Visual Beat Error and hairspring compression

[jdrichard]

I have an Elgin Veritas Pocket Watch that I have serviced and it has a 1.9 second beat error and a 270 deg Amplitude. When the balance swings clockwise and compressed the hairspring, the impulse jewel stops at the 8 o’clock position. When it swings in the other direction, the impulse jewel stops at the 2 o’clock position, which is around a 90 degree delta in where it should be if it swung the exact amount in clockwise  as it did in counter clockwise.

So the big question is, does the compression of the hairspring put additional resistance in the rotation of the balance then the decompression of the hairspring in counter clockwise motion? And if I correct the beat error, will the position in the impulse jewel actually be in the center of the banking pins. Illustrations attached. Red line is impulse jewel and green line is where the impulse jewel would be if all was symmetrical in where the balance stops in both rotational direction.

[Nucejoe]

49 minutes ago, jdrichard said:

 does the compression of the hairspring put additional resistance in the rotation of the balance then the decompression of the hairspring in counter clockwise motion? 

Do you mean additional load?   or     Additional resistance? 

Additional resistance due to additional load is infinitesimal and practically unnoticable.

53 minutes ago, jdrichard said:

 And if I correct the beat error, will the position in the impulse jewel actually be in the center of the banking pins. 

It will providing,

-Banking pins are positioned true.

-Fork is out.

Fork may stop somewhere else if somehow remains engaged with impulse pin or condition inside the fork crown, depending on the condition at the other end of fork " pallets" .

True elgin is good grade, but I don't know how good is the grade, you want to see near true pallet adjustment !!     have a look at an audemars piguet or Rolex.

Regs 

Joe

[praezis]

You must take account of the beat error: it is caused by an angle that makes the swing unsymmetrical.

1 hour ago, jdrichard said:

And if I correct the beat error, will the position in the impulse jewel actually be in the center of the banking pins. Illustrations

Just try 
I use to work the other way: center the pallet/impulse jewel -> get low beat error.

Compressing and widening the hairspring both takes energy. If exactly the same amount? I leave it to the physicists...

 

Frank

[jdrichard]

1 hour ago, praezis said:

You must take account of the beat error: it is caused by an angle that makes the swing unsymmetrical.

Just try 
I use to work the other way: center the pallet/impulse jewel -> get low beat error.

Compressing and widening the hairspring both takes energy. If exactly the same amount? I leave it to the physicists...

 

Frank

Thanks Frank, I do agree and then need to research the physics of the hairspring 

[JohnR725]

5 hours ago, jdrichard said:

if I correct the beat error, will the position in the impulse jewel actually be in the center of the banking pins.

One minor little concerned here?  This is an American pocket watch  the banking pins are movable  how do you know there where there supposed to be?  Otherwise try to find something else to reference the alignment with.

[Nucejoe]

1 hour ago, JohnR725 said:

One minor little concerned here?  This is an American pocket watch  the banking pins are movable  how do you know there where there supposed to be?  Otherwise try to find something else to reference the alignment with.

Yes, if banking pins are not positioned true, then where is their midpoint.

If memory serves me right, nickelsilver once said, ole American watches run accurate even if banking pins are not well adjusted.

[jdrichard]

6 hours ago, Nucejoe said:

Do you mean additional load?   or     Additional resistance? 

Additional resistance due to additional load is infinitesimal and practically unnoticable.

It will providing,

-Banking pins are positioned true.

-Fork is out.

Fork may stop somewhere else if somehow remains engaged with impulse pin or condition inside the fork crown, depending on the condition at the other end of fork " pallets" .

True elgin is good grade, but I don't know how good is the grade, you want to see near true pallet adjustment !!     have a look at an audemars piguet or Rolex.

Regs 

Joe

Additional resistance from the hairspring contracting.

[jdrichard]

1 hour ago, JohnR725 said:

One minor little concerned here?  This is an American pocket watch  the banking pins are movable  how do you know there where there supposed to be?  Otherwise try to find something else to reference the alignment with.

You are correct. I have adjusted banking pins in the past in the older American PWs, just to get the max amplitude.

[JohnR725]

1 hour ago, jdrichard said:

I have adjusted banking pins in the past in the older American PWs, just to get the max amplitude.

Then just as a reminder while people move the banking pins to adjust the amplitude their purpose is a safety feature and their supposed be adjusted for that function then you're supposed to play with the stones afterwords. Otherwise you'll have your safety features not set correctly.

[HectorLooi]

This has always been something at the back of my mind. Is the angular force equal in compression and expansion phase of the spring. Logically it might be symmetrical to a certain angle of travel. Then the geometry of the spring should make it different.

Maybe that's why when you visually centre a balance wheel to the banking pins, the timegrapher tells you otherwise. 

[JohnR725]

1 hour ago, HectorLooi said:

Maybe that's why when you visually centre a balance wheel to the banking pins, the timegrapher tells you otherwise. 

I'm attaching an image you don't actually need the banking pins to put the watch in beat. You're supposed to line the roller jewel up with that invisible center line that conveniently is easy to see if it was between the banking pins.

 

[Klassiker]

If I understand the question correctly, it is "is the hairspring stiffer (i.e. more powerful) when contracting than when expanding? Or put another way, is the force to cause a tiny deflection with the bend higher than to cause the same amount of deflection against the bend."

As long as it is the same deflection, then no. I think it will be the same force in both directions. However, I don't think the deflection of the spring remains in a constant ratio to the angular deflection of the balance wheel. The number of turns on the hairspring reduces during expansion, so each degree of deflection will cause proportionally more bending (or straightening in this case). This means the hairspring gets stiffer on expansion, and softer on contraction.

I've managed to convince myself at least. Does that make sense?

[JohnR725]

I was having strange probably unhelpful thoughts regarding Springs. I was thinking about the hairspring kinda like a mainspring we could look at the power curve of one of those but. Then of course the problem is we don't wind the spring up very much so mainspring power curves would be out.

I was wondering if you take the fork out and wind the balance wheel exactly one turn in one direction and let go and wait until it comes to a stop then wind that the other direction whether that would get us anything at all?

13 hours ago, Nucejoe said:

If memory serves me right, nickelsilver once said, ole American watches run accurate even if banking pins are not well adjusted.

It's actually amazing how well pocket watches run with thing is not right. Banking pins are only for horn clearance.

Then I'm currently being amused looking for something on adjusting banking pins. Mainly because all the references dear friend in? In other words the reference material in finding on adjusting banking pins all seems be slightly different and the order in which they're doing their adjustments is different.

I was thinking maybe we need one of these I wonder how hard it would be to clone one of these? Interesting machine that probably has a price we do not want to go into. You can measure lift angle but you really need it is still measured the amplitude whether it has lift angle or not.

https://www.witschi.com/en/group-of-devices/measurement-of-mechanical-watches/WisioScope S/WisioScope S.html

 

 

20200604_Witschi_WisioScope_Data Sheet.pdf

[praezis]

Action of the hairspring is the same as bending a clamped rod: as far as Hooke's law is valid, you will need the same force to bend the rod up or bend it down.

Characteristics of a hairspring is
D (Direktionsmoment in German) = M / α   (= applied moment / rotated angle)
and/or
D = EI / L  (= material and geometric properties)

D indicates no difference bending a hairspring in one or the other direction (just the sign will change).

Frank

[praezis]

Action of the hairspring is the same as bending a clamped rod: as far as Hooke's law is valid, you will need the same force to bend the rod up or bend it down.

Characteristics of a hairspring is
D (Direktionsmoment in German) = M / α   (= applied moment / rotated angle)
and/or
D = EI / L  (= material and geometric properties)

D indicates no difference bending a hairspring in one or the other direction (just the sign will change).

Frank

[Nucejoe]

17 hours ago, jdrichard said:

Additional resistance from the hairspring contracting.

If I am not misunderstanding your point again,

With the fork out, impulse pin ends up in same position by coming to rest of the oscilator , regardless of direction, amplitude, angular velocity, speed, acceleration etc of the oscilator, as long as the bend is within safe boundaries of elasticity of its material, within this limit hairspring displays perfect elasticity for all practical purposes.

 

 

[Watchfixr]

Obviously, there are many ways to skin this cat. I am not sure how much one is better than the other. My approach to this would be:

1. Remove the pallet fork.

2. Adjust the hairspring collet until the impulse jewel is visually in line with the escape wheel jewel and the pallet jewel. At this point the balance arms should be 90 degrees away from the line. In other words the roller table should be positioned so the impulse jewel is at a symmetrical 90 point from the arms. note the relationship with the balance arms and balance cock.

3. Install the pallet fork and visually check to see if the banking pins are equidistant from the fork with the impulse jewel in alignment with the jewels. If not, slowly rotate the balance back and forth and note the position of the fork when the impulse jewel clears the fork. Check the position of the pins at that point. Adjust as necessary to make them equidistant from the fork.

4. This should be a good starting point to check the beat error, amplitude and drop.

5. Fully wound and running, make slight adjustment to the pin positions for best amplitude. I don't open them up any more than is necessary to get the best amplitude.

6. If everything is working well, you should be able to minimize the beat error with slight adjustments to the pins.

This is based on the assumption that the pallet fork is free to move without any hesitation when the watch is moved through positions to check it falls freely in both directions. Your mileage my vary.

Edited December 16, 2020 by Watchfixr

[jdrichard]

5 hours ago, praezis said:

Action of the hairspring is the same as bending a clamped rod: as far as Hooke's law is valid, you will need the same force to bend the rod up or bend it down.

Characteristics of a hairspring is
D (Direktionsmoment in German) = M / α   (= applied moment / rotated angle)
and/or
D = EI / L  (= material and geometric properties)

D indicates no difference bending a hairspring in one or the other direction (just the sign will change).

Frank

I now understand the science and math you have presented and it makes sense. So with the same force in both directions there should be the same swing in the impulse jewel is centered.

[jdrichard]

1 hour ago, Watchfixr said:

Obviously, there are many ways to skin this cat. I am not sure how much one is better than the other. My approach to this would be:

1. Remove the pallet fork.

2. Adjust the hairspring collet until the impulse jewel is visually in line with the escape wheel jewel and the pallet jewel. At this point the balance arms should be 90 degrees away from the line. In other words the roller table should be positioned so the impulse jewel is at a symmetrical 90 point from the arms. note the relationship with the balance arms and balance cock.

3. Install the pallet fork and visually check to see if the banking pins are equidistant from the fork with the impulse jewel in alignment with the jewels. If not, slowly rotate the balance back and forth and note the position of the fork when the impulse jewel clears the fork. Check the position of the pins at that point. Adjust as necessary to make them equidistant from the fork.

4. This should be a good starting point to check the beat error, amplitude and drop.

5. Fully wound and running, make slight adjustment to the pin positions for best amplitude. I don't open them up any more than is necessary to get the best amplitude.

6. If everything is working well, you should be able to minimize the beat error with slight adjustments to the pins.

This is based on the assumption that the pallet fork is free to move without any hesitation when the watch is moved through positions to check it falls freely in both directions. Your mileage my vary.

Good advice to remove the pallet fork for the alignment, else you cannot see the resting position.  I need to invent a tool that transfers the angle error from a non centered impulse jewel directly to the stud position needed for perfect beat error.

[jdrichard]

8 hours ago, JohnR725 said:

I was having strange probably unhelpful thoughts regarding Springs. I was thinking about the hairspring kinda like a mainspring we could look at the power curve of one of those but. Then of course the problem is we don't wind the spring up very much so mainspring power curves would be out.

I was wondering if you take the fork out and wind the balance wheel exactly one turn in one direction and let go and wait until it comes to a stop then wind that the other direction whether that would get us anything at all?

It's actually amazing how well pocket watches run with thing is not right. Banking pins are only for horn clearance.

Then I'm currently being amused looking for something on adjusting banking pins. Mainly because all the references dear friend in? In other words the reference material in finding on adjusting banking pins all seems be slightly different and the order in which they're doing their adjustments is different.

I was thinking maybe we need one of these I wonder how hard it would be to clone one of these? Interesting machine that probably has a price we do not want to go into. You can measure lift angle but you really need it is still measured the amplitude whether it has lift angle or not.

https://www.witschi.com/en/group-of-devices/measurement-of-mechanical-watches/WisioScope S/WisioScope S.html

 

 

20200604_Witschi_WisioScope_Data Sheet.pdf 728.24 kB · 1 download

The pocket watch in question is a 23 Jewel PW and, with a 7ms beat error, it has been keeping perfect time for 3 days. So do i adjust it and risk introducing another issue.

