2024-11-19

Last modified: 2024-11-19 23:05:44

Encoder UI

I want the "tare" button to tare the time axis as well.

And I want something that can show me the rate of change of amplitude, or some other way to get an objective measure of the friction on the balance wheel when it's swinging freely.

We're losing about 4 deg/sec when the pivots are properly adjusted, and much worse if they are too tight.

I am still seeing the thing where odd-numbered ticks and even-numbered ticks have drastically different periods.

if I make it only update the period when crossing zero in one direction, then it works. But I don't see why, isn't it summing both directions' zero crossings just the same regardless of which order they came in?

Oh! The issue is that the positive period is gradually increasing and the negative one is gradually decreasing??

Is the issue that the tare isn't being applied before we detect zero-crossings?

The code looks fine to me. And the amplitudes are still pretty much symmetrical.

I'm going to twist the slotted disc slightly and see if that changes anything, maybe it is just to do with how the slotted disc is aligned? The 0 point is actually offset to one side slightly?

Lol, that was totally it. Now I have both the positive and negative one increasing, although the positive one is consistently slightly higher. I guess trial and error to randomly rotate the disc into a good position.

I have it down to about 1ms of "beat error" now, good enough for me.

Wow! Check this out:

At about t=240s I breathed on the balance spring!!

Heating up the balance spring makes it slightly weaker so it doesn't pull the balance around as fast so the period goes up.