5. Replacing a balance staff

16 04 2013

Clicking on the figures will provide you with an enlarged view. Use the back arrow (top left corner) to return to the text.

Recently, a Kirova MX6 chronometer, which had been keeping a very good rate within half a second a day, began to have a very irregular rate, sometimes amounting up to six seconds a day, so it was plainly time for an overhaul. The irregularity could have been due to variations in power, though by now the conditions required for an isochronous balance spring are well understood. If a balance is isochronous, it means that its period of oscillation  does not vary with the amplitude of oscillation. However, since the instrument plainly needed attention, I checked for causes of loss of power due, say, to a dry mainspring or rusted links in the fusee chain, but these areas seemed to be fine. I removed the mainspring and replaced the grease anyway, and checked each link of the chain for free movement. The chain too received a good clean and some fresh lubricant. The wheels and pivots of the train were also checked for defects.

When it came to the balance staff, I found one or two pieces of cotton fibre, visible only under a microscope and sticky, thickened oil which had spread from the upper balance pivot on to the shaft of the staff. After cleaning it up, it was then just possible to see with the naked eye that there was a groove in the upper pivot and this can be seen in Figure 1, which shows the isolated staff, with a scale of millimetres above, to give some idea of how it compares in size with a pocket watch staff. It is obviously much bigger and more robust, and so easier and safer to handle than the latter (you can enlarge the photos by clicking on them).

Figure 1 : Balance staff

Figure 1 : Balance staff

As an aside, when examining things like staffs under a microscope, it helps to make a little jig, assembled from pieces of microscope slide, to hold the part horizontal. In the jig on Figure 1, I glued two uprights to a slide, having first filed vee notches in the uprights to prevent the staff from rolling off. Thus under control, the mechanical stage of the microscope can be used to move the slide around for systematic examination. Figure 2 shows the condition of the upper pivot with dimensions added. A pivot diameter of around 0.2 mm is common for chronometers, while pocket watches typically are under 0.1 mm.

4 Cut pivot

Figure 2: Cut upper balance pivot.

Plainly, something is very wrong, as the pivot is “cut”, maybe as the result of rough handling or careless assembly causing the hole jewel to be chipped or cracked or perhaps due to defective heat treatment of the staff. However, the staff shown is hard chromium-plated and, although I could see no defect in the jewel, I replaced it anyway. For the staff there were three options of increasing difficulty: to replace the staff or re-pivot the existing one or make a whole new one, and as spare parts are still occasionally available for this type of chronometer, I took the easy way out and simply replaced the whole staff. Figure 3 shows the over-all length of the staff being checked using a bench micrometer. The dial indicator on the left is set to zero against a slip gauge, seen at bottom left, with the micrometer reading zero. When measuring a part, the micrometer head on the right is rotated until the indicator again reads zero and the reading on the micrometer head is added to the length of the gauge.

Figure 3: Checking new staff for length.

Figure 3: Checking new staff for length.

To make matters easier, it is helpful to take  photographs or make sketches of the relative positions of the balance spring and the rollers so they can be replaced in the same positions. Figure 4 shows the balance from above with a line added to show the position of the upper balance spring stud. The timing nuts are at three and nine o-clock. The lower balance spring collet is split as shown at about 8 o-clock to allow adjustment. To remove it from the tapered shaft, insert the tip of a screw driver into the split and rotate it while drawing it upwards. If it doesn’t want to move upwards, pry gently from below.

Figure 4: Angular position of balance spring

Figure 4: Angular position of balance spring\

Figure 5 shows the angular positions of the discharge and impulse pallets and rollers. Note that the acting faces are radial and that the discharge pallet is much smaller and therefore it is harder to be exact about its position. The position of the balance spring stud is shown by the white line. It is about 180 degrees from the impulse pallet and the latter is about 80 degrees from the discharge pallet. The balance wheel is removed after releasing two countersunk screws and stored safely with the spring.

Figure 5: Angular positions of impulse and discharge jewels.

Figure 5: Angular positions of impulse and discharge jewels.

