If a watch runs well—it is clean, free of flaws, and running with good amplitude—it can usually be adjusted to be more accurate across different positions. And then we have the great watches—the ones designed with precision and adjusting in mind. When those watches are in good shape, you can achieve adjusting excellence.

The Illinois Bunn Special, an iconic pocket watch, was designed with precision in mind. In this post, we’ll show how a century-old Bunn Special can outperform most of the cheap mechanical movements found in new budget watches and attain essentially perfect accuracy in all 6 positions.

If you have been reading and working your way through our “dynamic poising how-to” series that teaches the basics of adjusting watches to positions, this post will emphasize a couple points about the process of dynamic poising that are worth highlighting. We have seen these points before, scattered across many posts, but at this point they bear repeating and are worth reinforcing.

First, A Few Repairs

A while back I picked up a broken Illinois Bunn Special pocket watch that the seller had tried to fix but couldn’t figure out. Vexing Illinois watches with weird problems are my specialty, so I got it for cheap. It had one obvious fault and two subtle ones.

The obvious problem was a shattered lower balance hole jewel. That will stop a watch but is easy to fix.

One subtle problem was a distorted hairspring. You can diagnose hairspring issues more easily by removing it from the wheel and placing it in the stud holder. It looks fine from above: perfectly centered over the jewel hole.

From the side, though, you see the problem. The tilt caused the hairspring to rub against the balance wheel’s arm, which is a major fault.

What happened, I think, is that someone removed the balance bridge too quickly before coaxing the balance wheel from beneath the center wheel. The hairspring was distorted near the stud, causing it to rise too high. When the stud is installed, the hairspring is pushed out of flat and against the balance wheel’s arms. Hairsprings are my nemesis, but I have been reading and practicing, and I was able to reshape the spring back to normal.

And the third fault was a bent pallet guard pin. I’m not sure how this happened, but it was probably from the same inattentive, fast removal of the balance. The double roller probably caught the underside of the pin and bent it.

Watchmaking requires a great deal of mindful attention, without a doubt.

After the repairs, the watch was cleaned, fitted with a white alloy mainspring, demagnetized, and left to run for a couple days. The amplitude was excellent, and the timing machine traces were as straight and steady as west Kansas train tracks.

Now let’s adjust it while highlighting a few finer points of the dynamic poising process.

1. Reset the Regulator and Mean-Time Screws

Before adjusting, hit the reset button on the regulator and the mean time screws. For the regulator, shift it to the center of the slow/fast index. This gives you room to move, should you need to make the watch a bit faster or slower, but it is also an aesthetic point. One sign of superior adjusting is a highly accurate watch with a centered regulator arm.

For the mean-time screws, reset them to ensure they are equally distant from the balance wheel’s rim. Screw them all the way in and then unscrew them perhaps 1 or 2 full turns. Over the years, many watchmakers, deliberately or not, have set one mean-time screw a bit closer than the other. This isn’t necessarily wrong, but adjusting will be easier if you know what you’re dealing with and have set the screws equally.

2. Correct Big Poise Errors by Adding Weight

When a movement has a big poise error, correct it by adding weight. It’s easier, faster, and reversible if you miscalculated. Just as woodworkers and sculptors “rough out” a piece before the fine finishing, you can “rough out” the poise by adding weight and then fine-tune the movement by removing tiny amounts.

3. Be Strategic with the Mean-Time Screws and Regulator

Mean-time screws are invaluable for precision adjusting because they allow you to make the watch run much faster or much slower when you need it to. Getting excellent results requires being strategic with using the mean-time screws and regulator—knowing when to shift them to suit the adjusting process.

Let’s illustrate these three points with our Bunn Special.

First, I set the regulator to the center and reset the mean time screws to an identical distance before recording the initial rates. Here’s what we had for our Time 0, pre-adjusting rates.

These rates look okay. The dial-up (DU) and dial-down (DD) rates are the same, which is always a great start. Overall, though, there are big poise errors here. The biggest difference in the 4 vertical positions, for example, is 71 seconds (PU and PR).

