Bring down amplitude on Unitas 6498

[mbwatch]

I've been servicing a Swiss Unitas 6498 from a 1960's pocket watch. It's been cleaned, and I deep clean train pivots with EVEflex. Its original mainspring was a white alloy of the more concentric older type without an S curve. When assembling it, I decided to try the contemporary mainspring pulled out of a parts Chinese ST36.

With either mainspring, the movement runs dead flat on the timegrapher with at most a 10sec delta. The old mainspring achieves about 275° horizontal but drops way down to about 205°-210° in the vertical positions. A 70 degree drop seems excessive especially since it falls well below 220. There isn't much more I can do to improve the condition of the pivots with my skills and tooling and if I didn't have other options I'd probably leave it this way.

But with the modern ST36 mainspring, at full wind in the horizontal positions it races right up to >330° and starts rebanking. The verticals are healthy 275°-280°. I would like to use the stronger spring to take advantage of its longer power reserve and stronger verticals, but I need to rob it of maybe 10° to stay comfortably away from rebanking.

It's been oiled according to ETA 6497 spec sheets, so hp1300 on the center & third wheels, 9010 on the fourth and escape. I used 9451 on the pallet stones. Where would you recommend I try to impede it a little? On other movements, oiling pallet fork pivots has not made much of a reduction. I was thinking about cleaning off the 9451, but am concerned about wear on the escape wheel teeth. Maybe switch the rest of the train pivots to hp1300 or 9501 grease on the center wheel?

[Klassiker]

The chinese spring may not be correct for the Unitas movement. You could measure the chinese spring, then compare it with the recommended dimensions for the Unitas movement. If they are different, as I suspect, buy the correct spring.

Or you could just be satisfied with the old spring. A 10 second delta across the positions is not bad. What do you see after 24h?

[mbwatch]

6 hours ago, Klassiker said:

Or you could just be satisfied with the old spring. A 10 second delta across the positions is not bad. What do you see after 24h?

24h on the old spring has very little loss. Deltas are still around 10-12 and amplitude only falls to about 265 horizontal, vertical stays above 200. Perfectly acceptable. It does not start to deteriorate much until about 36 hours, where it begins slowing gradually. The last 8 hours of its reserve slow considerably and that's one of the reasons I was interested in trying a modern S curve spring.

 

[nickelsilver]

ETA lists the same spring for the 6497-1 and -2, at a thickness of 0.175mm (GR 4307-1). It's not uncommon though for a maker to use a stronger spring when going to a higher frequency on the same base caliber. The 6497-1 is 18,000bph, the -2 is 21,600. Perhaps the ST36, which is 21,600bph, does use a stronger spring. It doesn't take much to cause rebanking.

 

An old Unitas 6497 that's hitting 275 in horizontal really shouldn't see more than 40 degrees drop in verticals. Could be a pivot issue with the balance, or if the locks on the escapement are a bit heavy that shows up more in the verticals.

 

Curious what the size difference is between the mainsprings.

[mbwatch]

25 minutes ago, nickelsilver said:

An old Unitas 6497 that's hitting 275 in horizontal really shouldn't see more than 40 degrees drop in verticals. Could be a pivot issue with the balance, or if the locks on the escapement are a bit heavy that shows up more in the verticals.

 

Curious what the size difference is between the mainsprings.

Ah, I was not considering the difference in beat rate. I did not fully disassemble the balance so I could clean the pivots in EVEflex so that is a possible place I could improve the verticals on the old spring.

This weekend I'll pull out the ST36 spring again and measure both for thickness. It does calm down enough after about 4-6 hours from a full wind where it settles around 325° and stops rebanking. After that it runs great >50 hours.

[praezis]

Hello,

a stronger mainspring cannot heal a too big difference between flat and vertical amplitude, it will just shift values to a higher range.  Reasons are elsewhere as mentioned by nickelsilver.

And you do not want 275 deg vertical, for you will get differing rates in the vertical positions then - 220 is optimal.

Frank

 

 

 

[mbwatch]

1 hour ago, praezis said:

a stronger mainspring cannot heal a too big difference between flat and vertical amplitude, it will just shift values to a higher range.

I am observing much narrower differences between the springs though. That's why I wanted to explore a solution.

Old: 275h/210v (-65°) on full wind

New: 323h/290v (-33°) after 8hours

1 hour ago, praezis said:

And you do not want 275 deg vertical, for you will get differing rates in the vertical positions then - 220 is optimal.

I have not heard this before. Are you saying that if amplitude is too high it will cause greater variation among all the vertical positions than if amplitude is lower around 220°? Thanks for your input.

[praezis]

55 minutes ago, mbwatch said:

Are you saying that if amplitude is too high it will cause greater variation among all the vertical positions than if amplitude is lower around 220°?

Yes, if there is a poise error - and there always is.

We had discussed this in several threads about dynamic poising or just poising.

Frank

[nevenbekriev]

5 hours ago, praezis said:

a stronger mainspring cannot heal a too big difference between flat and vertical amplitude, it will just shift values to a higher range.

