8. Making a chronometer key

30 07 2013

A marine chronometer must never be wound clockwise except for those few that have a winding square on the front or side of the instrument. Winding in the wrong direction risks damage to the maintaining power mechanism and, if reverse power should reach the train, the locking stone and detent may be damaged. For this reason, all makers provided a key that would wind only in an anti-clockwise direction and which would slip if clockwise winding was attempted, rather than rely on the user’s memory. They are often referred to as “tipsy keys.” Most have a spring-loaded dog clutch, but many German Einheits-chronometers (Standard chronometers) and the Russian MX6 chronometers which closely followed their construction used a spiral spring which tightened and gripped the winding shaft of the key when turned anti-clockwise but which slipped when turned clockwise. This latter type is much easier to reproduce than the dog-clutch type and as second-hand chronometers, especially those sold without their cases, are often lacking a tipsy key, the collector is obliged either to source a new key on the internet or to have one made. One maker proposed a charge of 150 British pounds a few years ago, which seems rather a lot. This blog gives the details of one way to make a tipsy key and it may serve as a guide to others.

Figure 1 is a general arrangement drawing which no doubt violates many conventions, but which I hope makes the structure clear. 1 is the winding shaft with a square hole down the end. It is gripped by a helical spring, 5, which is a close fit over the shaft but an easy fit in the body, 2. The helical spring, which is anchored in a hole in the body at one end, must have a clockwise spiral if it is to grip the shaft when it is winding anti-clockwise. 3 is a knob, in this case a simple disc, but can be of any shape that takes the fancy. 4 is a taper pin that hold the knob in place and restricts the outward movement of the winding shaft. It is reamed after drilling the cross holes through the body and knob with the latter in the correct position. The drawings and photographs may be enlarged by clicking on them. Return to normal by using the back arrow, at top left on most browsers

General arrangement sectional drawing.

General arrangement sectional drawing.

Figure 2 is a drawing of the several parts with dimensions added. These of course may need to be varied to suit a particular chronometer, but should work with the majority of chronometers made in the twentieth century. For those not used to reading drawings, the main view is a plan view and the view to the side is what one would see standing on that side looking back.

Figure 2: Dimensioned drawing of parts.

Figure 2: Dimensioned drawing of parts.

Making the shaft.

This begins by facing a length of 8 mm brass rod, centring it and drilling it to a depth of 25 to 30 mm to a diameter that is the same as the across-flats dimension of the required square (Figure 3). The rod needs to have a length allowance to allow it to be gripped securely in the lathe chuck. The deep hole allows one room to manoeuvre when converting the round hole to a square one, using a square Swiss file. You file away, using the remnants of the hole as a guide to when the square is near to size and then try it on the chronometer. Once it begins to enter, it is then simply a matter of raising the filing hand a little so that the square is deepened until 4 or 5 mm will grip the winding square (Figure 4).

Figure 3: Facing, centring and drilling.

Figure 3: Facing, centring and drilling.

Figure 4: Converting a round hole to a square one.

Figure 4: Converting a round hole to a square one.

The proto-winding shaft is then returned to the lathe, the end supported with a back centre, the 6 mm portion turned down to size (Figure 5) and the whole parted off to length (Figure 6). The diameter may of course be adjusted a little to suit available springs. Most seem to be wound anti-clockwise, so your collection of odd springs may have only a few that are clockwise helices.

Figure 5: Turning to size.

Figure 5: Turning to size.

 figure 6: Parting off.


Figure 6: Parting off.

The Body

Making the body is mainly a straightforward drilling and turning exercise, but before parting off from the parent metal, a 3 mm wide slot must be milled across a diameter for the knob (Figure 7). The home workman is more likely to use a slot drill for this, but those well equipped with a horizontal milling machine and a suitable cutter may choose to use this, though the setting up time is probably longer. Of course, the job can also be done with a saw and file.

Figure 7: Milling slot for knob.

Figure 7: Milling slot for knob.

After parting off, the nose is tapered, as shown in Figure 8, taking care to taper the correct end!

Figure 7: Taper turning nose.

Figure 8: Taper turning nose.

Then, using a drill of about 0.8 mm diameter, a hole is drilled which will anchor the lower end of the spring to the body (Figure 9).

Figure 9: Drilling for spring anchor.

Figure 9: Drilling for spring anchor.

The Knob

The disc-shaped knob could be parted off a brass bar 36 mm in diameter, leaving the problem of how to finish the parted off face, so it is much simpler to make it out of sheet material. Those who like exercise can saw it to approximate shape and then file it round using a suitable cheater (see previous post for a note about these useful items), but I chose the easy way of mounting a square on a mandrel and turning off the corners(Figure 10). The edges can be bevelled at the same time.

Figure 10: Forming the knob.

Figure 10: Forming the knob.

