An Elgin 303, and Glue

This watch seemed like it would be straight forward, but the balance wheel did not move freely. The reason was not at first obvious.

On very close inspection, I found that the one-piece double roller was not straight. It was seated at a very slight angle. On removing the hairspring and the roller, I found that the roller table had been glued on, and a dot of glue to one side kept it from fully seating on that side. It was then too low, and touched the pallet fork, just in one spot. It was enough to make the balance very sluggish, but somewhat run.

The contact was so slight that it was not perceptible when watching the balance. I removed the glue and properly seated the roller table. Everything then worked fine.

This and the somewhat distorted hairspring lead me to believe that someone has changed the balance staff on this watch, but perhaps did not have the right tools for the job.

There was an interesting problem here. The dial was offset for some reason, just enough that the base of the second hand and the hour hand can touch the inside of their respective dial holes - and stop the watch.

The usual technique for "bumping" the dial over won't work here because the dial sits inside the rim of the base plate. This calls for a creative problem solving... It's always something.

This watch is an Elgin grade 303, 12 size, 7 jewels, made about 1919. 

Elgin Grade 287

Here we can see the pivots on the balance staff, broken.

The pivots are very small needles of hardened steel, and are easily broken if the watch gets a sharp jolt. Often the jewels that form the bearings of the balance also get cracked if something happens. But on this watch they are fine.

The old staff is removed in the lathe. The back hub is first cut down to almost nothing.

The replacement staff is riveted to the balance wheel.

This shows checking the balance, and the new staff, for free motion in the movement, prior to putting the hairspring back in place. In this case I had to slightly adjust the arms of the balance down, and then make sure it was true to round and flat again.

Then the roller table and hairspring are then installed and the balance assembly is complete.

The Separation of Static and Dynamic Beat

From The American Horologist magazine, July, 1944

The Separation of Static and Dynamic Beat
By Emanuel Seibel

To begin with, we are speaking of a good grade of watch in good mechanical condition.

What is good mechanical condition, you may ask.

First, it must be clean. Then we must have positive train freedom, but this freedom must be correct, not because pivots are sloppy but endshakes and sideshakes must be correct; these pivots must be well polished. Wheels must be absolutely round. Barrel arbor must fit as snugly as possible without bind. Center wheel must be upright; all wheel teeth straight without burs; pinions polished so that all frictions are reduced to a minimum.

Then your escapement must be set up as light as possible with safety.

Drops equal, lock and slide as light as you can make it, draw positive.

Roller jewel properly fit to fork slot and your guard sure, and the closer you can make pallet and balance shakes alike, the better you are off.

Now we are ready for poise. Most watchmakers poise to a square that is four points,-some to eight points or a double square. First, the balance must be absolutely true, round and flat. Before we attempt for poise, if the watch has the original staff in it, it should not require much, if any, poising; if a staff is being replaced, and old staff is turned out of wheel and properly staked fast, it should not need any, or very little, but poised it must be.

When you are satisfied it is as close to poised as it can be ,gotten, you have your static poise, or poise in rest.

Now, you put on your hairspring and your static poise is disturbed because your collet, to begin with, is not poised in itself and cannot be, for it has a slot in it, and if it is not a proper fit to Hairspring shoulder and is forced on staff, it is spread and the condition is aggravated. 

The spring, which is spiral, cannot be poised. Remember, the wheel itself is in poise at rest, but the complete assembly of wheel and spring has disturbed this static poise and we must get this assembly in dynamic poise (that is, in poise under power, or in motion). This is where the Watchmaster comes in.

Now, remember, the Watchmaster will not do it, and cannot do it. It can only indicate the trouble; and how you may correct it, is in your ability to interpret the readings of your Watchmaster and make the indicated adjustments.

Your balance under motion to function properly must have a turn and a quarter of vibration and we find that unless the hairspring is properly centered, positively true in round and flat and the pinning point correct so that it developes correctly, and breathes as it were, equally in all directions, you have unequal pushes and pulls in the spring, which disturb the rate and beat of your watch.

Just because your fork and roller are absolutely in line and in the absolute middle of the bankings, does not indicate that it is in beat, for the relationship of your pallets and roller may look correct in static condition, but when power is applied, your balance may pull to one side or the other and not start, and until you get this relationship of pallets to roller correct, it will not pull equally to either side and start. When this is correct, a watch in good mechanical condition will start off at the first twist of the crown. 

