Improved Experimental Pendulum

[Joseph Noci 1Participant @josephnoci1]

Dave, I played with one of those relays, stripped as you have, here on the bench now. Applying 12 to 15v pulsed gives a large increase in the magnetic field and the relay is just warm to the touch with a 100ms pulse every second after running for 4 minutes.

SInce you can't get the E_magnet closer to the bob for starting, if you have a 12/15v supply around the electronics, use a relay to select between +15v for starting and then switch back to +5v when running. That should work OK…

Edited By Joseph Noci 1 on 23/04/2023 20:30:11

[david bennett 8Participant @davidbennett8]

Dave, it would be useful to try to separate clock noise from electrical noise. In your earlier clock, did you notice much  noise difference when running on battery power (as in power cut mode) ?

dave8

Edited By david bennett 8 on 23/04/2023 21:15:00

[SillyOldDufferModerator @sillyoldduffer]

Posted by Joseph Noci 1 on 23/04/2023 20:29:27:

…

SInce you can't get the E_magnet closer to the bob for starting, if you have a 12/15v supply around the electronics, use a relay to select between +15v for starting and then switch back to +5v when running. That should work OK…

…

Now that's a good idea! I've made a holder to physically hold the electromagnet closer but been busy fixing mum#s plumbing to try it, maybe this evening/ If it doesn't work, I can up the volts instead.

Ta,

Dave

[SillyOldDufferModerator @sillyoldduffer]

Posted by david bennett 8 on 23/04/2023 20:50:09:

Dave, it would be useful to try to separate clock noise from electrical noise. In your earlier clock, did you notice much noise difference when running on battery power (as in power cut mode) ?

dave8

I ran it for about an hour on a battery and got much the same standard deviation. I have three main suspects:

1. Even if it was well-made, and it wasn't, the suspension required the clock to be level, which was difficult to do. So the pendulum was likely swinging on a twisted spring – not good, because it means the bob can;'t fly in a consistent straight line. The new version lets the pendulum naturally hang vertical, is better made, and the clock now has proper levelling screws.
2. The pendulum rod was flexible, not a good combination with a twisted suspension spring. I've upgraded to a stiffer one.
3. The beam break sensor is basic and my design made no attempt to create a sharp beam due to shortage of space. The sender and receiver are now mounted in long slotted tubes, which I hope the resulting constrained beam will make bob detection more consistent.

Electrical noise could contribute to false triggering, but I've not found any evidence yet. Lack of evidence may be because it was masked by problems 1,2 and 3 above. It's possible fixing them will reveal electrical noise is an issue at a lower level. The new circuitry already has some extra decoupling "just in case", but I still have to get the clock running before I can measure the pendulum again. Fingers crossed, the period will be much more consistent

Dave

[SillyOldDufferModerator @sillyoldduffer]

Posted by blowlamp on 23/04/2023 20:19:08:…

I haven't been following this topic closely, but I think the position of the electromagnet might be flawed.

It's a deliberate part of the experiment. In a mechanical clock where the escapement is usually at the top, it's definitely best to impulse when the bob is travelling at maximum speed because it minimises mechanical disturbance.

The same logic applies to magnetically impulsed pendula, so builders often often locate the coil directly under BDC. However, because a magnetic field doesn't deliver mechanical thump, I conjectured that the impulse can be applied at any point in the swing provided energy is added to the bob without disturbing it. The jury is still out – it's easier to impulse close to BDC, but I have run the clock successfully by impulsing close to end of swing. So much else wrong, with the earlier build though, that results are inconclusive.

Side mounting the electromagnet means it can also be used to start the pendulum swinging from a dead stop. This is useful when the pendulum can't be touched because it's in a vacuum! Self-starting and remote control also allows the clock to be physically located for best advantage – somewhere low vibration in the house where I won't trip over it! If the next version runs very well in a vacuum I might dig a safe hole for it in the back garden! Same hole might make a suitable grave if it runs badly…

Dave

[John HaineParticipant @johnhaine32865]

Posted by SillyOldDuffer on 24/04/2023 18:09:28:

Posted by blowlamp on 23/04/2023 20:19:08:…

 

……

The same logic applies to magnetically impulsed pendula, so builders often often locate the coil directly under BDC. However, because a magnetic field doesn't deliver mechanical thump, I conjectured that the impulse can be applied at any point in the swing provided energy is added to the bob without disturbing it. The jury is still out – it's easier to impulse close to BDC, but I have run the clock successfully by impulsing close to end of swing. So much else wrong, with the earlier build though, that results are inconclusive.