The knob is then held in the slot to be cross-drilled and reamed for a taper pin which will hold the knob in place (Figure 11). Taper pin reamers receive little use in industry these days and may be hard and expensive to find, so a parallel pin secured with a dab of industrial adhesive such as Locktite will do just as well.

Figure 12 shows the parts ready to be assembled. The compression spring shown is not ideal as the end tends to jump out of the anchoring hole, but it was the only suitable clockwise spring I could find at the time. I have since replaced it by fitting an anticlockwise extension spring over a 6 mm mandrel and unwinding it. I then cross drilled a hole in a 4.75mm (3/16 in) mandrel to anchor the end of the 0.7 mm wire and rewound the spring, clockwise.

Figure 11: Reaming for a taper pin.

Figure 11: Reaming for a taper pin.

Figure 12: Exploded assembly.

Figure 12: Exploded assembly.

After cleaning up the parts are assembled, tapping the taper pin lightly home and sawing it off to length. A brass pin would have been ideal, but I made do with a steel one, as the chronometer it is intended for will never go to sea again. Figure 13 shows the completed key.

Figure 13: The completed key.

Figure 13: The completed key.

If you enjoy reading this or other of my posts, let your interested friends know so that I will be encouraged by visitor numbers to write more. And of course buy “The Mariner’s Chronometer” for much, much more.





7. Making a replacement lid for the case.

28 07 2013

Nearly all chronometers were supplied with a three tier case, incorporating a base, a hinged glazed lid and a hinged cover for the glazed lid. On board a ship, the chronometer was often mounted beneath a glazed part of the chart table so that the time could be seen at a glance. Where more than one chronometer was carried, they were mounted in a special case with a single lid for the (usually) three instruments. This meant that the cover for the glazed lid was superfluous and it was usually removed with its hinges and eventually lost. The lid was usually provided with brass corners. This seems to be due to tradition rather than necessity, since a chronometer was handled so carefully that it had no need of protection, unlike the nautical sextant which often was and the case of which was never provided with brasswork.

The present-day lover or collector of chronometers may wish to replace the missing lid, so this post gives some details of dimensions and shows how brass corners are made. Chronometer cases were usually made of mahogany, rosewood or lesser timbers stained to look like them. I have a precious supply of well-seasoned African sapele, but beech or birch can be stained to look similar. Figure 1 shows the structure of a typical lid with dimensions which seem to be pretty much the same  whoever the maker, but it is as well to check against the box itself (All the figures can be enlarged by clicking on them. Use the back arrow to return to the text.).

Figure 1: Dimensions and structure of lid.

Figure 1: Dimensions and structure of lid.

The lid is in two main parts, a top 12 mm thick and a frame 15 mm thick rebated and glued into the underside of the top to give a total thickness of 20 mm. Figure 2 shows the underside of the lid to make this a little clearer. I use a simple mitre joint at the corners but those with more complex wood-working equipment may chose something more complex.

OLYMPUS DIGITAL CAMERA

Figure 2: Underside of lid.

Brass corners made by simply folding thin sheet do not have crisp edges and look amateurish, while most of those that I have seen for sale are not correct to form. The metal workers of the past would have filed a 90 degree groove in the brass sheet nearly all the way through prior to folding the sheet to a right angle and running solder into the groove. There is no reason why this should not be imitated today, starting with a saw cut to locate and guide the file along the correct line and, since the result will be out of sight, all manner of scratches and false starts can be made as long as the resulting fold is in the right place and looks crisp. I find it much easier to use a shaping machine and to mark out all four corners on a single sheet of brass so that it is easier to secure on the work table of the machine. Figure 3 shows the dimensions of a single corner and Figure 4 below it shows a groove being cut in a piece of 1.5 mm brass sheet which has been marked out for two sets of four corners.

Figure 3: Dimensions of a corner.

Figure 3: Dimensions of a corner.

Figure 4: Machining grooves.

Figure 4: Machining grooves.

The tool moves backwards and forwards, cutting on the forward stroke and deepening the grove with each stroke. Care is taken not to cut all the way through. I usually leave about 0.1mm uncut. The point of the tool is formed to an angle of 92 degrees so that when the metal is folded, the slight spring of the residual metal in the bottom of the groove results in an angle of 90 degrees and also leaves a little space for the solder to run into. Once all the grooves have been cut, the curves can be sawn nearly to size and the four corners separated prior to removing the little squares of material at the centre. Figure 5 shows me using a piercing saw to separate the corners. Note that the saw is tilted forwards a little, as this seems to make sawing in  straight lines easier, but when sawing curves, the blade must of course remain vertical.

Figure 5: Separating the corners with piercing saw.

Figure 5: Separating the corners with piercing saw.