After a balance has been poised on a poising tool, equally weighted at every degree of the circle, balancing perfectly in all positions around the axis that balance is in "Static Poise" or in a state of equilibrium. (See Page No.6 in the Book. "12 Thousand Hours").

Counter poise, in the "Static Sphere" means you actually take off or add weight.

The average watchmaker will use four corners, or one square when poising a balance. The more critical workman will use eight corners or two squares. (See Drawing at Top).

It is true that often in poising a balance you start with the screws that are too far in to the center, and when you do this the balance will be under-compensated; naturally if you start with the screws that are too far away from center the balance will then be over-compensated. You must have compensation in "Static poise" or balance that will match as near as possible, the ability of hairspring in "Dynamic Poise", after you have this combination you never destroy static poise by moving or changing screws. 

After hairspring and collet have been added, and placed in watch, we then have the "Moving Element", when this, takes motion we are then concerned with "Dynamic poise".
(See Page No. 6 in the Book. "12 Thousand Hours").

Counter poise, in the "Dynamic Sphere" means, control of equilibrium in motion, and the control of this force is attained by changing the level of hairspring and push and pull under dynamic power. (See Page No.7 in the Book. "12 Thousand Hours").

Take a look at simple drawing at top. First we use the "Static Poise" and accept it as being correct. Then we check the watch on the Watchmaster. In motion we see the rate is good in all positions except P. R. & P. 1. Records show P. R. 5 Sec. slow and P. L. 5 Sec. fast. We instantly know that the "Push & Pull" of hairspring are not equal at these positions, where the "Push" develops towards the rim more than the "Pull", at that position the watch will run slow. Naturally at opposite position the watch will run fast. You simply cause the hairspring to have same ability in both positions. You do this with "Dynamic" counter-poise and NOT by changing screws and distorting your already correct 'static' counterpoise. 

Information Please!

From The American Horologist magazine, July, 1944

Information Please
Directed by 
W. H. Samelius, Chairman 
Science of Horology and Technical Advisory Board

HRG: I have an Illinois 60 hour Bunn Special watch. Please advise if this watch should be wound every 24 or 48 hours?

Ans.- In all watches the maianspring power delivered varies throughout the running period of the watch. By designing the watch to run 60 hours, the variation of power during the first 24 hour period is less than the variation of power during the latter 24 hours so for that reason, as the mainspring delivers more constant power for the first 24 hour period, it is advisable to wind the watch each 24 hours. If you accidentally omit winding the watch after the first 24 hours, it naturally will keep running until the end of its running time, but the time piece is very likely to take a different rate during the latter part of the run, therefore, we repeat, to obtain a constant rate, it is best to wind your watch each 24 

JIT: When a watch is marked adjusted to heat and cold, 5 positions and isochronism, what is meant by isochronism?

Ans.- The word "isochronism" is derived from the Greek word meaning "equal time." As applied to a watch or timekeeping instrument, the term "Isochronism" means that the time or rate would be constant throughout the 24 hour period, that is, it will have a constant gain or a constant loss for each hour of the day. Theoretically, a balance and hairspring should oscillate at the same rate whether the arc of motion is low or high, meaning that the balance should return to center in the same length of time, regardless if the arcs are unequal during the 24 hours due to the variation of power. The resistance of the hairspring must be equal to the force of the balance. As all hairsprings are not isochronal; it is some. times necessary to test several hairsprings before one is found that will prove itself isochronal throughout the range of balance arc. The length of the hairspring has a great deal to do when striving for isochronism.

HMU: A friend informs me he always puts oil on a roller jewel? 'Is this good practice?

Answer - Oiling the roller jewel is bad practice as the oil soon turns black or will thicken or gather dust, which will soon have effects on the watch. If oiling the roller jewel was the proper thing to do, I am sure all watch factories would adopt that method, however, up to date, you will find all watches as delivered by the manufacturers, are left with the roller jewel dry.

YKG: When inserting teeth m clock wheels, is it advisable to use hard solder?

Answer - When inserting a tooth in a clock wheel, a dove tail slot is cut into the rim, a wedge filed to fit the slot and held fast by spreading or riveting the wedge to place. If hard solder· is used, the heat would take the hardness away from the teeth of the wheels when they would very likely bend over and cause trouble when the wheel is put into service. It is even advisable to avoid using soft solder when inserting teeth into clock wheels.