This then allows an approach to removal of the rollers, but first it is wise to check the longitudinal position of the discharge pallet. The end of that shown in Figure 6 is resting on the face of the impulse roller and the act of removal may cause it to jump out of its slot. It is as well to be prepared for this and perhaps save many anguished minutes looking for a tiny sliver of ruby. If it does fall out, it is a simple but ticklish job to re-shellac it into place and I will perhaps cover this in a future post. Note the flat on the roller. This makes adjustment using a close-fitting wrench easier.

Figure 6: Impulse roller and jewel.

Figure 6: Impulse roller and jewel.

Figure 7 shows the rollers and staff set up in a roller removal jig. That shown is simply a piece of 6 mm steel with a vee slot cut into it and the face of the vee relieved to a suitable thickness such that the vee will fit between the impulse roller and the balance wheel collet. A light, sharp tap with a hollow punch that fits over the pivot will usually cause the staff to drop out of the rollers, and it can then be put away safely to await re-pivoting.

Figure 7: Roller removal.

Figure 7: Roller removal.

The balance wheel is then attached to the new staff so that the rollers can be replaced in the correct angular relationships. Figure 8 is a posed photograph without the obscuring wheel, to show how the staff is set up on a stump in a staking tool with a hollow punch pressing the rollers back into place. They should not be hammered into place, especially not the discharge collet, as doing so may split it as it is forced down the taper of the staff.

Figure 8: Posed replacement of rollers.

Figure 8: Posed replacement of rollers.

Once the wheel and rollers are back in place, hopefully in the correct angular relationships, the balance may be poised as shown in Figure 9. Poising is perhaps less important in a chronometer than in a watch, since a chronometer is always face up with the balance staff vertical, but it is worth paying attention to detail with such a precise machine. There is a variety of  poising instruments including those with adjustable ruby knife edges on a mounting with a spirit level, but a pair of truing calipers used as shown in Figure 9 seems to serve as well, at less expense. One or other of the timing nuts (not the compensation weights) is adjusted until the wheel shows no tendency to come to rest at any particular position

Figure 9: Poising the balance.

Figure 9: Poising the balance.

It then remains only to replace the balance spring and refit the balance in the chronometer. At rest, according to most authorities (and there is precious little in print) the discharging jewel should be resting on the front face of the passing spring, just about to release a tooth of the escape wheel, but in the MX6 it is at rest about 10 degrees on the other side of the passing spring. Any adjustment to the lower spring collet alters its relationship to the discharge and impulse pallets. The position of the latter should be such that it and the escape wheel tooth are moving at the same speed when the tooth catches up with the pallet. If the angle is too little, the pallet may miss the tooth altogether and if it is too great, there may be insufficient impulse given.

Details of the action of a chronometer escapement are given in great detail in Chapter 2 of The Mariner’s Chronometer and some advice about its adjustment is in Appendix 1. I may attempt a systematic account in a future post. An alteration to any part of the escapement affects other parts and it is easy to go around in circles.





4: Hamilton Model 21 Escapement

12 04 2013

Note that figures may be enlarged by left clicking on them. Return to the text by clicking on the back arrow at top left.

At the end of the First World War, the United States of America had the largest of any navy, with nearly half a million personnel. A period of retrenchment followed the Treaty for the Limitation of Naval Armament of 1919 but in 1934, as world tensions seemed to be increasing, a program of ship building began to bring the US Navy up to the maximum size allowed by the London Naval Treaty of 1930. The Naval Act of 1938 contemplated a 20 % increase in building and the Two Ocean Navy Act led to a further 11 % increase, but this was soon followed by a plan to increase tonnage by 70 %, equivalent to 200 more ships. At that time, there were no manufacturers of chronometers in the USA, though there were firms that assembled them from imported parts and movements. The total annual world output of chronometers was probably in the region of 300 a year, principally from Britain and Switzerland. At the outbreak of the Second World War on 2nd September 1939, these sources became unavailable. Britain needed all her own chronometers while Switzerland, surrounded by belligerents, was soon to find that de facto limitations were place upon whom she could supply.

