Jake Anderson wrote: > well the book longnitude has them accurate to a few seconds a month > on a heavily pitching ship (though captain cook (discoverer of > Australia) had 17 of them on board) > traditional grandfather type clocks are probbably on the order of a > second or 2 per day The key to the accuracy of the shipboard pendulum clocks is that two opposing pendulums are used to "cancel" the error. For a spring driven clock these were amazingly accurate, but took many years to develop. The main issue with spring driven clocks is that the tiny impetus provided to the pendulum (to keep it swinging) is greater when the spring is wound tight than when it has run down a bit, hence they tend to run faster when first wound and then gradually slow down over time even though the pendulum doesn't change length. There have been many improvements implemented, but the basic issue remains. Many attempts at increasing the accuracy by limiting the range of the spring have been applied, such as placing the spring in a "barrel" so that it can't possibly unwind beyond a limit, and using a mechanism to limit the amount of winding that could be done to help keep it in it's linear range. These methods work reasonably well, but do not fully eliminate the problem. Weight driven clocks (like the grandfather clock in your example) tend to maintain a constant impetus with each tick of the escape wheel. This means the clock ticks at the same speed regardless of how high the weights are, naturally leading to far greater accuracy. The main accuracy limiting factors now becomes temperature and air convection. The air convection problem is easily solved by enclosing the pendulum, but the temperature factor is a little tougher. Many old laboratory grade clocks used vials of mercury on the pendulum to compensate for temperature changes, some of these weighed more than 10 Kilograms. As temperature increased, the pendulum would grow longer (normally resulting in it slowing down), but the mercury would expand in the vials raising the center of gravity of the pendulum to compensate. These clocks were accurate to less than 2 seconds per month (better than 1ppm) rivalling all but the most modern of timekeeping devices. Other methods were employed as well by using bi-metallic properties, but the mercury approach seemed to be the natural winner. Weight driven clocks have a natural advantage over spring driven when it comes to overall accuracy. This led to the next accuracy problem, when winding the clock, the impetus force would be removed as the weights were raised. Attempts at solving this issue involve using a spring to keep applying the appropriate impetus force while winding. While not perfect, it still improved the accuracy and demonstrates the length at which clockmakers would go to improve their product. As you can all tell by now, I'm quite fascinated by horology and mechanical clocks. I have over 20 clocks, many of which are roughly 100 years old. They all work well and most keep very decent time, especially the weight driven clocks. Most fascinating are the various methods used to accomplish the same task due to the ongoing patent wars. Methods for "striking" the hour are varied and quite ingenious IMO. PIC Material: I have an ongoing project to measure the the beat of a clock optically (accoustically when optically is impracticle) in order to quickly set the pendulum length adjustment and the balance of the "beat" over the course of a couple of hours instead of over several days of watching and tweaking. Devices that do this are already available, but are quite expensive as is the way with this particular hobby. PIC chips are cheap. ;-) Horological Tips of the Day: Most pendulum clocks that don't want to run (barring a broken spring) are merely out of beat. By listening to the time space between the tick and tock sounds, you can "level" the clock until it sounds even. Never, never, NEVER spray WD40 on a clock mechanism. Use only a quality clock oil and use it very sparingly on the holes in the plates where the pivots ride. Never oil the wheel (gear) teeth directly. Surface tension of the oil can make the clock refuse to run as the forces at the far end of the gear train are very small. Always remove the pendulum and support the hanger mechanism and crutch rod by using a rubber band (to hold it in one position) before moving a clock. SAFETY WARNING: NEVER attempt to disassemble a spring driven clock mechanism without properly removing the tension of the spring. There is more than enough stored energy in an 8 day mechanism to break bones, sever fingers and launch parts with enough force to embed them into your body and eyes. All this is aside from the incredible damage that will likely result to the mechanism. michael > ----- Original Message ----- > From: "Dave Tweed" > To: > Sent: Tuesday, April 08, 2003 11:39 PM > Subject: Re: [PIC]: A metronome project > > >> Olin Lathrop wrote: >>>> actually a 1% difference can be pretty noticiblle >>>> though mainly in concert with other players but it could throw the >>>> "feel" of a piece off. >>> >>> I doubt the old fashioned spring/pendulum metronomes were that >>> accurate. >> >> Why? Pendulum clocks (same basic technology) are at least an order of >> magnitude more accurate than that (1 min/day or better). >> >> -- Dave Tweed >> >> -- >> http://www.piclist.com hint: The list server can filter out subtopics >> (like ads or off topics) for you. See http://www.piclist.com/#topics -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics