Re: Magnetic vs gravity impulsing
toggle quoted messageShow quoted text
This may sound a bit "of the wall", but I have found that a good way to ensure that the impulse occurs in the centre of swing is to take the bob off! With just the rod, release the gravity arm while holding the rod to the left, then gradually move the rod to the right and you will feel the impulse (and it may take the rod out of your hand). I'm sure that with a bit of ingenuity, it should be possible to measure that point, which is much easier to locate without the bob.
The setting-up instructions are a good guide, but variations in manufacturing tolerance mean that they won't always put the impulse exactly in the right place.
Just a thought.
From: John Haine <john.haine@...>
To: synchronome1 <firstname.lastname@example.org>
Sent: Wed, 15 Jan 2020 11:24
Subject: Re: [synchronomeelectricclock] Magnetic vs gravity impulsing
Bepi mentioned that from data I presented over in the "reinventing" thread that the impulse was occurring too early. That was a considerable surprise to me because I though that the impulse timing was fixed in the pallet and suspension geometry. I think I have fixed this by shifting the pendulum to the right relative to the pallet (which I can do by adjusting thumb nuts that clamp the upper block of the suspension on the threaded rod that fits on the brackets). The downside of this is that when the pendulum is stationary the roller isn't half way down the slope of the pallet as I would expect but on the top of the slope just before it starts to roll down. This means that the theory of the slope profile (based on F H_J's writing) is wrong, or my implementation of it is - this needs more investigating. It also means that my "auto-start" routine not longer works so I have to get the pendulum swinging by hand, not so much of a problem really! To adjust the position I added some timing measurements to the Arduino code, it takes quite a long time to do the setup since the clock takes several hours to settle after each adjustment, so I haven't repeated it after reassembly in its case.
Bepi's clock seemed to have a late impulse. As he says, if the impulse can be put at the centre, variations in impulse strength ("noise") have minimum effect, since they will only affect the timekeeping through amplitude changes and circular deviation.