……

Dave

Basic oscillator theory says that the fractional change in period due to a phase difference of "phi" radians between BDC and the phase centre of the impulse is tan(phi)/2Q. Generally known as the "tangent rule" but actually one of "Airy's formulae". You can clearly see this in an occasionally impulsed clock such as a Synchronome if you look at the period of each cycle, it can be seen to have a significant error on each swing where the impulse happens (e.g. every 15 swings).

Edited By John Haine on 24/04/2023 20:06:36

[SillyOldDufferModerator @sillyoldduffer]

And I thought all I had to do was keep the amplitude constant….

[blowlampParticipant @blowlamp]

Dave.

It seems wrong to me because the magnet will cause the pendulum to accelerate within the deceleration phase of its arc.

However, if the bob was changed to be a permanent magnet and the electromagnet was polarised so as to repel the already (just) departing pendulum, then I can see how it might work, providing the on/off period is correct.

Martin.

[david bennett 8Participant @davidbennett8]

Posted by SillyOldDuffer on 24/04/2023 20:31:51:

And I thought all I had to do was keep the amplitude constant….

Dave (and John Haine), perhaps that phase change error would be seen by your clock as an amplitude error and be already compensated for?

dave8

[Michael GilliganParticipant @michaelgilligan61133]

Posted by SillyOldDuffer on 24/04/2023 20:31:51:

And I thought all I had to do was keep the amplitude constant….

.

At the risk of getting overly philosophical about it, Dave

… If you could do that, then the statement would be true

My understanding is that your system is working by a process of active adjustment, and therefore the amplitude cannot be kept constant.

MichaelG.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Michael Gilligan on 25/04/2023 00:39:50:

Posted by SillyOldDuffer on 24/04/2023 20:31:51:

And I thought all I had to do was keep the amplitude constant….

.

At the risk of getting overly philosophical about it, Dave

… If you could do that, then the statement would be true

My understanding is that your system is working by a process of active adjustment, and therefore the amplitude cannot be kept constant.

MichaelG.

The system is experimental, and I've tried both. The pendulum is simply that – apart from the suspension, there is no mechanical connection to it. It's position is detected by breaking an IR beam connected to a microcontroller, so what happens next is decided by a computer program. As I wrote the program, it can do almost anything, and since starting the project I've tried various timing combinations:

• Impulsing at top of swing.
• Impulsing at bottom of swing,
• Impulsing after every 'n' beats
• Impulsing when the amplitude falls below a trigger level
• Impulsing on every beat

I've also tried various impulse lengths – anything from 8uS up, so the impulse can be set from almost imperceptible to completely OTT.

By firing the impulse on every beat, it's possible to adjust the impulse such that amplitude doesn't vary much. Most constant by over-impulsing, but this disturbs period. Manually reducing impulse to slightly more than needed to keep the bob swinging produced a fairly constant period and amplitude. As both exhibited noisy random jumps the oscillator was a bit unstable. Not awful, but clearly in need of improvement.

My diagnosis was of mechanical issues in the design and build of the pendulum, notably a whippy rod, a badly made poorly designed spring suspension, and no sensible way of levelling the frame. Also, a strong possibility that the IR beam was too broad. The latest build addresses these issues, but I made a mistake with with the electromagnet such that the pendulum can't be started swinging by the microcontroller. There's a chance I'll fix that today, fingers crossed.

Once the pendulum and software are running, I shall revisit the earlier experiments. Indicative rather than conclusive because results were flawed by mechanical and software development problems. Now I'm more confident of the build and the software, it will be interesting to see how well the improved pendulum performs. I'm expecting better, which is why this version of the design has tackled vacuum containment seriously. The earlier version never performed well enough to be worth pumping out. Fingers crossed, this one will be!

All this is quite confusing, because a forum thread isn't a good way of documenting a skittish development programme with deliberate experimental features. The experimental aspect means I've changed tack several times when others pointed out flaws and improvements. Also confusing!

I'm doing a Zoom presentation to the SMEE Digital Group on the 13th May, which I hope will clarify the 'big picture'. Well worth joining SMEE just to see what Joe Noci thinks of my electronics and computing; hearing John Haine's critique of the theory; and enjoying Duncan Webster's no doubt pungent remarks on my approach to statistics! And there will be several other well-qualified engineers present. Should be fun.

Dave

[John HaineParticipant @johnhaine32865]

Posted by blowlamp on 24/04/2023 21:22:13:

Dave.

It seems wrong to me because the magnet will cause the pendulum to accelerate within the deceleration phase of its arc.

However, if the bob was changed to be a permanent magnet and the electromagnet was polarised so as to repel the already (just) departing pendulum, then I can see how it might work, providing the on/off period is correct.

 

Martin.