In the sixteenth and seventeenth centuries, an HOUR glass was standard equipment in the church, so the sermon wouldn't last too long. 

Waltham P.S. Bartlett

This is a Waltham 18 size, 15 jewels, P.S. Bartlett model. Key wind and key set.

One of the banking pin screws was broken. Again, I was able to find one (exactly one!). This is lucky. These parts are only found on the earliest Waltham 18 size movements. They are unusual for banking pins in that there is a wide head, rather than having threads that go all the way out to the top. On other watches, like a comparable Elgin for example, a variety of banking pins would work and look fine. On this movement, it has to be this one.

This movement key winds from the back. Here are some images of the mainspring ratchet on the dial side.

I think I have replaced one regulator pin on every one of these I have serviced. I have no idea why it would be so frequently missing.

This view slows the regulator, in place.

Replacing the roller jewel on one of these is a specially challenging due to the hairspring being directly above, on the same side. So I didn't take any photos of the process... It's ticking now though.

How I Turn a Staff

From The American Horologist and Jeweler magazine, January, 1946

How I Turn a Staff

Being in the service and away from material houses for long periods of time, I have to make most of my own material. The scarcity of material in the wholesale houses may make my experience worth while to the average, watchmaker at the bench.

Lacking a suitable set of micrometers to measure the length of the various seats on the staff, I have found the following very useful and quite accurate. 

I have an ordinary micrometer of the one inch outside diameter type, but this cannot get in the close quarters to measure all the seats 'on a staff while it is in a lathe. I use the following system:

After removing the old staffs, I measure diameter. (a) and reduce a piece of wire to this diameter. This can be done with the mike.

Then I use the balance wheel itself as a guage to turn the balance seat to its exact diameter. Next, I lay the old staff down along side the new staff in the lathe so the end of the balance seat just touches the large flange of the new staff. Now the flange on the old staff marks the exact spot to make the undercut for the balance seat of the new staff.

After cutting this and turning the new hairspring seat to the approximate size. I measure the old staff hairspring seat (measurement c) and turn the new hairspring seat to exact size. The diameter of (d) is not critical but it should be reduced below the diameter of (c) and polished; the exact length of (c) and (d) may be determined by the same type of measurement as is shown in drawing (b) by laying the end of each so as to butt against the next larger diameter and the end of the new hairspring seat; for example will be indicated by the position of the undercut part of the balance seat of the old staff. The shape of the pivots should, be duplicates of the original staff and maybe fitted so as to just enter the hole jewel in the balance cock (with endstone removed) without freedom, then polished, which will give them the necessary freedom. The jewel should tip about 15 degrees to each side after polishing and be free in the jewel with a little of the pivot projecting through; next turn the back side of the flange and the approximate shape of the roller seat; then cut off and, reverse in a new chunk. Turn the seat to the required taper and turn down so the roller goes on till the space between the top of the roller and the bottom of the flange is about equal to the thickness of the jewel pin roller table. This gives about the right tension to the roller after staking on. Finally turn and grind the lower pivot to the same jewel fitting as the top pivot and polish carefully. By using the old staff in the same manner as is shown in, illustration (b), the new roller seat may be gotten to the exact length of the old. I usually leave the pivots a trifle long and then cut to exact fit after riveting on the balance wheel. I find it best to cut the top pivot to the exact "length and leave the bottom pivot a trifle long as roller seat is difficult to chuck in the ordinary chuck. 

Elgin Grade 455

This watch has a broken balance staff. I have posted many photos of parts of the process of changing a staff, so here we'll just look at a one or two things that aren't common.

For example, this is a two piece double roller, shown here removed from the balance assembly. These take extra care to recognize and deal with. The "guard" part is small, not to mention remarkably well finished and polished.

For a replacement staff, it's good to check the endshake by testing the lean of the staff in the balance jewel. This will do, shown after trimming down the pivots some.

Here is the replacement staff, good pivots, riveted to the balance wheel.

This double roller staff has a very large hub where the main roller table goes. Instead of using the stack with the notch for the roller jewel to press the part on, a stump that gives more space has to be used, and the job is done upside down from the usual.

The second, smaller, part of the roller can be put on normally though.

This movement is Elgin's grade 455, 16 size, 19 jewels, made about 1922. 

It is a high quality product, with a micro-regulator and some nice extra touches to the finish.