The curves may now be filed to the finished shape, either by simply filing to the marked line or, more easily, using a “cheater”. This is simply a piece of steel of the required form that can be clamped to the work piece to guide a file. Figure 6 shows one in use. When filing brass, a very sharp file is needed and one that has been used on steel will cut brass only with difficulty, so a deft touch is needed to sense when the file has cut all the brass and is sliding over the steel without cutting into it.  Notice the yellow tape on the handle of the file to remind me that it is to be used only on brass.

Figure 6: Using cheater to guide file.

Figure 6: Using cheater to guide file.

Figure 7 shows corners at various stages of formation, from bottom to top, cut out, cleaned up with square cut out, and folded ready for soldering. When cutting out the square, care must be taken to leave a little metal for cleaning up with a small file, so that a nice clean mitre is left at the corner without any line of solder showing. Once the folds have been made, check that the sides form right angles with the tops and with each other. Solder can then be run into the joints as shown in Figure 8.

Lid replacement 007

Figure 7: Stages in forming corner.

Figure 7: Soldering joints.

Figure 8: Soldering joints.

Seats for the corners must now be cut into the top of the lid. The rebates for the sides of the corners are relatively simple, though a sharp chisel is needed for cutting across the grain and care must be taken not to cause large splits at the end of the cut when doing so. Small blemishes do not matter, as the corner will conceal them. The cut-out for the curve, however, is more difficult and requires a chisel of the right curvature. Gouges of this size are expensive, especially if they are unlikely to be used again, so I made a rough chisel out of a scrap of steel and case-hardened the business end. Note that the bevel of the chisel must be on the inside of the curve. Figure 9 shows the cut made by the chisel and this provides enough guidance for a fairly narrow chisel to remove the waste wood.

Figure 8: Chisel for cutting curves.

Figure 9: Chisel for cutting curves.

Figure 10 shows the rebates finished ready to receive the corners, but first the wood must be stained and varnished to a similar finish to the rest of the case,  Originally, the corners were pinned into place with narrow brass pins whose heads were then filed off flush, but a modern gap-filling adhesive such as a two-part epoxy like Araldite makes this unnecessary. Note that the corners of the rebates must be bevelled so as to allow room for any excess solder that might otherwise prevent a snug fit. Minor fitting errors can be concealed with mahogany paste.

Figure 9: Corner ready to be fitted.

Figure 10: Corner ready to be fitted.

Figure 11 shows a completed lid. Fitting of the hinges is routine, but getting them so that the lid sits squarely is not. The rebates in the lid for the hinges can be marked out from the rebates on the glazed top and new hinges fitted to the lid. It is as well to plug any holes in the glazed top with hardwood pegs if they are likely to be anywhere that new screws are going to go. If you then roll out a thin layer of Plasticine and cut slices to place where the new hinges will go, position the lid correctly and press, the screw holes will be marked out on the Plasticine. You can then drill pilot holes through the Plasticine into the wood in the correct places and get the lid to fit correctly first time. All is not lost if the lid ends up askew. Simply drill out the screw holes in the glazed top to about 5mm and fill the holes with epoxy putty. When dry, it can be drilled for pilot holes without any danger of the screw following the grain of the wood. If you reserve the last coat of varnish until the corners are fitted, you can give the corners a coat that will keep them shiny for a good few years.

Figure 10: Finished lid in place.

Figure 11: Finished lid in place.

The Mariner’s Chronometer is for sale through www.amazon.com and its European equivalents such as www.amazon.co.uk and www.amazon.co.fr ;and this is probably the cheapest way to buy it. I have been greatly encouraged by the  reviews left by buyers. I have been making minor modifications to the book as I find better illustrations, especially in the historical section.  If you have even a passing interest in marine chronometers, I think you will find the book is well worth buying.





7. Repivoting , part 2 : an escape wheel arbor.

22 07 2013

A few weeks ago, I received a Russian MX6 chronometer that had not travelled well in a journey half-way around the world. Despite having been packed with care, partly wound and with the balance wedged, it arrived with the upper balance and the upper escape wheel pivots broken. As I had a spare balance staff by me, replacing it was without problems (see post 5: Replacing a balance staff), but re-pivoting is for me always a little nerve-racking and I am in good company as one authority on repairing has remarked that “…replacing a pivot is never a pleasure.” There is more than one way to skin a cat and in the preceding post I described how to replace a pivot by drilling down the existing arbor and inserting a new piece. The drill used is necessarily slender and if it breaks off in the hole it is practically impossible to remove or to turn away, so the surrounding arbor has to be turned away, shortening it in the process. In this post, I describe a method used by clockmakers of old to replace worn pivots, but scaled down for the much finer chronometer pivots.

In this process, a new piece is fitted over the old arbor using what was picturesquely called a “muff”. It is rather difficult to illustrate photographically and I have shown it diagrammatically in Figure 1.

