Weird Behaviour on Timegrapher (rebuilt ETA 2390)

[Pauly]

Hi, I recently stripped/cleaned/rebuilt an old ETA 2390 movement for practice, (not the one in my '1st completed practice' post... another one from the same ebay lot!) it wasn't running at all when I began on it, and after the rebuild it is running. But the timegrapher results are weird... I will strip and rebuild again to investigate, but any ideas as to where to look for the cause for this would be appreciated!

Basically over the course of a minute or so, the timing varies dramatically, but almost in a predictable way (it is predictably erratic if you will), it runs 5-10s slow for about 30s (but never really 'settles' it is always tracing a curved plot), then races away to around 3 minutes fast for 20s before returning back to 5-10s slow. It was doing this on a full wind, in all positions, and is still doing it around 18 hours later (so towards half wound down). Over the course of the behaviour, amplitude and beat error barely change (around 240, 1.5ms respectively). All the screenshots are of dial up, in the other positions amplitude is down to around 180-190, but the slow/race behaviour is the same.

There is nothing obvious different visually (to me at least, which may not be saying much!) when it is running fast vs slow, i.e. no noticeable skipping of train wheels or anything.

I have tried demagnetising it, but in any case not sure magnetisation would cause this repeatable slow/fast/slow/fast behaviour?

I was thinking about bent staffs/pivots cracked jewels etc, but wouldn't they all generally slow the watch down?

Because it seems to repeat at set intervals I was thinking something in the train or motion work, but again, can't think what would cause it to be super fast (as opposed to slow?)

[transporter]

crikey mate that does look predictable doesn't it, i think you are on the right track thinking train, is there a bent pivot somewhere that is causing drag, maybe even a high spot on a wheel that is causing depthing issues, im sure some one with far more knowledge that me will be along soon to help you out even more

[Pauly]

20 minutes ago, transporter said:

crikey mate that does look predictable doesn't it, i think you are on the right track thinking train, is there a bent pivot somewhere that is causing drag, maybe even a high spot on a wheel that is causing depthing issues, im sure some one with far more knowledge that me will be along soon to help you out even more

Thanks for the reply! I'm kind of hoping I'll find something obvious if I strip it again, what is 'depthing'? (pardon my ignorance, I've only been playing with watches for about a month)

[Nucejoe]

I check the center wheel, that is the biggest wheel on its arbour the seconds hand gets mounted, check its arbour, gear teeth, pinion, jewel, for wobble as it may rub, also check the center tube.

Did you remove and check its end stone? its pivot.  How near the bridge is center wheel turning, it can rub under the bridge.

[Pauly]

Yeah, thanks, I just double checked and it is doing it at exactly 1 minute intervals, goes crazy for about 15-20s then backs off during the other 40s, so 4th (seconds) wheel would be a good bet since it is turning once a minute? I'll probably strip it again and take a look tomorrow evening, will report back what (if anything) I find.

[Pauly]

Ok, I stripped, cleaned, rebuilt, behaviour was slightly different but still similar, repeatable time gain/loss on the minute.

So went through it all again and I finally managed to diagnose the issue, and partially fix it... but then made it worse! It wasn't the 4th wheel itself, but the tube it passes through that is part of the main plate, it was slightly out of true, by pushing it in various directions I could remove the erratic running. It then just ran consistently fast and with higher amplitude... however, I should have left it when it was 'good enough' rather than trying to get it 'perfect'. In pushing it back and forth I loosened it (it is a push fit in the plate) sufficiently that now it barely runs at all  I have a spare main plate from another 2390 from the same ebay lot, so will strip everything again and rebuild on to that.

It seemed incredibly sensitive to the alignment of this, I guess at some point it got damaged while fitting hands or something?

A good learning experience anyway, another thing to bear in mind for future projects to keep an eye on.

[JohnR725]

On 6/8/2020 at 9:00 AM, Pauly said:

It was doing this on a full wind, in all positions, and is still doing it around 18 hours later (so towards half wound down). Over the course of the behaviour, amplitude and beat error barely change (around 240, 1.5ms respectively). All the screenshots are of dial up, in the other positions amplitude is down to around 180-190, but the slow/race behaviour is the same

The drop down to 180° is unacceptable. When you're timing you really should actually do it in more than one position dial up or dial down at the minimum may be like crown down. Because when you're resting on the end of the pivots in one of the dial positions you getting the least friction the best results. Which is why always putting in a crown position to see how good or bad things are is really good.

