Centering the impulse stone didn't fix beat error

[GPrideaux]

The Russian 2602.1 I'm working on had a beat error of about 4ms (plus or minus about 0.5 depending on position) before I serviced it. After cleaning that hadn't changed, so I took out the pallet to examine where the impulse stone was to see which way to adjust the collet. I was surprised to see it sitting very nearly central in line with the escape and balance jewels. However, it was ever so slightly to the right of centre so I tweaked it until it sat dead centre in line with the jewels. Feeling pretty pleased with myself I put the pallet back in and reinstalled the balance, only to find that the beat error was now worse! Even without the timegrapher on it I can see that the swing in one direction is taking longer than the swing in the opposite direction. So, what else could the cause be? Please educate me!

The balance spring is not as sitting as flat as I'd like and it is running fast, especially dial down, but I'm confident that no coils are touching each other nor is anything touching the balance or the stud or the cock. The overcoil is between the regulator pins and nothing else is touching the pins.

I'm thinking there might be an issue with the pallet. I think the lock depth is greater than it should be. If this is true, and especially if it's more true for one stone than the other, could that affect beat error?

I'll take some pics tomorrow. It's too close to my bed time now.

[nevenbekriev]

Yes it is possible that wrong escapement depth  to be reason for fast working and shortening of the swing in one of directions. This is when there is no locking (here it is known as drop lock) in one of directions of pallet fork movement. For sure it is situation with lo amplitude and no timegrapher is needed to see what is hapening. Another possibility is hairspring shortening (touching) when swinging balance to one direction, this is problem of the balance/hairspring as resonator. Thou it seems that the hairspring doesn't touch in static, but actually when the balance is rotated, then it begins to touch. Another possibility is the balance arm (one of it's halves) to touch something and this brings hard limit of the balance movement in one direction. Such thing can be the head of the pallet fork cock screw if it sits higher than cock surface (it can happen when wrong screw is used) or similar mechanical reason.

It will be best if You make video of the movement working in dial down position and may be of the free oscillations test.

 

 

[GPrideaux]

It seems I was too confident that I had centred the impulse stone correctly. When I came to take some photos this morning it was not centred at all.

I'll improve that this morning.

I attempted to take some videos of the balance spring in action. It's tricky to get it adequately lit and I'm waving a torch around to try to make things visible. (There are some rolling shutter artifacts that make the impulse stone look crooked but that's all they are; it's not really wobbling around as the balance swings.)

This is what the lock looks like on the pallet stones.

Entry stone looks ok to me, I think:

(The stone has a damaged trailing edge but I don't think that's going to have much effect on anything.)

Exit stone lock looks excessive, I think:

I've taken the pallet out to improve the positioning of the impulse stone, and figured I should do a free oscillation test while it was out. It's a fail. Starting with a 180 degree displacement the balance only swings about 40 times before coming to a standstill, so I need to do some more work on the balance to get that sorted. I should probably start with getting the spring sitting a bit more level. I didn't notice it being out of level when I had it out and disconnected from the cock, so I think I need to start by tweaking the stud positioning.

 

[praezis]

4ms is not very much and it is easy to adjust in the wrong direction or adjust too much, beyond zero. Probably what you did.

NB. your hairspring is tilted severely, I would be surprised if it didn‘t rub on the balance arms.

And the staff suffered from a poor replacement, now wobbling severely, too.

Frank

[nevenbekriev]

Hm, the chipped entry stone might be a problem.

See, evaluation of the escapement adjustment correctness by position of tooth edge against the edge of rest surface of pallet in banking position is not correct. This depends much on correct position of banking pins, and they are easy to bend.

Wat You really need to check is if there is correct locking (the drop lock exists and is reliable) on every tooth of the wheel an on both pallet stones. When tooth moving on the chipped pallet impulse surface, it will leave it sooner as the surface is shorter. Thus the next tooth can drop not on the rest surface, but directly on impulse surface of exit pallet.  Well, when this happens, the sound is different and I didn't hear such sound in the videos, but for sure even if the locking is still present, some amplitude is lost due to the chipped pallet.

You have to better adjust the balance position by collet rotating, but deal with the hairspring position adjustment first. Every change of the hairspring position will change the 'beat error', so the beat error adjustment is the last adjustment to do on the hairspring

 

[GPrideaux]

So, attending to the hairspring. I tried to tweak it a little with the balance in the movement, but didn't achieve much so I've removed the balance again and taken some pics of the side and top view of the spring on the balance.

From the side:

That's not perfect ... still a little conical, but otherwise pretty parallel with the balance wheel. The stud, however, does not appear to be perpendicular to the spring, and I think that's what's throwing it out when it's attached to the cock. The out of square stud is elevating one side and consequently depressing the other. Not sure how to address this ... should I loosen the pin, straighten, then tighten the pin again?

From above:

Part of the overcoil doesn't look quite right to me. Where it interacts with the regulator pins should be concentric with the pivot but isn't, quite. Coming from the inside it crosses to mid-way between the third and fourth coil as it approaches the arm on the right, then its radius increases until it is outside the third coil , then comes in again and where it meets the stud it lines up with the third coil.

