Seiko 5606 positional accuracy troubles

[guiri]

I bought a nice cosmetic condition Seiko Lord Marvel 5606 a few months ago (movement bright and shiny too), with no date wheel issues.

It's been a long debugging process ever since. it was running when I bought it at a local vintage market, but it would randomly stop. After putting it on the timegrapher, it was running pretty low amplitude (160-ish), so I figured I'd try servicing it. I've been stuck for a while now after trying a number of things, and I'm looking for some input.

(everything has been lubricated with appropriate Moebius 9010, 9103, and 8217 for barrel walls, per Seiko tech manual)

Steps I've already taken so far:

- Replaced barrel and mainspring with NOS parts (no major improvements)

- Replaced pallet fork with NOS parts (didn't want to run, probably needs pallet stone adjustments, which feels above my skill level)

- Replaced pallet fork, escapement wheel and escapement bridge with used parts (back to running poorly after this swap)

- Replaced balance & hairspring assembly with NOS parts (definitely improved amplitude, in dial up, at least) 

At this point, my problem is simple to describe:

- In dial up position, it's running 0-.5 beat error, 200-210 amplitude, and I can zero it in timing and it's super stable.

- In every other position, amplitude is 160-170, beat error jumps around a but (still well under 1), and timing is all over the place, changes every few seconds.

I've closely inspected the balance jewels with a 25x loupe, cleaned/assembled (naptha & alchohol + toothbrush, and Rodico) the running gear numerous times (with less oil each time), and inspected gear teeth/pivots and endplay (which is fairly tight everywhere) as closely as possible. The original pallet fork looked a little weird, which is why I replaced it already.

My only thought from this point is that I should find another mainplate, and give that a shot. Perhaps there's a bad jewel I'm not seeing.

Thoughts?

P.S. I know a pro could fix it...I'm trying my best to learn as a hobby.

[rogart63]

How does the holes for the mainspring barrel look? Both in the mainplate  and the trainwheel bridge? That could be a problem with Seiko movement. I don't think you need to adjust the stones on a NOS pallet fork. 

Check if the hairspring touches something. It self or the bridge. 

Edited August 6, 2017 by rogart63

[clockboy]

Probably to much end shake with the balance. If I get a great difference in different positions then this is the first area I look at. Check for play with the mainspring barrel which can be wear with the  manipulate. In my experience Seiko,s do not give a great amplitude but seem to run just fine with a 220/250 amplitude.  

[guiri]

Sorry I forgot the mention...the mainspring arbor port seems within reason (I watched a lot of Spencer Klein videos to learn what OK looks like  ). That could be it, but I didn't see any brassing etc. It also seemed a bit tighter after swapping in a new barrel.

 

Balance endshake makes sense to me. Does that imply I should replace the endstones diashock frames? I guess I'm just not sure right off the bat how I would tighten up excessive endshake.

[Cmmb8519]

Try to put the all movement in the demagnetizer. I've seen surprisely big improvements with the Seiko movements by just doing this. It does not match necessarily with diferent behaviour on different positions, but it will not harm.

Sent from my SM-G950F using Tapatalk

[guiri]

7 hours ago, Cmmb8519 said:

Try to put the all movement in the demagnetizer. I've seen surprisely big improvements with the Seiko movements by just doing this. It does not match necessarily with diferent behaviour on different positions, but it will not harm.

Sent from my SM-G950F using Tapatalk
 

It seems like almost every watch I get in the mail has been at least partially magnetized, it's one of the first things I check for. This one has been demag'd early on, even before replacing the hairspring.

[clockboy]

Reducing end shake is skill that takes practice and the use of the correct tool.  If the balance staff is OK & the end stones in good condition then you will need a Seitz jeweling tool or a Horia Jeweling Tool. The link below  shows how the guy reduces end shake in train wheels but the method applies to the balance as well. I have adjusted end shake in that past by moving the lower jewel setting in a tiny amount. I presume this is the correct method and it works. 

 

http://raulhorology.com/2012/07/eta-6498-the-gear-train-adjusting-end-shake-truing-wheels-lubrication/

[rodabod]

Out of interest, if it is indeed excessive balance endshake which causes poor timing across positions, what is it about the change of position of the balance which typically affects it's rate? Is it maybe the interaction of the impulse Jewel and the pallets?

[clockboy]

37 minutes ago, rodabod said:

Out of interest, if it is indeed excessive balance endshake which causes poor timing across positions, what is it about the change of position of the balance which typically affects it's rate? Is it maybe the interaction of the impulse Jewel and the pallets?

I think so but not sure all I know reducing end shake reduces positional error. What I don't understand is, if the jewels are OK & the balance is OK why there is to much end shake. Perhaps the balance bridge has become distorted or a wrong length balance staff has been fitted or maybe the balance & jewels over time just wear. I do know that the Rolex 3135 allows for  this by an adjustments fitted to the plate. 

[rodabod]

I've had a worn balance staff before, so I guess end-shake can increase there. I've just been re-jewelling an IWC balance cock, so am obviously interested in the potential effects on positional variation. In my case, I had to increase the endshake by moving the jewel 1/100th of a mm at at time until it was completely free, so the endshake certainly won't be excessive.

[guiri]

Correct me if I'm wrong, but with the Seiko diashock system, the endstones and diashock frames determine the endshake. As such, if they're seated properly, and the balance bridge isn't distorted, it should be correct?

P.S. I noticed in the technical manual, there are distinct part numbers for the upper/lower diashock frames. I can't visually see any difference...perhaps I have the frames on the wrong side? Does anyone know how to tell them apart? (I didn't keep them separate during cleaning).

[clockboy]

Not sure if you are correct. However I have encountered this issue with many watches (normally vintage) and with Seiko's. In theory there should not be any end shake if the endstones /jewels are OK and the balance is OK but it happens.
I recently restored a vintage Smiths watch for a friend and positional error was an issue but after moving the lower jewel setting a tiny amount the fault was eradicated. Someone with  more experience may have the answer.