[HectorLooi]

Another question that has been troubling me is, if a watch has been set in beat previously how does it go out of beat, assuming that nobody fiddled with the collet or dropped the watch?

[Nucejoe]

1 hour ago, HectorLooi said:

assuming that nobody fiddled with the collet or dropped the watch?

Any regulation with an imperfect terminal curve, alters the position of impulse pin.

Anytime you touch the hairspring coil to center, even level it, etc,  may practically alter the position that impulse pin comes to rest at.

 

 

[JohnR725]

7 hours ago, HectorLooi said:

Another question that has been troubling me is, if a watch has been set in beat previously how does it go out of beat, assuming that nobody fiddled with the collet or dropped the watch?

Theoretically what you're saying should be correct if the factory adjusted their watch correctly in the first place.

But in this particular example were talking about a American pocket watch.

I will give you an example of my fun yesterday at work. A nice Hamilton hopefully in the final phases of getting it to keep time. I was looking at the balance wheel and was horrified that it look like my upper jewel is now broken? This is because it was way way too much end shake and the upper pivot appeared to have way too much site play like maybe it was missing? Take it apart look at the balance it's perfect. Look at the hole jewel it visually looks fine push on it with my tweezers it wiggles around. The two screws holding on the capstone nice and tight so there was no problems there. Disassembling the hole jewel in the brass setting visually looked fine but had way too much side to play. Measuring it's diameter it did not correspond to what it's supposed to be. Fortunately we had replacement jewels replacing  it things look much better in timekeeping improved. Then whoever had put the jewel setting in had whacked it a couple of times with a staking  set punch in an attempt to spread it to work and it probably would work for most people who weren't paying attention.

Moral of the above story with an American pocket watch anything that can be changed moved played with conceivably has been unfortunately.

Then even if things are reasonably good as a reminder our modern timing machines with  liquid crystal displays nifty digital numbers of exact milliseconds beat didn't exist when these watches came into existence. Didn't even exist for most of their life. Really up until relatively recently the timing machine of choice spit out paper and how close the lines were together how closely you were in beat. It's only really now with our nifty digital machines do we get really obsessed with all of the stuff. 

As this discussion is really how to adjust the escapement I thought I would attached a hand that was some light reading.

 

Escapement handout wostep nscc.pdf

[jdrichard]

7 hours ago, Nucejoe said:

Any regulation with an imperfect terminal curve, alters the position of impulse pin.

Anytime you touch the hairspring coil to center, even level it, etc,  may practically alter the position that impulse pin comes to rest at.

 

 

Totally correct.  I know that there are many adjustments you can do on the hairspring to speed up or slow down a watch, and that would move the position of the impulse jewel.  I am going the take off the plate holding the 4th wheel and Escapement in place and have a look at where the impulse pin is at rest.  going to try to do this without removing the pallet fork as I think it is not touching the impulse jewel whet they are perfectly centered.

[jdrichard]

On 12/16/2020 at 7:56 AM, JohnR725 said:

I was having strange probably unhelpful thoughts regarding Springs. I was thinking about the hairspring kinda like a mainspring we could look at the power curve of one of those but. Then of course the problem is we don't wind the spring up very much so mainspring power curves would be out.

I was wondering if you take the fork out and wind the balance wheel exactly one turn in one direction and let go and wait until it comes to a stop then wind that the other direction whether that would get us anything at all?

It's actually amazing how well pocket watches run with thing is not right. Banking pins are only for horn clearance.

Then I'm currently being amused looking for something on adjusting banking pins. Mainly because all the references dear friend in? In other words the reference material in finding on adjusting banking pins all seems be slightly different and the order in which they're doing their adjustments is different.

I was thinking maybe we need one of these I wonder how hard it would be to clone one of these? Interesting machine that probably has a price we do not want to go into. You can measure lift angle but you really need it is still measured the amplitude whether it has lift angle or not.

https://www.witschi.com/en/group-of-devices/measurement-of-mechanical-watches/WisioScope S/WisioScope S.html

 

 

20200604_Witschi_WisioScope_Data Sheet.pdf 728.24 kB · 1 download

Excellent Document