A mechanical phase locked loop!

And that is what it is, not at all unlike a PLL using a bang-bang phase detector.

I believe you are thinking of the Foucault pendulum. This had nothing to do with elliptical paths of pendulums. This was a pendulum free to swing along any axis. As the earth rotates the pendulum continues to swing in its original path and the earth turns beneath it. Of course the pendulum appears to rotate the plane of swing.

Yes, that is right. The change in frequency (phase change rate) is only momentary.

Hadn't heard of that one. At the BHI lecture there was mention of another correction of circular error by a colied spring attached somewhere at the bottom, but I wasn't paying full attention at that point.

There were also other means such as cycloidal cheeks around the suspension spring.

It has everything to do with the circular error and the variation in frequency that comes with varying amplitude of the swing.

... which is virtually the range where sin( theta) = theta.

You seem to be completely misunderstanding the operation of the Shortt clock. The slave pendulum has no need for correction of circular error. It is a good pendulum, but not a great one. It doesn't need to be great, it is corrected every 30 seconds by the electromechanical escapement of the master pendulum. It only has to be good enough to provide an appropriately timed release of the gravity lever.

So the small circular error has no bearing on the slave pendulum.

Exactly. This *is* the range where sin(theta) = theta. Anywhere other than zero it is an approximation.

I'm sorry, but you totally misunderstood what I was saying, which was that because all pendulums exhibit circular error, when the hit occurs in the hit and miss synchroniser and foreshortens the swing, then, for that half-cycle, and only that half cycle, the frequency is changed, as it must be.

Just as in the electronic PLL, instantaneous changes of phase have instantaneous changes of frequency, no matter how short lived, associated with them.

Wrong! It does NOT measure the relative phase, it makes NO measurement of the phase difference. All it does is detect if there is a phase lag of any degree. It could be a fraction of a degree or 180 degrees, the same correction is then applied regardless.

Wrong again it is open loop, there is no measurement, just the same adjustment regardless of the phase difference.

Jeff

Might be easier to define a set entitled "Locked Oscillators, of which the phase locked loop, injection locked and hit and miss synchronised are all members.

Are there other candidates ?...

Chris

From pre-war, the Goyder Lock?

Which raises an interesting point; before the 3-tier coffer-filling fiasco was the spawn of the RSCB, the candidature for the RAE tended to know all about the history of amateur radio before getting their licence, but now they seem to know sweet FA even after getting their licences, such as the difference between sideband and sidetone.

House!

Hadn't heard of that, so looked it up and found:

Which was an interesting read, but not enlightening.

Some of the early scope timebases, puckle, for example sounded interesting, but they were effectively injection lock, of course. I guess a triggered timebase is a variation of the hit and miss model.

Couldn't grok the relevance of the following paragraph above :-)...

Chris

What you say about frequency vs. phase is true and how the Shortt clock adjusts phase, but it has nothing to do with circular error of the pendulum. The correction of the phase is from the added spring resistance shortening the time as well as the travel of the pendulum. The fact that the swing is shorter and the second order circular error will create a tiny error in the timing is pretty much irrelevant. The real change is from the added spring constant changing the first order effect in the pendulum equation. The coefficient of the gravitational constant is effectively changed by the spring.

Is that more clear?

Jeff wrote on 8/6/2017 6:38 AM:

...and that is a measurement. It determines if the relative phase is plus or minus, a binary measurement. This is exactly the same as the measurement taken by a 1 bit ADC. Even though it is one bit it is still a measurement.

Totally wrong. The phase adjustment varies from a constant about to ZERO! Again it is a binary adjustment.

If there was a three level range of measurement and adjustment +, 0, -, would that be enough to constitute a measurement and adjustment so it becomes a PLL? If not, how many bits are required? If any number of bits can't do it are digital PLLs not PLLs?

It relates to the abysmal lack of technical acumen amongst those who are today's would-br radio amateurs, most of whom are really CBers-masquerading-as-radio-hams, identifiable by their M3 and M6 callsigns past and present.

You continue to misunderstand. Any pendulum swinging with circular error speeds up for shorter amplitude; speeding up means increased frequency. Therefore, for the half cycle inwhich there is a hit, a shorter amplitude and hence instantaneous higher frequency exists.

I understand perfectly and explained it for you in excruciating detail. The change in phase of the Shortt clock slave pendulum is due to the FIRST ORDER change in the effective gravitational constant in the pendulum equation by engaging the leaf spring. While the reduced amplitude of the swing *will* cause a SECOND ORDER effect in the motion of the pendulum, it will be MUCH SMALLER than the FIRST ORDER effect.

What part of this do you not understand or not agree with?

Nothing in any of rick's posts he does understand the above, or anything else. Plus, what you have posted is exactly what I explained to you earlier.

It is clear you are on the edge of resorting to your normal abuse.

It's not that I do not understand nor disagree with you, it's that you're off on a complete tangent to what I was suggesting, and do not realise it.

No he isn't, you are not keeping up.

If he's not keeping up then he needs Viagra.