Is there a reason why you haven't put the watch in beat  yet?

Then the center pipe as I think it's called that you were playing with appears to be a replaceable part?

http://cgi.julesborel.com/cgi-bin/matcgi2?ref=q[cJ

[rodabod]

I doubt it was still knocking out 245 degrees when it started to run fast. The timegrapher probably didn’t manage to register a stable reading to update it. Sometimes you have to look at the balance wheel itself to gauge amplitude. 

[rodabod]

Just to add - you must be a natural at fault diagnosis if you’ve only recently started. That’s potentially quite a tricky and confusing fault, so well done. I have many 2390 spares as it’s the base caliber for a type of Hamilton watch which I occasionally restore. Let me know if you need any spare parts. 

[Pauly]

5 hours ago, rodabod said:

Just to add - you must be a natural at fault diagnosis if you’ve only recently started. That’s potentially quite a tricky and confusing fault, so well done. I have many 2390 spares as it’s the base caliber for a type of Hamilton watch which I occasionally restore. Let me know if you need any spare parts. 

Hi, thanks, maybe, more like dumb luck though I think! Unfortunately my luck ran out, latest rebuild has made things much worse, think I've damaged the balance or escapement/pallets now somehow as it is all over the place, running 800+s fast and beat error of 6-7ms, re: your earlier point of amplitude drop, you may be right, but I couldn't see anything at all obvious different on the balance during the 'fast' 20s (not to say it wasn't dropping, I wouldn't necessarily see it as I'm new to this).

Thanks for the parts offer, but no need on this occasion, I have 4 of these 2390s from a job lot bought from ebay (all orphaned, no case movements, none of which were running), they have no value to me, just using for beginners practice. I've switched a couple of parts back and forth between them, but the 1st one I rebuilt (superficially the one in worst cosmetic condition) is running the best of the 3 I've got going. I'll put that one in the case that was in the lot too, and I will continue playing with this one to see if I can work out what I've done to it!

[Pauly]

On 6/12/2020 at 10:54 AM, JohnR725 said:

 

Is there a reason why you haven't put the watch in beat  yet?

 

I haven't quite got the confidence (or knowledge of exactly what to do, e.g. which way or how far to turn it) to adjust the hairspring collet on a fixed stud balance yet, I've only started this as a hobby in the last month, and this is only the 3rd or 4th 'practice' movement I've stripped/rebuilt. Also, as I'm a bit inexperienced/clumsy, the beat error on the movement seems to change every time I remove/replace the balance (I only seem to get the impulse jewel on the correct side of the pallet about 1 time in 3!) so I don't know what to use as a baseline.

I did have a look at this, I've seen other posts where the advice is to set the impulse jewel directly between the banking pins with no power on the mainspring, but try as I might I can't see where the impulse jewel is located once the balance is installed, it is obscured by the pivot and I can't find an angle where I can see it.

Edited June 14, 2020 by Pauly

[JohnR725]

2 hours ago, Pauly said:

I haven't quite got the confidence (or knowledge of exactly what to do, e.g. which way or how far to turn it) to adjust the hairspring collet on a fixed stud balance yet,

we need a really good picture of the top of the balance bridge assembly and your watch may have a floating stud which is easy to adjust which is why I might've pointed this out or asked?

8 hours ago, rodabod said:

I doubt it was still knocking out 245 degrees when it started to run fast. The timegrapher probably didn’t manage to register a stable reading to update it. Sometimes you have to look at the balance wheel itself to gauge amplitude. 

always good with a timing machine to look at three separate things they should all look the same. Visually the balance wheel motion should agree with whatever the timing machine says it is. The graphical display and the numeric display should agree. Unfortunately it's not like you get the pick whichever one looks the best and that one's going to be right. Sometimes the graphical display will look like garbage you'll actually see numbers and the numbers mean nothing. Or visually the balance wheel is barely moving and once again the timing machine gives you happy nice numbers then you have a problem.

