Bushing a worn pivot hole without making it worse

I

Wear runs in one direction and says so

Pivot holes do not wear round. They wear along the line of pull, because the force on every arbor in a going train comes from one direction and stays there for decades. The mainspring pulls the great wheel, the great wheel leans on the centre pinion, and each hole elongates towards the wheel that is driving it. The direction is useful, because it means the shape of the hole tells you which train it belongs to and whether the wear is honest or whether somebody has been in before and changed the geometry.

The first thing I do with a stripped plate is hold it up against the north window and look through each hole at the light. A worn hole shows as a bright oval with a dull shoulder on one side where the pivot has been polishing the brass. On this particular movement, a French eight-day of the sort that sits under a dome, the third wheel hole in the front plate was so oval I could read it from a metre away. The second wheel hole was a little over, the centre wheel hole was barely touched, and the escape wheel hole in the back plate had been bushed at some point by somebody else and the bush itself was already oval, which tells you the earlier job never dealt with the reason it wore.

The temptation at that point is to bush everything and call it a rebuild. I have done that and it does not help. A hole that is round and tight is not improved by being drilled out and filled with new brass, and every bush you press in is a chance to shift the depthing of two wheels that were previously meshing correctly.

II

Measuring, because the eye flatters you

I measure every pivot with a bench micrometer that reads to a hundredth of a millimetre, and I write the figure down before I form an opinion. On this movement the escape wheel pivot came in at 0.48 mm, the third wheel at 0.62 mm and the second wheel at 0.94 mm. Then I find the largest smoothing broach that will pass through the hole on its narrow axis and the smallest that will not jam on the long axis. The difference between those two is the wear, and on the third wheel hole it was a little over a quarter of a millimetre, which is far more than it looks like through a loupe.

The pivot itself decides half the job. If the pivot is polished, straight and unscored, a new bush gives it a clean bearing and that is the end of the matter. If the pivot has a ridge or a groove cut into it, bushing the hole only hands that groove a fresh surface to chew, and you will be back inside the movement in three years. The third wheel pivot on this one measured 0.62 mm at the tip and 0.59 mm where it met the shoulder, with a ridge you could feel with a fingernail.

That pivot went in the lathe and got about four minutes against a burnisher, turning slowly, with the burnisher held flat and moved along the length rather than held in one place. It came out at 0.61 mm, parallel, and bright enough to see the window in. Only then did I know what size bush the hole wanted.

III

Broaching square is most of the work

A five-sided cutting broach cuts on whichever side you lean on, and a plate is easy to hold at an angle without noticing. The failure I see in other people's work, and made plenty of times in my own, is a hole opened square to the hand rather than square to the plate. The result is a bush that sits in a cone, a wheel that is not upright, and a depthing that is tight through one quarter of a turn and loose through the next.

So the plate goes flat on a small steel block with a clearance hole under it, the broach stands vertical, and I check it against a small square from two directions every quarter turn. I open the hole to the smallest bush that will clean up the oval completely, which usually means the bush is a size larger than I would like. Then the bush goes in with a flat-faced punch in the staking tool, one firm stroke rather than five hopeful taps, because five taps will walk it over.

After that the new bush gets opened with a smoothing broach until the pivot has a shake of roughly five hundredths of a millimetre. That is enough to hold a ring of oil and let the arbor turn, and not enough to let the wheel wobble under load. The bush face wants to sit a whisker below the plate surface on the oil sink side, so the sink still holds oil rather than letting it run out along the brass.

IV

The two I pressed in crooked

I did the third and fourth wheel bushes on a Thursday afternoon in November with the window light going and the bench lamp not yet switched on, and that is the actual root cause of everything that followed. The third wheel bush went in with a lean I could not see on the bench but could see immediately when the plates went together, because the wheel stood visibly out of upright and the train would run through half a turn and then stop dead.

Getting a pressed bush out without ruining the hole is a slow job. I ran a cutting broach into it from the oil sink side just far enough to break the grip, then pushed it out with a flat punch from the other side, and checked the plate hole afterwards for a burr pulled up by the broach. The hole was a tenth larger than it started, so the second bush was a size up again, which is exactly the creeping damage I try to avoid.

The fourth wheel hole went wrong in the opposite direction. I opened the new bush one stroke too far with the smoothing broach and finished with a shake around twelve hundredths, loose enough that the wheel visibly shifted when the movement was turned over. That one came out too. Two holes took a whole morning. The note in the job book says: if the light is going, put the broach down.

V

Fourteen days on the board before it goes back

Once the plates are together and the train runs free under a light push, the movement goes onto the wall board behind the bench with a weight or a wound spring and stays there for a fortnight. A movement that has been bushed will almost always run beautifully for the first two days. The question is what it does on day nine, when the oil has spread and the mainspring is delivering a good deal less torque than it did on the first morning.

This one gained eleven seconds on the first day and lost four on the seventh, which is an ordinary spread for an eight-day going barrel with no fusee and nothing wrong with it. By day fourteen it had settled to a loss of about two seconds a day, and the oil rings at the newly bushed holes were still circular rather than smeared. I pulled it down once more to check the shake at each bearing, found it unchanged, and only then wrote the job up as finished.