NavList:

Robin Stuart, you wrote:
"However my impression was that it was a bit slower than the Greenwich version. At Greenwich there appears to be something close to an initial free fall followed by some braking."

Reading a little further, the Greenwich system is fairly simple. The ball is attached to a metal staff. When the ball is dropped, the staff drops into a fluted opening to an air cylinder. The cylinder is designed with a small escape valve at the bottom. So it drops in, bounces on the compressed air, then sinks in the rest of the way to a gentle landing as the air escapes the cylinder. Something like that...

"The Port Chalmers time ball seemed to descend at a pretty uniform rate but, again, I wasn't really paying particular attention to that aspect of it."

Maybe braking by eddy currents? That's how I would build it (mostly because eddy currents are cool, but I suppose I could come up with a more practical reason when applying for funding). 

As long as some changeis visible from a couple of miles and potentially timed to the nearest quarter of a second, that should have satisfied any historical need presumably. So if a ball moves by, let's say, one minute of arc as seen from the most distant location in the extended harbor area (one mile range maybe) and covers that angular distance in one second or so, then that should be enough??

For anyone reading along, why a fraction of a second? In celestial navigation, an error of four seconds in the time yields an error of only one minute in the longitude which is generally regarded as good enough. A fraction of a second sounds like overkill. But the point of these observations would be to "rate" the chronometer in a circumstance when it's not possible to take it onshore for a few days visit with a proper rating service.

If you're in port on Friday the 16th, and the chronometer is fast by 37.5 seconds while on Thursday the 22nd it's fast by 39.0 seconds, the implied rate is 0.25 seconds per day ...(39-37.5)/6. A quarter of a second is nothing, but if the rate is a quarter of second per day when you previously believed the rate was a third of a second per day, that difference would add up to a mile error in longitude after seven weeks --enough to worry about.

Frank Reed