Best way to adjust positional variation in etachron systems

[nickelsilver]

  On 4/23/2022 at 10:53 AM, ifibrin said:

Would it be correct to say I should turn the regulator stud in the direction that the inner prong of the stud (the prong on the inner radius, less distance to the pivot hole) so that the inner prong will be closer to the hairspring stud? That is the impression I get from the awci article.

Expand  

Yes!

[ifibrin]

  On 4/23/2022 at 10:58 AM, nickelsilver said:

Yes!

Expand  

Ok, I will try rotating the regulator stud now. Will post the results in awhile.

[ifibrin]

  On 4/23/2022 at 10:58 AM, nickelsilver said:

Yes!

Expand  

Unfortunately, it appears the delta between CL and CR is here to stay. For some reason, when I rotate the regulator stud, the inner prong contacts the hairspring, but there is still a gap between the hairspring and the outer prong. Further rotation of the regulator stud doesn’t help. Interestingly, if I rotate the regulator stud even more so that the hairspring doesn’t bounce between the prongs of the regulator stud, the amplitude isn’t affected, but when I push the regulator arm closer to the hairspring stud, the amplitude suddenly drops from 290 to 180 suddenly at a certain point. I suspect there’s  a slight twist in the terminal curve somewhere, since this isn’t supposed to happen...

[Klassiker]

  On 4/23/2022 at 12:52 PM, ifibrin said:

For some reason, when I rotate the regulator stud, the inner prong contacts the hairspring, but there is still a gap between the hairspring and the outer prong

Expand  

This doesn't make any sense to me, as the gap must get smaller (or larger) as you rotate the block. It cannot stay the same! The last photo you took was pretty good. Can you take a series of pictures like that one as you gradually rotate the block?

  On 4/23/2022 at 12:52 PM, ifibrin said:

Interestingly, if I rotate the regulator stud even more so that the hairspring doesn’t bounce between the prongs of the regulator stud, the amplitude isn’t affected, but when I push the regulator arm closer to the hairspring stud, the amplitude suddenly drops from 290 to 180 suddenly at a certain point.

Expand  

I think you have rotated the block too far at this point, and trapped the hairspring between the pins. Then, when you move the regulator arm you are pulling the hairspring out of shape.

[ifibrin]

  On 4/23/2022 at 4:24 PM, Klassiker said:

This doesn't make any sense to me, as the gap must get smaller (or larger) as you rotate the block. It cannot stay the same! The last photo you took was pretty good. Can you take a series of pictures like that one as you gradually rotate the block?

Expand  

The gap does get smaller as I rotate the regulator stud, it’s just that the inner prong touches the hairspring before the outer prong does. In other words, the two prongs of the regulator don’t close in evenly on the hairspring as you rotate it, such that the hairspring leans on the inner prong while the outer prong doesn’t touch the hairspring. I’ve tried taking pictures of the regulator stud but I can’t get it to focus on the two prongs (the depth) to show what I mean.

Edited April 23, 2022 by ifibrin

[Klassiker]

OK, I can imagine what you mean now. That suggests the hairspring wasn't quite centred between the pins in the first place. It's a very tricky process of trial and error getting it perfect. Close the gap slightly by turning the block, then very slightly turn the stud to move the hairspring a fraction away from the pin, so that is not in contact with either of them. Get the gap as small as you can without pinching the spring, then check the timing variations again.

[Nucejoe]

You might have moved the regulator stud or its just the gap, thats if you are sure end curve is OK, Should fix the fault , not correct it by creating another fault.

[ifibrin]

The end curve looks fine, and follows the regulator arm, at 10x magnification. I think I will hold off any further hairspring work until I get s microscope so that I can be sure what I am doing. Thanks for all the help! I will consider it before I proceed any further.

 

  On 4/23/2022 at 5:09 PM, Klassiker said:

Close the gap slightly by turning the block, then very slightly turn the stud to move the hairspring a fraction away from the pin, so that is not in contact with either of them.

Expand  

You mean the hairspring stud right?

[Klassiker]

I am using stud to mean the (silver-coloured) part where the end of the hairspring is anchored, and block to mean the (brass-coloured) part containing the regulator pins.

I absolutely understand your decision to wait until you have better optics. It is crucial that you can see what you are doing. Even if you decide to leave the watch as it is, it is in a perfectly serviceable condition.

Edited April 23, 2022 by Klassiker

[JohnR725]

Something bothering me from your pictures? The first picture looks right but the second picture doesn't look right? So I snipped out the problem it looks to me like your regulator is up a little bit versus the original picture where it's down where it's supposed to be?

[ifibrin]

  On 4/23/2022 at 7:26 PM, JohnR725 said:

Something bothering me from your pictures? The first picture looks right but the second picture doesn't look right?

Expand  

I’ve looked at the regulator stud, and it’s flush on top of the regulator arm. I think it’s due to the shadow that’s why it looks above the arm.

After repeated attempts to adjust the hairspring stud, and the regulator stud, I realized something odd.

I now think there’s a slight twist in the hairspring somewhere, for the following reasons:

in the horizontal positions (both dial up and dial down), the balance wheel shows a slight wobble. I think the slight twist causes the balance wheel to be pulled up and down against gravity as the hairspring contracts and expands, against the (necessary) endshake as the balance oscillates, causing a slight wobble. You can see in the video that the balance staff is wobbling/jumping between the horizontal positions.

https://drive.google.com/file/d/1kw_UPEEZU7Ihsq9AUfiXB1a_3JfkwdEA/view?usp=sharing

 

In the vertical positions, the balance wheel does not wobble, as the slight twist in the hairspring is not acting against gravity as the balance wheel oscillates.

https://drive.google.com/file/d/1AdK9uGQqLfqxIqYA-OrVvOSqcPIneJ6i/view?usp=sharing

 

Edited April 24, 2022 by ifibrin

[Klassiker]

I wouldn't read too much into that slight wobbling, to be honest. It doesn't have anything to do with the positional variation you are seeing, and doesn't indicate a problem with the hairspring. A flat hairspring, even If it is perfect, will always Impact radial as well as turning forces on the balance, so a slight wobble due to side-shake is common. If the movement were axial as you are proposing, then surely it would also be present in the vertical positions?

[ifibrin]

  On 4/24/2022 at 11:16 AM, Klassiker said:

I wouldn't read too much into that slight wobbling, to be honest. It doesn't have anything to do with the positional variation you are seeing, and doesn't indicate a problem with the hairspring. A flat hairspring, even If it is perfect, will always Impact radial as well as turning forces on the balance, so a slight wobble due to side-shake is common. If the movement were axial as you are proposing, then surely it would also be present in the vertical positions?

Expand  

Oh... I always thought that the balance wheel should not have any wobble at all, assuming correct values of side shake or end shake.

Anyway, I did try your suggestion of turning the hairspring stud slightly to center the hairspring better, but it didn’t make much difference to the positional variation. Visually at 10x, it appears to me that the hairspring is making contact with the outer prong of the regulator block somewhere in the middle of the hairspring, as opposed to where it should be, at the edge of the hairspring.

Another odd thing is that the beat error also changes slightly if you rotate the hairspring stud too much, even if the hairspring is still centered over the pivot.

[Klassiker]

  On 4/24/2022 at 11:53 AM, ifibrin said:

I did try your suggestion of turning the hairspring stud slightly to center the hairspring better, but it didn’t make much difference to the positional variation.

Expand  

Did you get it centred - i.e. daylight visible between both pins and the hairspring? Did you follow up by reducing the gap between the pins (turning the regulator block)? If you did all that, and still didn't see any improvement in positional variation, then you can cross it off the list.

  On 4/24/2022 at 11:53 AM, ifibrin said:

Visually at 10x, it appears to me that the hairspring is making contact with the outer prong of the regulator block somewhere in the middle of the hairspring, as opposed to where it should be, at the edge of the hairspring.

Expand  

Sorry, I've read this through a few times, and I don't get what you mean. Maybe a photo or sketch would help?

  On 4/24/2022 at 11:53 AM, ifibrin said:

Another odd thing is that the beat error also changes slightly if you rotate the hairspring stud too much, even if the hairspring is still centered over the pivot.

Expand  

You can only check if the hairspring is centred by removing it from the balance, and remounting it alone onto the cock. Then you can see if the collet lies concentric with the jewel hole. If you are turning the stud a lot, you are probably putting the hairspring way off-centre, but the regulator is keeping the terminal curve in place, so you are not noticing it. If you look carefully, you will see the spacing of the coils changing. I'm talking about relatively large rotations of the stud here, not the tiny adjustments you need to make when to centre the spring between the regulator pins.

[ifibrin]

  On 4/24/2022 at 2:08 PM, Klassiker said:

Did you get it centred - i.e. daylight visible between both pins and the hairspring? Did you follow up by reducing the gap between the pins (turning the regulator block)? If you did all that, and still didn't see any improvement in positional variation, then you can cross it off the list.

Expand  

To the best of what I could see at 10x, I adjusted the gap in the regulator block to the following:

  On 4/24/2022 at 2:08 PM, Klassiker said:

You can only check if the hairspring is centred by removing it from the balance, and remounting it alone onto the cock. Then you can see if the collet lies concentric with the jewel hole. If you are turning the stud a lot, you are probably putting the hairspring way off-centre, but the regulator is keeping the terminal curve in place, so you are not noticing it. If you look carefully, you will see the spacing of the coils changing. I'm talking about relatively large rotations of the stud here, not the tiny adjustments you need to make when to centre the spring between the regulator pins.

Expand  

I was thinking that if the hairspring was shifted off center from the pivot hole by rotating the hairspring stud, I would see bunching of the hairspring coils in one direction, and that it wouldn’t be necessary to take the hairspring off the balance wheel to see it... But I suppose that if it was off center only a tiny bit I wouldn’t notice the coils bunching up.

 

 

 

 

[Klassiker]

My understanding is this. There are two very similar versions of this system, one from ETA and one from Seiko. Looking at your photos, your regulator is the ETA version with the two pins or tapered "prongs" either side of the hairspring. It looks like this when correctly adjusted and the balance is at rest:

You've drawn the other version, which is the Seiko version, a large round pin with a slot through it. However, it is relieved in the middle section to effectively form two pins, with a similar cross-section to the ETA version in the area where the hairspring operates. That round "cap" on the end with the slot in it which you have sketched should not contact the spring. When correctly adjusted, it looks like this:

Whatever the case, with the balance at rest there should be no contact between the hairspring and the pins, and the two gaps should be as small and as equal as possible.

[ifibrin]

  On 4/24/2022 at 5:07 PM, Klassiker said:

You've drawn the other version, which is the Seiko version, a large round pin with a slot through it.

Expand  

The regulator block I have is the eta version with the prongs; I only sketched the cross section of the regulator block where the hairspring resides.

The odd thing is even if I open the regulator block gap further, or close it, the difference between CL and CR is always about 50seconds. The absolute values will change, but the difference is always about there.

[ifibrin]

I reran the low amplitude readings, and the results are:

Dial up -40s; 155deg; 0.2ms

Dial down -63s; 150deg; 0.1ms

after adding a bit of wind to get the vertical positions to around 150deg,

Crown up -39s; 148deg; 0.2ms

Crown up-right +28s; 149deg; 0.0ms

Crown right +21s; 150deg; 0.1ms

Crown right-down -22s; 153deg; 0.1ms

Crown down -81s; 152deg; 0.0ms

Crown down-left -154s; 149deg; 0.2ms

Crown left -157s; 150deg; 0.5ms

Crown left-up -123s; 149deg; 0.6ms

The interesting thing is that the delta between CL and CR is reversed at low amplitude, compared to at high amplitude. This does suggest a dynamic poising issue. I did some reading up on dynamic poising, and it appears that hairspring shape and positioning does appear to affect the overall dynamic poise of the balance wheel, among other factors. Meaning that a hairspring not centered properly, or out of flat can give an apparent heavy spot in an otherwise statically poised balance wheel.

