The Count Mode

MicroSet has three methods that will allow you to find the correct rate of a running clock. One way is to measure the time it takes the minute hand to go around. Click here to see that procedure. The following description will show you how to count the correct number of Beats Per Hour from the swinging pendulum. This method can be done relatively quickly and is amazingly accurate. The third method is the Rate Finder, which is a refined application of the Count Mode described here.

A clock or watch movement is a gear train. There is an exact relationship between the ticks of the escape wheel and the motion of the hands. Movements are rated in terms of their Beats Per Hour. A given movement has a specific number of Beats Per Hour. It doesn't matter if the timepiece is running fast or slow; it always beats the same number of times to move the minute hand around once. If we could count the beats in one rotation of the minute hand we would know the correct number of Beats Per Hour of that movement.

It's not safe to count the sound of clock ticks for an hour. In that much time, some other noise is almost certain to occur and advance the count erroneously (especially in striking clocks!). But MicroSet has an optical detector which can count swings of the pendulum reliably for hours on end with no false signals. Each swing of the pendulum corresponds to one beat of the clock. If we can count the number of pendulum swings in one rotation of the minute hand we will have the correct number of Beats Per Hour of that movement. The same method can be used with the acoustic sensor on watches, which run quietly enough that they are not likely to get false counts.

MicroSet has a special mode for this operation, called the Count Mode. In the Count Mode, MicroSet does not measure beat times, it simply counts beats and displays the total on the LCD screen. To count Beats Per Hour, start by arranging the optical sensor on the pendulum of a movement so it will trigger on each beat (at the center of swing). For watches, place them in a quiet place and arrange the sensor to trigger consistently. With MicroSet running in the Count Mode, each beat will advance the counter. All we need now is a way to judge one rotation of the minute hand.

You can look closely at the minute hand as it moves around the dial. Start the count at some specific point, then stop the count when the hand returns to that point. This is not very precise because it's hard to judge the exact position of the minute hand. With watches you often have a second hand. By watching both the minute and second hands you can be quite accurate. You can dramatically increase the accuracy of this method by letting MicroSet count for several hours. For example, if you let a clock run for five hours, you can stop the count when the minute hand reaches the same point at which you started the count. Divide the total count by five and you'll have the average count for each hour. This is the number of Beats Per Hour of the movement.

If the movement is a striking clock there is a better way. The clock will strike at nearly the exact same point each hour. If you start the MicroSet counter as soon as you hear the strike begin, then stop it as soon as you hear the strike begin the next time (an hour later), you will be very close to the correct beat count of the movement. As before, you can increase your accuracy by letting the clock strike several times and then divide the total count by the number of hours that have elapsed.

To illustrate, I have a Seth Thomas 8-day mantle clock that is rated at 9777.86 Beats Per Hour. I happen to know the correct rate for this movement, and we can use this knowledge to see how close we come with the MicroSet Count Mode. I used the strike of this movement to start and stop the counter. Shown below are the counts over several hours. As you can see, the accuracy increases with each count. After three hours the value is within 11 seconds a day of the correct value.

Hour

Count

Average per Hour

Error

1 9770 9770.00 -70  sec/day
2 19561 9780.50 23  sec/day
3 29330 9776.67 -11 sec/day
4 39115 9779.00 8 sec/day
5 48893 9778.60 7 sec/day

To test the method on a pocket watch, I started the Count Mode as the minute and second hands passed a known point. I stopped the count exactly half an hour later (according to the watch), and it showed a count of 9,002. Clearly this watch has a beat time of 18,000 BPH. Out of curiosity I let the count run overnight. After 14 hours I stopped the count 252,002. If you divide this by 14, you get 18,000.14 BPH. The answer is no more accurate than the half hour reading, but it demonstrates that MicroSet was able to count 252,000 beat sounds without any false triggers. Often you can tell what the correct rate for a watch is by a simple measurement. If it's very near 18000 BPH, the watch is probably 18000 BPH. But there are many watches with other rates, and the Count Mode can identify them.

I'll make one more point that's not obvious. The Count Mode works whether the clock is fast or slow, and it doesn't matter if the clock is REALLY fast. For this reason you can remove the bob from the pendulum and let the crutch "wig wag" at a greatly accelerated rate. Set the optical sensor to count swings of the crutch, and a "10 hour" count can be made in three or four hours. I performed this test on the same Seth Thomas used above, and ran several automatic "10 hour" tests with a speeding crutch. Here are the results:

Test

Count

Average per Hour

Error

 1  97781 9778.1 2 sec/day
 2  97773 9777.3 -5 sec/day
 3  97776 9777.6 -2 sec/day

Click here to see a description of the automated Rate Finder beat counter.