Jump to content

Eta 900 timegrapher trouble shooting


Recommended Posts

IMG_20200930_214535.thumb.jpg.89d714811b40c5f46f877cd492d8e495.jpgI finished cleaning and reassembling an Eta 900 movement, it got a new mainspring and balance complete, both NOS. The timegrapher shows a number of oddities, as shown in the photo.

- The watch was at perhaps 1/4 wind for the timegrapher photo, at full wind the amplitude was 310-320 which seems too high.

- The beat rate is recognised by the Weishi as 19800, but the movement is listed as 18000. 

- The watch runs either really late or slow.

- A clear abnormality in one of the 2 vibrations.

- BE is huge, but that seems the least of the problems for the time being.

After seeing the high amplitude I've removed the grease from the barrel arbor to increase friction. I've cleaned the escape wheel and pallet fork again, and re-checked for abnormalities, can't find any.

Any ideas what might be going on?

 

 

 

Link to comment
Share on other sites

Thanks for the input so far, I will see what a fixed 18000 bph will show on the timegrapher. I will dive further in the ETA 900 beat rate, see if it is consistently mentioned as 18000. I will check the hairspring, no idea how I would check the vibration frequency of the balance (I just started on this watch repairing journey).

Link to comment
Share on other sites

Then just a reminder notice your beat error of 6.9 ms and the graphical display looks pretty decent? Because of the size of the graphical display it will only display up to about 4 ms in that it has a rollover error Which gives impression that you beat looks really nice but the numeric rate is probably correct And the display was much bigger those lines would be really far apart.

Out of curiosity how did you lubricate the escapement?

Then I was looking at the parts list for this watch to see if there is any variation in or indication of different beats and there didn't seem to be any variations that way. There were variations for height of parts on the dial and from the link below the As usual with an older movement looks like variations but no indication of variations for timekeeping.

Why did you get a balance complete for the watch? Then a little about the history of its condition before you started servicing it? Didn't run for instance?

 

http://www.ranfft.de/cgi-bin/bidfun-db.cgi?10&ranfft&0&2uswk&ETA_900

  • Like 1
  • Thanks 1
Link to comment
Share on other sites

Check the picture of the example on fanfft for some clues as to how the balance should look. Is the regulator arm miles out from the example. 

Check for fouling hairspring, sticky hairspring, end shake etc.

Finally, if you suspect the balance may be the wrong one for the watch, and you have the original balance, If you have some very accurate scales, then compare the weights of the two balances (removed from the cocks, as these may skew the results).

If they differ significantly in weight, but are of similar construction and diameter, then it is may be the wrong balance. This is not necessarily going to be an accurate test, as it takes not account of the strength of the hairspring or the geometry of the balance. You really would need to check the balance more accurately to determine how it oscillates to be sure.

  • Like 1
Link to comment
Share on other sites

Stud holder is fixed in eta900 and hairsprings came in flat and bereguet overcoil versions, either way if your BE is stemming from wrong H/S and impulse pin mal- adjustment then adjusting would be a pain. 

Just because it is a NOS doesn't mean its the right one,  the new mainspring can be too strong, usually MS coil should cover a third of barrel bottom surface area, if strong the mainspring is thick therefor covers more than third of the area.

A NOS balance complete can easily have got magnetized or dirty/sticking.

I test the tg, just test a healthy watch on it.

You can also bench test the watch, put the hands on, despite no dial and cover the watch, check its time kerping over a day.

So, I make sure balance complete is the right one, demaged and clean.

To check for sticking which mostly is a case of intermitent, video it while running and check at slow motion is best, but to sit there and count the beats on video shorten your own life.

Good luck

 

  • Like 2
Link to comment
Share on other sites

The watch wasn't running, but no apparent damage when I received it. The new mainspring I installed because the Stowa is a watch I'd like to start wearing myself. I just read that a newer alloys MS can provide more power and than what the older movements were built for, that could explain the high amplitude. I demagnetised the movement. The HS is a regular flat type. I ordered a NOS Balance complete as well because I had managed to ruin the previous hairspring...doh!