When WWII began, the US Navy had 394 ships of all types and by the war’s end she had 6,768, many of which  required chronometers, two or more for larger ships, one of more for smaller ones and navigational watches for vessels like patrol boats. The Hamilton Watch Company of Lancaster, Pa. was, among others, invited to commit to manufacture of the necessary chronometers and expressed interest in a letter of 2nd July, 1940. The Company was then provided with two Swiss Ulysse Nardin chronometers to examine. E W Drescher  said he thought Hamilton could make the instruments provided that certain design modifications could be made to allow mass production and the first order was placed on 16 May 1941. This accounts for the “1941” that appears on the face of the chronometer. It is not the date of manufacture but of the design. The first deliveries began on 27 February 1942 and 58 had been delivered by the year’s end, rapidly increasing thereafter, so that 8902 had been delivered by the end of the war. In most respects, the Model 21 chronometer closely followed the Ulysse Nardin design, except for a revolutionary balance design and the use of a pre-formed Elinvar-type balance spring, to obviate the time consuming and somewhat intuitive spring adjustments previously necessary.

Thus it came about that the escapement design is that of Nardin and uses a slightly more complex detent than that of the standard Earnshaw spring detent used by practically every other maker. The escapement is the part that communicates with the movement and lets it run down in equal steps while the movement gives up energy via the escapement to keep the balance wheel oscillating. A plan view is shown in Figure 1 and a perspective view in Figure 2, and you may find it helpful to refer back and forth between the two diagrams while reading the explanation. In Figure 2, modified from a figure in the official overhaul manual, the horn of the detent obscures some details of the discharge jewel and so in the inset, the horn has been displaced so this can be seen.

Figure 1: Plan view of escapement (After Rawlings)

Figure 1: Plan view of escapement (After Rawlings)

The escape wheel, driven by the movement, rotates in clockwise steps. The impulse and discharge rollers and their attached jewels are mounted on the balance staff, below the balance wheel and they are shown rotating anticlockwise. In both diagrams the discharge jewel is pressing on the end of the passing spring which in turn is pressing on the horn of the detent. Upon further anti-clockwise rotation, the detent will move its attached locking jewel out of engagement with tooth A of the escape wheel, leaving it free to rotate. Meanwhile, the discharge jewel slips free of the passing spring, allowing the locking stone to fall into the path of tooth B, ready to lock the escape wheel again, and the rollers continue their anti-clockwise rotation. Tooth C catches up with the impulse jewel, delivering some energy to the balance wheel to keep it going, and then tooth B arrives at the locking jewel and the rotation of the escape wheel is arrested.

The balance wheel and the rollers eventually reach the end of their anti-clockwise rotation and return, clockwise, but this time the discharge jewel simply lifts the delicate passing spring off the horn of the detent and continues past to complete the clockwise rotation before returning to recommence the cycle. Thus, when listening to a chronometer going, one hears a loud tick as an escape wheel tooth is locked against the locking jewel and a much softer tick as the passing spring drops back on to the horn of the detent. Unlocking, impulse and locking occur only once per full oscillation back and forth of the balance wheel and for the rest of the time the it is free of interference or detached.

Figure 2: Perspective diagram of escapement

Figure 2: Perspective diagram of escapement

The structure of the detent is most easily seen in Figure 2 and a photograph of an isolated detent is shown in Figure 3.

Figure 3: Model 21 detent.

Figure 3: Model 21 detent.

A foot is attached to a support block with a screw and washer. Two guide pins engage in a groove in the support block and allow only longitudinal movement when the depth adjustment screw is turned. This adjustment determines how deeply the passing spring engages with the discharge jewel, which in turn determines for how long the locking jewel is lifted out of the path of the escape wheel teeth. If it is out of engagement for too long, locking on the next tooth may fail and the chronometer is said to trip. It will then go twice as fast as it should. The support block allows the detent to be removed from the chronometer and replaced without disturbing its adjustment in relation to the rest of the escapement.

