be a combination of it being too stiff and maybe reduced power in the train. One positive seems to be that the accuracy is very good. The movement is maintaining OK and is plumb level - which is where this started from. I could not get any operation unless the case was heavily canted up to one side. Eventually I reversed the spring and it needed to be canted the other way. That led me to change the suspension spring, hence my references to the compliance, the "bias" - and its adjustment. I began to realise that I don't have a very deep understanding of the suspension spring and the part it plays in the operation. Yes of course, when prompted, I do acknowledge the cuckoo clock trapeze alternative, but the silk used by some French movements is not something I have ever seen or come across before. Yes I can see a compliance advantage but it plays no part in reducing circular error or promoting a cycloidal swing. Not, I think, that is worth chasing for on the kinds of clocks that I have anyway!
I have taken a copy for my files of your excellent treatise on these more practical aspects of pendulum operation and setting up and can only thank you once more.
Best regards, BerryG

 

[R. Croswell]

Thank you so much for that very full and prompt reply which I am in your debt for. I have ordered some more suspension springs as I think this must be a combination of it being too stiff and maybe reduced power in the train. One positive seems to be that the accuracy is very good. The movement is maintaining OK and is plumb level - which is where this started from. I could not get any operation unless the case was heavily canted up to one side. Eventually I reversed the spring and it needed to be canted the other way. That led me to change the suspension spring, hence my references to the compliance, the "bias" - and its adjustment. I began to realise that I don't have a very deep understanding of the suspension spring and the part it plays in the operation. Yes of course, when prompted, I do acknowledge the cuckoo clock trapeze alternative, but the silk used by some French movements is not something I have ever seen or come across before. Yes I can see a compliance advantage but it plays no part in reducing circular error or promoting a cycloidal swing. Not, I think, that is worth chasing for on the kinds of clocks that I have anyway!
I have taken a copy for my files of your excellent treatise on these more practical aspects of pendulum operation and setting up and can only thank you once more.
Best regards, BerryG

Here's a French silk thread suspension. Can't get much simpler than that. Spool up more thread and the clock runs faster. The silk has zero bias. Both legs of the loop go through the brass post so the pendulum always is positioned correctly.

RC

 

[Berry Greene]

Wow-ee! An eye-opener to the simple. Got it!
Whats the down-side? Learned so much from this site.
Thanks again - BerryG