The pendulum decelerates for two reasons – it's climbing against gravity, and losing energy. If the attractive force just balances the friction force (or rather the energy given equals the energy lost), the pendulum won't know the difference except that if the "centre of gravity" of the impulse is not at BDC the timing will be affected. If the magnet was repelling it's exactly the same principle, except that you don't want permanent magnets anywhere near a steel bob or iron base as the magnetic force will affect the effective gravity.

I'd love to see Dave's presentation but alas it clashes with the AHS AGM where they have some talks on the Big Ben restoration.

Edited By John Haine on 25/04/2023 11:09:16

[Joseph Noci 1Participant @josephnoci1]

Posted by SillyOldDuffer on 25/04/2023 10:59:48:

By firing the impulse on every beat, it's possible to adjust the impulse such that amplitude doesn't vary much. Most constant by over-impulsing, but this disturbs period. Manually reducing impulse to slightly more than needed to keep the bob swinging produced a fairly constant period and amplitude.

I don't understand that…If amplitude is constant, so must period be..?

I'm doing a Zoom presentation to the SMEE Digital Group on the 13th May, which I hope will clarify the 'big picture'. Well worth joining SMEE just to see what Joe Noci thinks of my electronics and computing; hearing John Haine's critique of the theory; and enjoying Duncan Webster's no doubt pungent remarks on my approach to statistics! And there will be several other well-qualified engineers present. Should be fun.

Mm – Think I will keep my mouth shut on that – make a fool of myself to easily…

Dave

Comments inline above…

Joe

[blowlampParticipant @blowlamp]

Posted by John Haine on 25/04/2023 11:07:43:

The pendulum decelerates for two reasons – it's climbing against gravity, and losing energy. If the attractive force just balances the friction force (or rather the energy given equals the energy lost), the pendulum won't know the difference except that if the "centre of gravity" of the impulse is not at BDC the timing will be affected. If the magnet was repelling it's exactly the same principle, except that you don't want permanent magnets anywhere near a steel bob or iron base as the magnetic force will affect the effective gravity.

I'd love to see Dave's presentation but alas it clashes with the AHS AGM where they have some talks on the Big Ben restoration.

Edited By John Haine on 25/04/2023 11:09:16

Hi John.

I understand that the impulse is required to replace only the lost energy and should be as 'unobtrusive' to the natural movement of the pendulum as possible. This is where I think Dave's current setup may be lacking.

I also understand that a steel/iron structure will interact with permanent magnets and so will cause problems with timekeeping. My example was just to illustrate that I think a repelling force might be more appropriate for what Dave would like to achieve.

We know that a perfectly free pendulum will tend to be a better timekeeper, because of the lack of interference by impulse mechanisms. By impulsing in the deceleration phase of the pendulum's movement, I think he is causing more disruption to its natural movement than is optimum. Once past BDC, a free pendulum will gradually decelerate until it stops and reverses. Unfortunately, Dave's setup changes the process and causes the pendulum to accelerate when it would normally be slowing down, by the fact that the closer the bob gets to the magnet, the stronger it will be attracted to it.

In contrast, if the electromagnet were to be fitted directly under the bob, it would be easier to impulse on either every swing, alternate swings, or even sporadic swings. The pendulum would only see the impulse as a slight increase in gravitational force, rather than an occasional pull to one side.

Martin.

[duncan webster 1Participant @duncanwebster1]

Posted by John Haine on 25/04/2023 11:07:43:

I'd love to see Dave's presentation but alas it clashes with the AHS AGM where they have some talks on the Big Ben restoration.

Edited By John Haine on 25/04/2023 11:09:16

It will probably be recorded so SMEE members can watch later. Impulse should be equally disposed either side of centre. This is what you get with synchronome/pulsinetic. The only way I can see of doing this non mechanically is using an aluminium vane and a linear motor. I'm not going to attempt it.

[John HaineParticipant @johnhaine32865]

I'm hoping there will be a recording.

Several types of very successful escapements impulse after BDC when the bob is decelerating. Perhaps the best example is the grasshopper, which is a recoil escapement where the impulse torque increases from just before BDC when the "legs" interchange right up until the peak excursion on the "other side". A good example is on Clock B which doesn't suggest that this is a disadvantage.

[duncan webster 1Participant @duncanwebster1]

Grasshopper impulses every time, so it falls into my 'doesn't matter what you do as long as it's the same every time' category. I do realise this is a generalisation!

I used to have a copy of 'the science of clocks and watches', but I let it go. This explained very well best position of impulse.

[duncan webster 1Participant @duncanwebster1]

Posted by SillyOldDuffer on 25/04/2023 10:59:48:

………… Duncan Webster's no doubt pungent remarks on my approach to statistics! And there will be several other well-qualified engineers present. Should be fun.

Dave

I wouldn't presume to comment on statistics. I once proved beyond doubt hat the 2 test rigs we had for testing production machines were giving consistently different results. The boss of development took 2 machines and tested them on both rigs. No difference. Lies, damn lies and statistics!