I provided a photo of the timegrapher after manually setting the bph to 18000, the display makes more sense in terms of BE, but after 7-8 seconds the graph is reset and the process starts again! I have tested the timegrapher with watch that keeps decent time.

I will have a good hard stare at the Ranfft reference images, and I will have a closer look at end shake. Thanks for all the responses so far, supercharges the learning process!

IMG_20201001_130158.jpg

Link to comment
Share on other sites

4 minutes ago, Nucejoe said:

Surely you know smaller wheel doesn't neccessarily mean different beat.

This new mainspring is excessively strong.

As for the BE, I think johnR hit the nail on the head.

 

All things being equal, a larger radius will increase inertial momentum. However I have no sensitive scale at hand to see if there's a weight difference between the balance wheels. At the very least I felt the difference in balance wheels was worth pointing out. 

I can imagine the movement running fast when the high amplitude causes knocking; The oscillation is effectively shortened.

I can see there is rollover error on the display, no doubts about the BE. 

Is there a non-invasive way to increase friction on the main barrel to absorb part of the MS power?

Link to comment
Share on other sites

 Not that I know of, no non- invasives way. 

I agree all things being equal......        but at production line, balance wheel are vibrated with hairsprings which may differ in length or CGS to produce a desired beat.

I am not sure if oiling fork pivot would reduce force of impulse enough to get an acceptable amplitude. What should the amplitude be? Do you have a datasheet?

  • Like 1
Link to comment
Share on other sites

I don’t understand how we can conclude that the new mainspring is too strong. 
 

Your timegrapher is unable to capture a trace which quite possibly means that the vibrations are either very far out of range in frequency, or the impulses of the escapement are too weak to distinguish. Either way, the trace is close to useless. What you may be seeing for the brief period of time when it does appear to capture is the effect of aliasing. An example of aliasing is when a video of a helicopter rotor looks as if it is turning very slowly as a result of the capture sample rate which may be 25 or 30 frames per second.

If you are unable to monitor the rate of time via the hands then is it easy enough to listen to the watch beating with your ear? You could compare to a known-good 18,000 movement and see if it sounds way off  

 

 

 

  • Like 1
Link to comment
Share on other sites

As you're new to watch repair I will start off with the basics. The basics are people new to watch repair look for problems were they don't exist in other words it's usually the simple things it's usually the basic things but those are not the things you're ever going to see.

Then changing components doesn't always have the desired effect of improving things. Unfortunately if you mutilate the balance wheel you have to change that. But changing components has consequences unless the watches super new were manufacturing skills are almost perfect and you can swap with usually zero problems.  Soon always empty careful in changing components as you can introduce new problems.

Then the mainspring yes the white springs can be stronger than the blue Springs  but it usually isn't a consequence.  It's a much bigger problem in American pocket watches where once you might have five different sizes  and now you just have one and it's probably too strong but even there it usually isn't a problem  most of the time.. I snipped out a section of the mainspring catalog to the sizes listed agree with the size the mainspring you put in?

6 hours ago, GaspardColigny said:

The new balance wheel is slightly smaller than the old one!

Where exactly did you purchase your new old stock balance wheel? Then unfortunately I don't have a cross reference to the number you have plus if it's a different size it's not the balance wheel for your watch. As the diameter is wrong I don't suppose the roller parts are wrong in size also? Like the safety roller the roller table the roller jewel are they physically different? Also it be really nice if you could get a close-up picture of the balance wheel in the watch for us preferably one looking straight down and one looking sideways.

Along the lines of  basic problems balance wheels fall under that category it is really easy to do things  undesirable here and it is exceedingly hard to see those things by just about everyone on this group..  It's one place where you have to really really pay attention because there's a lot of faults that are very bad under very hard to see.

If your balance wheels physically smaller is the hairspring physically smaller? Visually when you look at your balance wheel in the watch does they hairspring appear to be centered, is they hairspring going to the regulator pins is it flat is it touching anything it's not supposed to. If they hairspring is physically smaller you going to have issues with it being where it's supposed to be for the regulator pins you're going to have issues.

7 hours ago, GaspardColigny said:

The watch wasn't running, but no apparent damage when I received it.