Thanks to @Klassiker, @nickelsilver, @JohnR725, @Nucejoe  for the help.

Edited April 26, 2022 by ifibrin

[ifibrin]

I decided to try some dynamic poising, using the results I obtained at 150deg amplitude. However, while I did manage to find the heavy spot by the dynamic poising method, I did not want a permanent solution, so I added a small amount of weight opposite the heavy spot instead, using a small amount of rodico. Astoundingly, the results were impressive, again timing at low amplitude:

Crown up -122s, 149deg, 0.7ms

Crown up-right -121s, 151deg, 0.5ms

Crown right -114s, 155deg, 0.2ms

Crown right-down -127s, 158deg, 0.1ms

Crown down -126s, 152deg, 0.3ms

Crown down-left -121s, 154deg, 0.6ms

Crown left -139s, 151deg, 0.9ms

Crown left-up -130s, 149deg, 0.9ms

However, since I think the balance is out of dynamic poise because of mishandling of the hairspring, I don’t want to make permanent alterations to the balance wheel by removing weight for now.

 

Edited May 2, 2022 by ifibrin

[Klassiker]

Excellent detective work! Do you want to show us the hairspring, or are you going to leave it alone now? (Only for a while, obviously, until that itching in your fingers becomes unbearable).

[ifibrin]

  On 5/3/2022 at 5:26 AM, Klassiker said:

Excellent detective work! Do you want to show us the hairspring, or are you going to leave it alone now? (Only for a while, obviously, until that itching in your fingers becomes unbearable).

Expand  

I’m not confident of removing such a tiny hairspring from a tiny balance wheel without distorting the inner coils of the hairspring currently. The etachron system of attaching the hairspring to the balance cock also makes me a bit nervous, since the only way to check the hairspring is flat and true is to attach the naked hairspring to the balance cock.

As a hypothetical exercise, what is the best way to dynamically poise a balance wheel? Do you remove weight from the balance wheel with the hairspring and balance cock still attached? I imagine taking the balance wheel or hairspring off the balance cock can introduce new dynamic poise issues if you take it on and off, especially in the etachron system. But how else can you safely remove weight from a smooth balance wheel?

 

[nickelsilver]

  On 5/3/2022 at 5:53 AM, ifibrin said:

.

As a hypothetical exercise, what is the best way to dynamically poise a balance wheel? Do you remove weight from the balance wheel with the hairspring and balance cock still attached? I imagine taking the balance wheel or hairspring off the balance cock can introduce new dynamic poise issues if you take it on and off, especially in the etachron system. But how else can you safely remove weight from a smooth balance wheel?

 

Expand  

On a solid rim balance like this, you would lay the balance with cock on the bench upside down, grab the rim with sturdy tweezers and bring the balance edge where you want to remove weight over to sit on a low bench block, then with a fine drill, remove a but from the underside. Having a bench block that is only 5 or 6 millimeters thick helps. Remove less than you think you should- you can't put it back, haha!

[ifibrin]

  On 5/3/2022 at 7:14 AM, nickelsilver said:

On a solid rim balance like this, you would lay the balance with cock on the bench upside down, grab the rim with sturdy tweezers and bring the balance edge where you want to remove weight over to sit on a low bench block, then with a fine drill, remove a but from the underside. Having a bench block that is only 5 or 6 millimeters thick helps. Remove less than you think you should- you can't put it back, haha!

Expand  

Where did you find such a thin bench block? Did you make it yourself?

[NigelTufnel]

I hope you all don't mind a noob jumping in to ask questions.  I have a Seiko 4R36A I am replacing the balance whee and hairspring on and I was looking at the manual, which discusses turning the regulator pins.  I'm curious since I'll barely be able to see it when in, how can I actually tell the correct setting? Is this something easy to tell with a timer? Like, from pins parallel to hairspring, to just touching.  WHat happens?