The detent spring in the Hamilton Model 21 is in two parts, unlike most chronometers which have a simple leaf spring, typically about 0.04 mm thick by 2 mm wide. Having two parts cannot be to allow a thicker spring to be used, as the stiffness of a spring varies directly as its breadth and as the cube of its thickness. Possibly it was felt that the increased effective breadth resisted torsional forces better. In any case, the spring must be stiff enough to resist buckling as the escape wheel tooth locks on the locking jewel while being weak enough so as not to interfere significantly with the balance wheel when unlocking.

The slender and flexible passing spring is mounted on a Z-shaped bracket. However, it must not be so flexible that it cannot cause unlocking, nor should it be so strong that on resuming its seat on the horn of the detent the percussion causes the locking jewel to release a tooth. The lock adjusting screw determines the position of the stop button which in turn determines how deeply the locking jewel engages with the escape wheel teeth. It must be deep enough to lock despite any shake in the escape wheel bearings. Depth of the passing spring and lock, and the angular position of the discharge jewel in relation to the impulse jewel are interdependent; if any one is disturbed, the others may need to be adjusted to regain an efficient action of the escapement. The official Manual for Overhaul, Repair and Handling of Hamilton Ship Chronometer explains in detail how to carry out adjustments for the Model 21 and Appendix 1 of my The Mariner’s Chronometer does so for chronometers in general. Without a clear understanding of the action of the escapement and the effects of individual adjustments, one is best advised to leave well alone, and if the instrument has fallen out of adjustment through wear or accident, expert help and wide experience may well be required to put it right.





3. Hamilton Model 21 chronometer: Getting started

11 04 2013

Note that all illustrations in the blog may be enlarged by clicking on them; and reading of the text may be resumed by clicking on the Back arrow (top left)

I have described in “Getting started” how to remove wedges from the balance of a chronometer of standard design and at first the Hamilton watch company immobilised the balance for transport the same way. By the time that the official Manual for Overhaul, Repair and Handling” of the Hamilton Model 21 chronometer was published in 1948, a new method of immobilising the balance was being fitted to chronometers by the US Navy, as the chronometers became due for overhaul or repair. This locking arm is shown in Figure 1.

Figure 1: Balance locking arm

Figure 1: Balance locking arm

To lock the balance the locking screw is loosened and the locking arm swung until the hole in it traps a large timing nut on the balance rim, when the screw is re-tightened. Unlocking is the reverse of locking. Although this places the balance and its pivots at less risk than when inserting wedges it was still possible for the ham handed to cause damage, and it was still necessary to remove the instrument from its gimbals and open the case. At some time after 1948, a balance locking fork was fitted as chronometers were overhauled, allowing the balance to be unlocked without opening the case. If on inverting the chronometer a small hole in addition to the winding hole is present in the dust cover, a fork is probably fitted, provided the movement is in its own case. Figure 2 shows  the hole and on rotating the cover, the socket of an Allen screw is revealed, before the winding square comes into view in its own hole

Figure 2: Locking balance

Figure 2: Locking balance

If the original key is absent, any 1/16th inch AF (across flats) Allen key will do to loosen or tighten the screw and, as the screw is captive, you need not fear that you will over-tighten it and damage the movement or unscrew it completely so that it gets lost in the works. Figure 3 shows the fork itself. There are two soft plastic pads on the ends of the arms and when the screw is tightened against the springy stem, the pads bear on the balance rim and bring it to a halt.

Figure 3: The locking fork.

Figure 3: The locking fork.

When servicing a Hamilton Model 21 chronometer, the fork and the bridge that straddles it have to be removed before the upper balance cock and balance can be removed. As Figure 4 shows, the bridge is attached to the plate by two screws, while the fork itself is secured with a single screw. Once the bridge has been removed, the fork is supported with tweezers while its attaching screw is removed.

Figure 4: Attachment of the locking mechanism

Figure 4: Attachment of the locking mechanism

By 1970, 13,086 Model 21 Chronometers had been produced, about 11,000 of them for the US Navy, the US Maritime Commission and the US Air Force during WWII. The majority of the ones I have seen are fitted with the locking fork.