One of the things nice to do with incoming watches is to attempt to determine why it's not running. As opposed to assuming that cleaning will fix all problems because many times there were pre-existing conditions as to why the watch wasn't running and cleaning doesn't necessarily fix those problems.

Then a new updated picture of the timing machine where it's showing you nothing  that's because the timing machine is trying to reach out and communicate with you unfortunately it does not have artificial intelligence to do that correctly. The same as your original pictures you need to visually verify that whatever the machine is telling you the watch appears to be doing the same thing.. So if the amplitude looks like it's super superhigh visually it should look like at superhigh. Occasionally it's more common problem with Chinese machines  but  it will happen to all timing machines if the amplitude is super low for instance it won't numerically look to be super high. This is where  missing features  from the Chinese machine would be helpful if we had such as the oscilloscope where you can look at the waveform  and see where the timing machine is attempting to  pick up the signal. Because oftentimes you'll find is looking at the wrong part of the waveform so if your timing machine says you're doing 300° that balance wheel better look like it's doing 300° otherwise you have a problem..

So the basics would be  let the power off get us some photographs..  Visually verify the watches in beat. Look carefully at the hairspring if necessary look at another watch or look at several watches as I said it is really really hard to tell if they hairspring is flat and where it's supposed to Be.

Then one little warning about  stock? Typically and watch repair  especially if you buy from eBay  if the package has been opened there is no guarantee that at actually is new old stock despite what the seller might think.. It's a common practice of watch repair people  to put the old part  back into the package so any opened packages should always be suspected that they're not correct.. Unfortunately this also does not apply to watch repair there's been other things I purchased off of eBay that were listed as new old stock  and they were definitely not because the seller wasn't paying attention in my case I wasn't paying attention to the photos either I took them at their word.

 

 

 

 

eta mainspring 900.JPG

  • Like 2
Link to comment
Share on other sites

Thank you for the extensive responses, I will make some photos, I will have to buy a cheap macro lens for my smartphone first. BTW I fully expect to encounter issues with these vintage watches, if one runs decent after service I'll consider myself lucky haha. It's just hoping that the issues I do come across aren't too complex, but I now that's asking for much. 

I also will install the hands and see if the watch keeps time. Upon visual inspection it is clear that the balance oscillates more vigorously than movements that run with 250ish amplitudes. 

So far in my short watch journey I've certainly noticed layer upon layer of intricacy, and that's what appeals to me; much to learn! 

Link to comment
Share on other sites

2 hours ago, rodabod said:

I don’t understand how we can conclude that the new mainspring is too strong. 
 

Your timegrapher is unable to capture a trace which quite possibly means that the vibrations are either very far out of range in frequency, or the impulses of the escapement are too weak to distinguish. Either way, the trace is close to useless. What you may be seeing for the brief period of time when it does appear to capture is the effect of aliasing. An example of aliasing is when a video of a helicopter rotor looks as if it is turning very slowly as a result of the capture sample rate which may be 25 or 30 frames per second.

If you are unable to monitor the rate of time via the hands then is it easy enough to listen to the watch beating with your ear? You could compare to a known-good 18,000 movement and see if it sounds way off  

 

 

 

OP tested his timegrapher with another watch that is keeping decent time, I guess amplitude and beat error has shown normal with the other watch.

  • Like 1
Link to comment
Share on other sites

1 hour ago, Nucejoe said:

OP tested his timegrapher with another watch that is keeping decent time, I guess amplitude and beat error has shown normal with the other watch.

The issue I have with this is that (in my opinion) it doesn’t show good use of logic. I believe that is one of the points that JohnR is trying to make. 
 

I would say it’s fair to comment that the new mainspring *may* be too strong. But we do not have proof of that - too many variables. Again, the point that JohnR is making. 

  • Like 1
Link to comment
Share on other sites

22 hours ago, GaspardColigny said:

 

Is there a non-invasive way to increase friction on the main barrel to absorb part of the MS power?

 If it came to it, regardless of the cause high amplitude can be lowered about 20degrees by oiling fork pivots.

Link to comment
Share on other sites



×
×
  • Create New...