Big poise errors should get corrected by adding weight. The movement is currently running fast, so adding weight is ideal—it will help all six positions converge toward zero. But what if it were running slow? In that case, we would be strategic with the mean time screws—by screwing them in, we would make the watch run fast so we could add weight.

We won’t go step-by-step. Having read the prior posts, you know that we identified the heavy spot by running the watch at low amplitude. After doing so, we added a heavy timing washer to the light spot.

Here’s what happened after adding weight. the poise is much better—the biggest vertical difference shrank from 71 seconds to 44 seconds—but there is still a long way to go.

Once again, a big poise error is better handled by adding weight. But because the movement is now running slow DU—an effect of adding a big timing washer, of course—adding weight would cause the horizontal (DU, DD) and vertical (PU, PL, PD, PR) rates to diverge, which would be bad.

Here’s where we see the need to be strategic with our mean-time screws. By threading them inward around a half-turn, I can set the watch to run fast overall. As a result, adding weight would help with poise while also bringing the horizontal and vertical rates closer.

After another timing washer, we can see that we’re really getting somewhere. The largest difference between the vertical positions is only 17 seconds. Frankly, this watch would keep great time in use as it is.

Nevertheless, we can get a high-grade watch like this even closer.

For fine-tuning, you can more easily handle tiny poise errors by removing weight, ideally in a way that is inobtrusive. In this case, one spot is slightly heavier, so removing a tiny amount of weight at that spot would improve poise. But the watch is running fast, so we need to slow it down.

For small changes to the overall rate, it’s better to use the regulator than the mean-time screws. Turning the mean-time screws an exactly equal amount is easy when it is a 1/4 or 1/2 turn, but much harder if it is 1/8 or 1/3 of a turn. If those dense mean-time screws aren’t turned identically, the wheel is thrown out of poise, thus undoing much of your hard work.

For this watch, I set the dial-up rate to -2 seconds by turning the regulator. I then removed a small amount of weight from the heavy spot with a screw slot file, a nice method for unobtrusive changes.

At this point, adjusting is a matter of making a tiny change, observing the result, and repeating the cycle until you get the result you want. I probably filed the same screw 3 or 4 times, slightly deepening the slot each time. After each time, I would set the watch to be very slightly slow, -1 or -2 seconds when dial-up, to ensure that all 6 rates were converging around zero.

And here’s where we ended up—perfection is not too shabby, if I do say so myself.

To keep some perspective, after some hours pass and the amplitude is different, the rates won’t be perfect. And if you watch the rates for a long time, they will occasionally blip up or down by a second or two.

But for a 20 minute period in February, 2021, this watch ran dead zero in all 6 positions, so I’ll claim it and own it. (Pics are at the end for the doubters and haters.)

Wrapping Up

In this post, we wanted to reinforce a few key points about the process of dynamic poising. In practice, it is best to first add weight to handle large poise errors and then remove small amounts for fine tuning. To accomplish this without causing a bad horizontal/vertical divergence, however, you will need to be strategic with the mean-time screws and the regulator by using them to control the global rate in a way that affords doing the adjustment you want to make.

Clearly, by this point in the series, you should be convinced that you don’t need an old poising tool to achieve excellence in watch adjusting. To discern really tiny poise errors, you need a timing machine. Spinning the balance wheel in the jaws of a poising tool simply can’t magnify the tiny poise errors like a timing machine can. If anything, this only reinforces our respect for the tremendous skill of the watch adjusters who worked on these watches back in the day.

Finally, I should probably quit adjusting old watches, shut down the blog, and find a new hobby. There’s nowhere to go but down now.

Some Pics

Because anyone on the Internet can say anything on a blog, here’s some documentary evidence of this watch’s rates. They were measured using a 20-second moving window and a lift angle of 50 degrees, which seems about right for how this watch’s escapement was set up.

If you’re familiar with timing machines, you can see how, if watched over the course of a few minutes, the rate would occasionally pop up or down by a second or so, based on the minor inflections of the timing traces. These are smoothed out with a longer moving window.

DIAL UP

DIAL DOWN

PENDANT UP

PENDANT LEFT

PENDANT RIGHT

PENDANT DOWN