Thanks all for the further replies. I have used Mitchell's AHS notes from 1978 previously to set the clock up although they did not really mention the latch side of things. I am happy the pivots are in good nick and nicely lubricated. Also the gathering jewel
conforms to the dimensions that Mitchell specifies. Having now checked, I can see that the face of the latch is a bit pitted so it clearly needs some polishing. I will give this a go next.
Out of interest there does seem to be a difference between various setting up guides with respect to the position of the roller on the slope when the pendulum is at rest. Mitchell says it should be half way down the slope whereas other guides say it should
be at the top?
It sounds as if your clock has never really been running properly, as the amplitude you describe has always been very low.
More often than not in all things horological, a problem is not simply down to a single isolated cause, but is rather down to a whole combination of factors. As regards the mechanical mechanism of you clock, you need to work through the entire thing in a logical
and methodical manner, taking care of each parameter in turn - don't just focus on a single aspect. Might I suggest the following:
- Disassemble and clean everything, so that all parts are properly de-greased.
- Peg out all pivot holes: count wheel, armature, latch, gravity arm, back-stop, impulse roller.
- Check that the back-stop roller is perfectly free, and in good condition.
- Closely examine the condition of all pivots - they should be bright, perfectly cylindrical and dead smooth. Any scoring. pitting, rust or other damage must be dealt with.
- Check that all bushings/pivot holes are a good fit for their corresponding pivots. All bushings must also be tight in the back plate.
Now, to deal with your latch/gravity arm noise, I personally have seen these parts in very rough condition. Examine with a watchmaker's loup the condition of the right-hand edge and lower face of the wee steel part which is screwed into the gravity arm, and
which is in contact with the latch. Also look at the corresponding right-hand edge of the lower part of the latch. These edges rub against each other during impulse and reset. These edges must be absolutely dead smooth, without any burrs or sharp corners,
lumps or pitting etc, and polished to a mirror finish. Then a tiny smear of clock oil will assist the action, although it will tend to spread around the latch. The point is that if these edges are damaged, rough or pitted then some of the gravity arm's gravitational
energy will be wasted in getting the wee steel part on the arm to slide down that edge of the latch, and this could also be producing the noise you mention.
Next, look at the tension of the latch spring? Is it original, or has it been messed with? It should be fairly light (although too light will cause mis-locking on reset of the gravity arm).
Once you are certain this is all good, then re-assemble each component in turn, and check that it runs perfectly freely. Once you are happy that all is well, reassemble everything and oil the pivots with clock oil.
DIsassemble and check the door slave in the same way
The moment of unlocking of the gravity arm relative to the position of the impulse pallet on the pendulum is also critical. This is controlled by the length of the wire for the gathering jewel, the position of rest of the count wheel as dictated by the back-stop,
and the position of the wire vane on the arbor of the count wheel.
Then, the rest of the set-up needs to be checked methodically: the case firmly screwed to the wall with at least 3 decent screws, perfectly vertical in 2 planes; the position of the impulse roller with the pendulum at rest, all of the buffer gaps, the current
etc etc. All of this can be found, in FHJ's "Electric Clocks and How to Make Them", or in Miles' book, or on the hvtesla website:
I know this may sound like a lot of work, but I would argue strongly that a methodical approach is the fastest way to produce positive results, unless you are sure exactly where the problem lies.
Best of luck