Something to say here. The losses in balance system increase proportionally to the square of the amplitude Thus, there is effect of limitation of the amplitude due huge increase of losses, while the income of energy is constant, and this leads to smaller differences vertical/horizontal when bigger amplitudes

[JohnR725]

On 9/25/2024 at 7:53 PM, mbwatch said:

I would like to use the stronger spring to take advantage of its longer power reserve

The running time of the watches by the design of the watch with the length of the mainspring being very important. So in other ways we can get you a really powerful mainspring that's really short and you're going to have no run time at all.

On 9/25/2024 at 7:53 PM, mbwatch said:

I decided to try the contemporary mainspring pulled out of a parts Chinese ST36

This is a interesting solution because?

On 9/26/2024 at 5:30 AM, nickelsilver said:

ETA lists the same spring for the 6497-1 and -2, at a thickness of 0.175mm (GR 4307-1). It's not uncommon though for a maker to use a stronger spring when going to a higher frequency on the same base caliber. The 6497-1 is 18,000bph, the -2 is 21,600. Perhaps the ST36, which is 21,600bph, does use a stronger spring. It doesn't take much to cause rebanking.

I would be really curious as to where you found the listing for ETA and the mainspring. The problem is for the 6498 it's been around forever and specifications of the mainspring are readily available. But the – two version does not come up in the typical parts listing that I look at. But let's look at the technical documentation. In the technical communication for the 6498 –1 We get a minimum running time of 46 hours with atypical of 53 hours. With the problem of the modern tech sheet not even giving a part number other than a note of variations? It's the problem of the modern documentation that even if we did get a part number it's typically going to be for the barrel complete I need to find a much older documents. But let's see what happens with the newer version for the running time they just say 53 hours.

For better idea currently looking at bestfit online and I know we discussed the mainspring before and I know they're supposed to be different. So according to the listing the mainspring barrels themselves have differing part numbers and the barrel arbor is different. Five over going to get as is what typically happens which is why there's an assumption that they're both the same when I'm surely not all we do get is a mainspring listing for the original one which is this  MSPG 6X5X18.5 48-16L.

Now for nitpicky manufacturing information for the 1 version is slightly different than the second version but not much.

- 2 version

One last thing to look at which is somebody has a list of mainspring winder sizes versus watch calibers. The tool diameters the same at 14.6 mm but the winders themselves have a minor size difference I wonder if perhaps there's a slight difference in length of the Springs?The -1 4.5 mm versus the -2 at 4.0 mm.

 

On 9/26/2024 at 6:05 AM, mbwatch said:

I did not fully disassemble the balance so I could clean the pivots in EVEflex so that is a possible place I could improve the verticals on the old spring.

Sometimes I wonder about these modern newfangled cleaning methods that How many of the rest of us are using this? One of the reasons I bring this up is I saw a nice article on refinishing clock pivots where they use the electron microscope. Where sometimes things that are bright shiny in reality are not actually good pivots they only look bright and shiny but looking at the website here of providing you get the right grit it does indicate you can polish but the wrong Grit you could start doing damaged your pivots

https://eternaltools.com/products/eveflex-mounted-rubber-polishing-burrs

Then I'm not sure what you meant by you didn't disassemble but you do need to expose ideally your cleaning fluid to the pivots themselves and disassemble the jewel assembly to get them clean.

On 9/26/2024 at 8:49 AM, praezis said:

And you do not want 275 deg vertical, for you will get differing rates in the vertical positions then - 220 is optimal

Amplitude versus positional errors is really quite interesting. Which becomes quite amusing for this group where they would be very unhappy if your watch did 220°. But if you can have a watch running at 220° for 50 hours it would providing nothing else was incorrect would keep phenomenal timekeeping. Versus this group that wants to see a certain minimum amplitude or else of very very high and almost never pays attention the amplitude 24 hours

Here's an interesting image out of the Greiner timing machine manual. We can say that has the amplitude goes beyond 220° there is less influence on positional errors then at lower Amplitudes

On 9/25/2024 at 7:53 PM, mbwatch said:

With either mainspring, the movement runs dead flat on the timegrapher with at most a 10sec delta. The old mainspring achieves about 275° horizontal but drops way down to about 205°-210° in the vertical positions. A 70 degree drop seems excessive especially since it falls well below 220

Now let's go back to the original problem is your watch behaving as the manufactured specified? So for the – one version they're not as obsessed with timekeeping as much as they were with the – two version. In other words here they're only looking at into positions versus three for the newer and the amplitude is different at least for the low amplitude here the older version as long as you're over 180° at 24 hours that would be acceptable providing a watch is still keeping time contrary to this discussion group the most important thing to a consumer or typically a watch owner is does my watch keep time. Then by keeping time that does not mean impressing the timing machine that means 24 hours later your watch is still on time

now back to your original problem of I assume your winding up the watch nice and tight are probably not following a timing procedure of letting it run about 15 minutes before placing it on the timing machine? So which case you're going to artificially high numbers. As you can see ETA could care less about amplitude at the maximum other than not exceeding the maximum but your vertical positions are dropping way too low and my guess of me at 24 hours you're going to fall below the minimum to your watch is having a problem.