Pivots 2 drwg

Figure 1: Fitting a muff.

A in Figure 1 shows the broken pivot which in B has been reduced in size by turning, for about one and a half diameters. The MX6 arbor is about 1mm in diameter and I turned it down to 0.7 mm using my improvised Jacot tool as shown in Figure  2 of post 6. Turning is made much easier by annealing the end of the arbor by heating it beyond blue to a black heat and letting it cool down, since the hardness of the arbor itself is no longer important. The heat is prevented from travelling down to the pinion by holding the shaft in a small toolmakers clamp or a crocodile clip. The muff shown in C must now be made from pivot wire or silver steel and although the latter is usually supplied annealed, it is best to make sure of it by annealing, as occasionally hard batches are encountered.

Holding the steel in a collet chuck, it is centred, in my case by using a 1 mm carbide drill which has been sharpened by the 4-facet method and is therefore self-centring. This is followed by a 0.7 mm drill to a depth of a little more than the turned down length of arbor, taking great care to back out the drill frequently, so that swarf does not accumulate in the hole and jam the drill. If it does break off, it is not a disaster of the same order as if it were to break off in an arbor, but tiny drills of this size are not cheap. The outside of the muff and the pivot is then rough turned to, say, 0.2mm oversize for the body and 0.1 mm for the pivot, before cutting off perhaps 0.5 mm over-length, reversing in the chuck and facing off to exact length by repeatedly trying it on the arbor and measuring the over-all length using a micrometer with care.

Drills seldom drill exactly down the centre line of a work piece, so to ensure that the hole and the finished outside are concentric, it is necessary to make a tiny mandrel, turned down at its end to the same diameter as the arbor and the hole, and to fit the rough-turned muff to the mandrel without removing the latter from the chuck, unless you can be sure that your chuck is accurate to very close limits. On this occasion, I heated the end of the mandrel with a soldering iron and applied a flake of shellac until there was sufficient heat to melt it, at which point I fitted the muff to the mandrel and allowed everything to cool down. This takes us to point D in Figure 1 and is shown in Figure 2, when the muff can be turned down to its finished size and the pivot burnished. Burnishing is often described as a process that both smooths and work-hardens the surface of the pivot, but with the sort of pressures that can safely be applied, it is unlikely that any work-hardening ever takes place, so the finished muff must be heat-treated to increase its hardness without too much reducing its toughness (or increasing its brittleness, which amounts to much the same thing). Since the muff, being close to the chuck, is well supported, it can be burnished without the use of Jacot tool and it is helpful to hold the burnisher under the pivot so that progress can be monitored more easily until the pivot appears to have an even polish. Application of the soldering iron then releases the pivot from the mandrel.

Figure 2: Finish turning muff.

Figure 2: Finish turning muff.

As the pivot is now only 0.2 mm in diameter, heating it directly in a flame to harden it may cause it to flare up to white heat and disappear, so I buried mine in a little pile of case-hardening compound (Kasenit) on a fire brick and heated it slowly until it melted all around my muff, when I brought it to red heat and decanted the bleb of compound into cold water. The compound protects the steel from oxidation and in this instance also refines the grain structure near the surface, reducing the likelihood of cracking in service. It is now very hard, but brittle, so must be tempered to increase its toughness. The easiest (and safest) way is to use a domestic oven turned up to 260 degrees Celsius (500 F). I polished the end of the mandrel to witness the tempering colours and fitted the muff to the end as a convenient way of not loosing a tiny piece of steel barely 2 mm long and 1 mm in diameter. As a precaution, I also monitored the temperature with a thermocouple thermometer. After tempering for 30 minutes and allowing the parts to cool, I then cemented the finished muff to the arbor, again using shellac. I have used Locktite in the past, but to make it release its grip it needs to be heated to a much higher temperature than shellac, which melts at about 140 Celsius, so any heat treatment of the adjacent metal is put at less risk with shellac.

Figure 3 shows the finished product. The tempering colour is just visible, a dark brown verging on purple, so it is harder than ordinary blue pinion wire and than the deep wine colour recommended by Marvin Whitney in his “The Ships Chronometer”, but as it shows the degree of toughness and hardness used in the past for punches and reamers, it should stand up well to service in the chronometer. I have left the tempering colours on the body of the muff as a witness to the repair for any future servicer of the instrument.

Figure 3: Hardened and tempered muff fitted to escape wheel arbor

Figure 3: Hardened and tempered muff fitted to escape wheel arbor

Although the muff appears to be of larger diameter than the tapered rest of the arbor, it is in fact 1.01 mm in diameter, and the escape wheel boss fitted over it without problems. It remained only for it to be fitted to the chronometer. It was a trifle over length, so I reduced it by about 0.01 mm using a diamond lap so that there is now barely discernible end play in a freely running wheel and the chronometer is performing as it should.