[duncan webster 1Participant @duncanwebster1]

Posted by SillyOldDuffer on 25/04/2023 10:59:48:

………… Duncan Webster's no doubt pungent remarks on my approach to statistics! And there will be several other well-qualified engineers present. Should be fun.

Dave

I wouldn't presume to comment on statistics. I once proved beyond doubt hat the 2 test rigs we had for testing production machines were giving consistently different results. The boss of development took 2 machines and tested them on both rigs. No difference. Lies, damn lies and statistics!

[SillyOldDufferModerator @sillyoldduffer]

Breaking news: the clock is running!

No celebrations because I made two silly mistakes with the in vacuo circuit board.

• A connector plug rises vertically from the board, no problem until the female is attached and doesn't leave enough room to slide the soil pipe on. This can be fixed with a right-angled connector.
• Worse, I completely forgot about the pressure/temperature sensor! Hope there's enough room for it on the existing board and I don't have to make another one from scratch. Doh.

Meanwhile I'm letting it run until there's enough period data in the log file to measure the standard deviation.

Dave

[Joseph Noci 1Participant @josephnoci1]

So, how did you make it run? Give it a bash? tilt it? Or is it not in vacuum yet?

 

Edited By Joseph Noci 1 on 25/04/2023 22:36:45

[Michael GilliganParticipant @michaelgilligan61133]

Posted by SillyOldDuffer on 25/04/2023 10:59:48:

Posted by Michael Gilligan on 25/04/2023 00:39:50:

Posted by SillyOldDuffer on 24/04/2023 20:31:51:

And I thought all I had to do was keep the amplitude constant….

.

At the risk of getting overly philosophical about it, Dave

… If you could do that, then the statement would be true

My understanding is that your system is working by a process of active adjustment, and therefore the amplitude cannot be kept constant.

MichaelG.

The system is experimental, and I've tried both. […]

.

Sorry, Dave … but I think you might have missed the ‘philosophical’ point I was trying to make:
If there is any form of feedback being usefully involved, then the pendulum is an ‘agile’ device in a constant state of jitter … ergo the amplitude cannot be constant.

MichaelG.

[SillyOldDufferModerator @sillyoldduffer]

Posted by Joseph Noci 1 on 25/04/2023 22:36:14:

So, how did you make it run? Give it a bash? tilt it? Or is it not in vacuum yet?

Not in a vacuum yet. I altered the chariot (the 3D-print assembly that holds the magnet and IR sensors) to move the magnet closer to the bob. Not entirely happy with the geometry because it means the bob is impulsed later than I would like, so I'll have to print another one, which moves the sensors closer to BDC.

Ideally the chariot would allow the sensors and electromagnet to be adjusted independently relative to bob BDC, but I'm trying to keep the build simple.

I've ordered the pipe and fittings needed to exhaust the container, but I need to correct some silly physical mistakes on the veroboard before the soil-pipe can be installed. I forgot the BME280 and positioned a connector so the soil-pipe can't get past it to bed in the base.

Dave

[SillyOldDufferModerator @sillyoldduffer]

Progress report. I'm running the clock without compensation to see how much impulse the revised pendulum needs – the bob is nearly 5 times heavier than the Mk 1.

A few problems:

• Self-start isn't totally reliable (failed once)
• Despite being mounted in a brass tube, the IR receiver was still sensitive to ambient light entering from the rear. Even though it shouldn't be necessary when the container is fitted I've added a shroud, which ensures good behaviour whilst testing the uncovered clock in daylight.
• Software bug in my code, where compensation was applied even when told not too. As the temperature sensor isn't fitted yet, the compensated period was wrong.
• Had to write new code to report on reduced data (no temperature etc.) Made more mistakes!

However, results are promising:

Nice straight line comparing my clock with NTP: it's consistently slow, which is easy to correct.

Dispersion is good (Q nearly 4000), but has two peaks, which is consistent with over-impulsing. I found balancing impulse so amplitude stays constant (approximately – see MichaelG's comment!), to be tricky. 18 is too low, 20 is too high, and the stats show 19 is slightly too high. The hardware timer's resolution is a little too coarse.

Something odd with the waveform, which displays a comb effect. Values that jump much further from the mean. The comb exaggerates the error – 88 times in 43153 is 0.2%, and it's not that big – standard deviation is 0.27uS

The wild jumping about at the end is due to me eating breakfast next to the clock and vibrating it.

Annoyingly have to go out now, because I have several changes to make!

Dave

[Michael GilliganParticipant @michaelgilligan61133]

That’s great progress, Dave

… good to see.

MichaelG.

[Author]

Posts