 

[mbwatch]

On 9/26/2024 at 7:30 AM, nickelsilver said:

Curious what the size difference is between the mainsprings.

I measured them today and the ST36 spring is a bit stronger.

Original 6498 mainspring: 0.17mm

ST36 spring: 0.185mm

13 hours ago, JohnR725 said:

On 9/25/2024 at 9:53 PM, mbwatch said:

I would like to use the stronger spring to take advantage of its longer power reserve

The running time of the watches by the design of the watch with the length of the mainspring being very important. So in other ways we can get you a really powerful mainspring that's really short and you're going to have no run time at all.

Ah, I did not mean to imply this ST36 spring would run longer because it was stronger. I expected it to run longer because it IS longer. About 430mm vs 455mm.

 

13 hours ago, JohnR725 said:

But let's look at the technical documentation. In the technical communication for the 6498 –1 We get a minimum running time of 46 hours with atypical of 53 hours.

So on my initial full-wind-run-to-stop of the original 6498 spring, It lasted 48 hours to dead stop. Rate was declining precipitously after about 44h. I own watches running a recent ST36 which all consistently run about 52-53 hours, and I was hoping the additional length on the ST36 spring might translate to another 2-3 hours of runtime.

BUT: That does not turn out to be the case. This ST36 spring actually lasts just 46 hours in this watch.

Ranfft (not official source of course) lists the mainspring as 1.5 / 0.18 / 450. The ST36 spring measures just marginally thicker and longer though I doubt 5mm of length makes a difference. But the 0.005 additional thickness could certainly be enough to push it over the limit into rebanking.

14 hours ago, JohnR725 said:

Where sometimes things that are bright shiny in reality are not actually good pivots they only look bright and shiny but looking at the website here of providing you get the right grit it does indicate you can polish but the wrong Grit you could start doing damaged your pivots

https://eternaltools.com/products/eveflex-mounted-rubber-polishing-burrs

Then I'm not sure what you meant by you didn't disassemble but you do need to expose ideally your cleaning fluid to the pivots themselves and disassemble the jewel assembly to get them clean.

I started using the EVEflex finest grit to lightly clean pivots on the recommendation of an Alex Hamilton video. I have been satisfied with the results on several older pocket watches. I do just a couple of rotations on train pivots to clean off anything my cleaning process misses.

When I say I didn't disassemble the balance - I did not remove the balance complete from the balance cock. I merely soaked the whole assembly in naphtha with the inca jewels removed and separately cleaned. Because I don't have a rotational cleaner and do not put hairsprings in the ultrasonic, that means these balance pivots did not get much cleaning fluid agitation. If I do remove the balance complete from the balance cock, I will be able to do a quick extra clean with the EVEflex pin polisher and any additional residue or debris the naphtha did not clean should be removed. If there's an issue with the balance pivots affecting vertical amplitude, the first thing I must do is clean them to the best standard I'm currently capable.

(I don't have a jacot tool, cannot burnish them properly)

14 hours ago, JohnR725 said:

now back to your original problem of I assume your winding up the watch nice and tight are probably not following a timing procedure of letting it run about 15 minutes before placing it on the timing machine? So which case you're going to artificially high numbers. As you can see ETA could care less about amplitude at the maximum other than not exceeding the maximum but your vertical positions are dropping way too low and my guess of me at 24 hours you're going to fall below the minimum to your watch is having a problem.

No - I always let it run at least 20 minutes on a full wind before recording the TG result. Today I have the original spring installed again and it is on about 12 hours. I don't recall recording it at 24 hours before but I do think it was still over 190. Right now at 12 hours it's performing acceptably 270h / 225v and if it tomorrow morning it's still in spec above 180v I'm going to stop worrying about this and use the old spring.

Anyway I'm going to report tomorrow morning what 24 hours reads like. I think I am ruling out the ST36 spring because it isn't actually delivering the runtime I expected, while also causing rebanking headaches. I still might like to replace it with the proper strength and an S curve because I expect the last 4-6 hours of runtime would see a more consistent rate.

Also this is all academic and skill building for fault finding because I'm building a watch for myself and am very unlikely to ever wear it more than 16 hours at a time or even on 2 consecutive days.

[mbwatch]

26 hours on the old spring and it is running at 245° horizontal, 205° vertical so well within ETA's spec. PD is a little lower than PU at around 198°. And the rate deltas are still <15s.

Now I also did note that when it was more fully wound yesterday that the vertical amplitudes had risen from the 210 I was seeing in its first 3-4 days running up to 230. Horizontal did not change much, maybe +5. But at this point the watch has been running almost continuously for a full week and the unacceptably low vertical amplitudes have improved. I'm going to leave it now.

[mbwatch]

I did clean the balance pivots with a pin polisher and got a marginal increase another 3 degrees.