2 hours ago, Pauly said:

Hi, thanks, maybe, more like dumb luck though I think! Unfortunately my luck ran out, latest rebuild has made things much worse, think I've damaged the balance or escapement/pallets now somehow as it is all over the place, running 800+s fast and beat error of 6-7ms, re: your earlier point of amplitude drop, you may be right, but I couldn't see anything at all obvious different on the balance during the 'fast' 20s (not to say it wasn't dropping, I wouldn't necessarily see it as I'm new to this).

one of the thing is really really hard to see for all of us from time to time especially so with beginners is issues. Hairspring issues are the most common place touching the balance wheel touching the balance bridge things that aren't quite right. just really really hard to see.

it would be nice if you could get me a picture of your movement you're working on I need to see a couple of things so I can see if I see what you're seeing. Then I'll explain a little about the beat whether we want to change it or not.

when you're putting the balance wheel back in the watch if you want to make sure the roller jewel's always in the right place to engage the fork usually what helps is there should be a little power on the watch you locate the pallet fork C no word is typically I haven't pushed towards the inside of the watch. You bring the balance wheel in with the roller jewel greatly exaggerated away from the fork not sure if the wording makes sense. Bring it in set it in place rotate the balance bridge because the balance is usually hanging under the bridge the balance bridge in place in the watch should start off. So basically you're trying to get the roller jewel on the same side as the fork and when it rotates back to that position it will engage the fork if you try to engage the fork your chance of getting the roller jewel where it's supposed to be is problematic. Hopefully this makes sense

 

 

 

[Pauly]

Picture of the balance, just looking at it again now, the hairspring is distorted and touching underneath the cock on the left of the picture, I should have said in earlier comment, on the excessive rate and beat error, amplitude is very low now too (150), stud is fixed on this balance. So this one may be done for :(.

I know the theory of the approach to installing the balance, but I really do struggle! The most difficult bit for me is locating the lower pivot initially, either I can't get it to locate, or the balance pivot hits the pallet fork and causes it to flick back the other way If I do get it in, on this movement you can't go in too far off the angle as the regulator lever would foul the center wheel as you rotate ther cock around. Often as I rotate the cock/bridge around the lower pivot often jumps out again (so going to the wrong side of the pallet again), or I do everything what seems correctly, but struggle to locate the upper pivot and when I finally do I find the impulse pin is 'wrong side' again... I just need practice! 

 

[Nucejoe]

The tube seconds arbour goes through is call center tube/pipe and is replacable, the hole  into which center tube is fitted ( center hole) may wear out. Routinely examine this tube in all ETA, hold the tube in tweezers and shake, it should stand firm in place and perpendicular to the plate , on the other side  bottom of the tube should flash level with  mainplate. Run a piece of thread inside the tube and clean the hell out of it. Replace the tube and/or bush the mainplate as needed. 

In case of fixed stud holders, I turn the roller table to put impulse in beat, that is a great deal less work than the alternative.

 

[JohnR725]

2 hours ago, Nucejoe said:

In case of fixed stud holders, I turn the roller table to put impulse in beat, that is a great deal less work than the alternative.

even though this is what he does it's not the preferred method. If you are into ultra precision timekeeping it screws up to timekeeping.

it appears to be a three armed balance wheel? If you look at the bottom side line the roller jewel visually out to the edge doesn't correspond to the arm? Because sometimes you can use that third arm and placing it directly over the pallet fork pivot that's where it should be when it's in beat.

once you've verified where it should be when it's in beat in your holding the balance wheel there you can site from the stud to the rim of the balance wheel and put up tiny mark where the stud should be when it's out of the watch. On some older watches the factories will actually mark the balance wheel so you know where the stud should be.

then ideally for rotating the collet there's a special tool. Basically it's a really thin blade that slides into the slot with out spreading the slot. If you shove a screwdriver in their you risk spreading the collet and that requires another tool close.

On the other hand I wouldn't worry about that right now your pictures really interesting notice when you look at the hairspring you go in a couple of coils and why does it look really really dark?

[Pauly]

27 minutes ago, JohnR725 said:

 

On the other hand I wouldn't worry about that right now your pictures really interesting notice when you look at the hairspring you go in a couple of coils and why does it look really really dark?

Think it is just the angle/lighting on the blue tinged spring, combined with a slight gap in the coils at the point in the swing when the picture was taken (balance running when picture taken) this is it from a different angle, looks more normal? I need a better camera, my phone does a reasonable job but a macro lens would be easier!

Thanks for the tip of marking the pallet position on the balance and using that as a reference when turned over, hadn't thought of doing it that way!