[GPrideaux]

Update: I think I have the balance oscillating freely at last.

It's not always one thing.

I adjusted the angle between spring and stud without unpinning the spring and did minor adjustments to the curve of the overcoil then put it back in the watch. It looked better, but didn't swing any better. I discovered the angled section of the spring where it transitions from overcoil to the flat coil was sitting too high and rubbing on the cock. After half an hour or so of patiently tweaking the spring and I was happy with how it was sitting. But it still failed the free oscillation test . I'm just about to leave it for a bit when I notice a tiny black fibre on the balance. I go to lift it out and find it is actually anchored at one end under the barrel bridge and I have to give it a decent tug with the tweezers to get it out. But such a tiny fibre, it's not even visible without the microscope. How much difference could it make? How about the difference between 40-ish oscillations and almost 140 oscillations from a 180º start!

I need to make sure it's still good with the pallet bridge back in place, then I can get back to aligning the impulse jewel and hopefully getting it running with decent beat error and a modicum of regularity.

[nickelsilver]

1 hour ago, GPrideaux said:

 I'm just about to leave it for a bit when I notice a tiny black fibre on the balance. I go to lift it out and find it is actually anchored at one end under the barrel bridge and I have to give it a decent tug with the tweezers to get it out. But such a tiny fibre, it's not even visible without the microscope. How much difference could it make? How about the difference between 40-ish oscillations and almost 140 oscillations from a 180º start!

 

I had the exact same thing happen a few years ago, lost about a day trying to figure it out. Almost invisible little fiber sticking out from under the barrel bridge.

[GPrideaux]

After confirming that the free oscillation test was still good with the pallet bridge in place I readjusted the position of the impulse jewel until it was pretty nearly exactly in line with the pivots then put the pallet in and reinstalled the balance. I'm getting over 300º amplitude in horizontal positions, but still getting around 2.5 ms beat error. Inspecting visually, I think the end of the overcoil, where it bends down to join the rest of the spring, is still brushing the underside of the balance cock so that I have a slight asymmetry between the anticlockwise swing (which moves this away from the cock) and the clockwise swing (which moves it towards and under the cock). It isn't apparent on a free oscillation test because the overcoil only goes under the cock when the amplitude is large enough so starting with 180º results in it approaching the cock for only the first few oscillations. On the TG-TIMER software it also looks like it has perhaps 10 or 15 degrees more amplitude in one direction than the other which is consistent with the hairspring touching the cock in only one direction.

I think I'll leave it for a few days, probably until next weekend, then I'll give the hairspring another tiny twist to try to get the overcoil sitting a few hundredths of a mm lower so it properly clears the cock. Which will, of course, necessitate re-checking that the impulse jewel is still central, and possibly tweaking the collet again.

So it goes.

[GPrideaux]

I need some more advice with this balance and its spring.

I think I've been looking at the wrong part of the hairspring in this movement. I've been focusing on the overcoil, including where it joins the stud, but the spring is not parallel to the balance at all, suggesting an issue with the other end. This pic shows how one side (the RHS in the pic) of the hairspring is higher than the other. The coil sits nicely in a single plane all the way in to the collet but that plane is not correctly parallel to the balance and perpendicular to the balance staff.

Next steps?

The issue seems to go all the way in to the collet itself, with the lowest point about 180º from the point where the spring joins the collet. Should I try to adjust it there (if so, I'm not sure how), or go one full turn out and give it a twist to bring the low side up, or perhaps only go half a turn out and give a twist to bring the high side down?

[nevenbekriev]

OK, You must twist the spring between the points marked with red circles in order to lift the lower part of the overcoil up. The part to be twisted has to be diametrical to the lowest point.

Now about the internal end. First You must observe the spring position when the balance staff is horizontal, as the stud is heavy and bends the spring. If I understand the matter of the issue correctly, You just have to lift a little the spring close to the place where it is pinned and this will get it to correct position, but first observe and try to figure if this is correct and logical as I can't see on the picture the position of the point where the spring is pinned.

[GPrideaux]

Thanks, @nevenbekriev. That photo is with the balance staff horizontal: it's looking vertically down on the wheel standing on its rim on some Rodico.

Here's the pic from above I posted before with a line to show roughly where the tilt lies, with the low point below the stud. I've circled the point where I think I need to adjust, but it might be possible to go 180º before that, near the slot in the collet.

 

 

[nevenbekriev]

I guess the reason for this bent is the hairspring end was pressed down right in front of the collet slit. This of course happened when attempting to turn the collet with screw driver for adjusting the beat error. Of course, the spring now can be  made parallel by twisting where the red circle is, but this will position it a little higher than it used to be originally. Twisting in diametrical point will position it a little lower. Of course, this is not problem at all as such shifting is too small to make any difference. But, the most logical and correct way to bring the spring to parallel is to apply some lifting of the beginning  curve right in front of the slit, just in reverse to the way it was bent down.

Edited February 4 by nevenbekriev