EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: micha1102b on March 23, 2022, 03:06:39 pm
-
Hi there,
does anyone know how this electronic "eternal pendulum" described here
https://www.mikrocontroller.net/topic/434344 (https://www.mikrocontroller.net/topic/434344)
(unfortunately in German language) can be made to work? I found it by chance during search. It seems promising due to the simple dual use single coil used in the schematic as well as the easy way to scale up the coil and its driver circuitry for use in more power demanding applications, which I'm actually looking for.
Any idea is much appreciated.
Michael
-
Google Translate can translate full webpages, too, not only words. Works pretty well.
- Go to https://translate.google.com/
- In the lefthand side select German, in the righthand side select English
- copy/paste the link of the German webpage into the leftside of the translate page.
- click the link shown in the righthand side textbox, and the translated page will appear.
https://www-mikrocontroller-net.translate.goog/topic/434344?_x_tr_sl=de&_x_tr_tl=en&_x_tr_hl=en&_x_tr_pto=wapp
-
Many years ago, I found an article in Scientific American about how the Foucault Pendulum exhibit at Chicago's Museum of Science and Industry drives the pendulum in order to compensate for the energy lost in friction and air resistance, but have been unable to find anything recent about the circuit. It involved some kind of coil set below the bob, and had to avoid distorting the direction of swing.
-
Many years ago, I found an article in Scientific American about how the Foucault Pendulum exhibit at Chicago's Museum of Science and Industry drives the pendulum in order to compensate for the energy lost in friction and air resistance, but have been unable to find anything recent about the circuit. It involved some kind of coil set below the bob, and had to avoid distorting the direction of swing.
I would guess there is a solenoid-shaped electromagnet below the center with axis oriented vertically. It just needs to be turned on anytime past the peak excursion on the way to the center. And off just as it crosses the center to the opposite peak. ie. on the trailing-edge of the sine wave (95° to say 180°, 275° to say 0°).
The above assumes the electromagnet attracts the metal bob and center detection is done optically. The implementation linked by OP may work by repulsive force on a magent embedded in the bottom of the bob. Most of the time the electromagnet is used as a sensor which detects the bob as it approaches the center and just passes it, then the MCU turns on the electromagnet, giving the bob a brief kick away from the center. ie. on the leading-edge of the sine wave (5° to say 40°, 185° to say 220°) .
-
It doesn't need to know the exact moment when the pendulum and the coil are aligned.
It's enough to sense some voltage increase induced into the coil by the incoming pendulum
- as soon as some voltage is detected, a short enough pulse will energize the coil and put some extra energy into the pendulum
- the kick pulse only needs to be placed somewhere on the incoming trajectory, anywhere
- the pulse needs to be short enough so it won't act as a break on the outgoing trajectory
-
Thanks for your input so far.
If I got the auto translation correctly, the circuit only works with a dedicated piece of controller firmware which I still haven't found in that linked thread yet.
-
.
-
If I got the auto translation correctly, the circuit only works with a dedicated piece of controller firmware which I still haven't found in that linked thread yet.
Yes, any microcontroller will need a dedicated firmware to work, though the algorithm needed for a perpetuum pendulum is very simple: If a voltage increase is detected, pulse the coil for a very short time.
To do that, it doesn't even need to be a microcontroller. Can be done as well with a 555 timer, or with just 2-3 transistors and no integrated circuits at all.
-
It's generally a bad idea to influence the swing of a pendulum on every oscillation, that takes you away from the concept of the 'free pendulum', where the natural movement of the pendulum doesn't suffer any interference (energy gain or loss).
The traditional master clock uses a gravity escapement, where the pendulum is excited by small mass falling under gravity, at the point of maximum velocity, every 30 to 60 oscillations to minimise interference, you can fairly easily get to 2ppm accuracy that way... https://www.eevblog.com/forum/chat/what-did-you-buy-today-post-your-latest-purchase!/msg1154556/#msg1154556 (https://www.eevblog.com/forum/chat/what-did-you-buy-today-post-your-latest-purchase!/msg1154556/#msg1154556)
Even better, was the Shortt–Synchronome clock, which was used as an international time standard from the 1920s to the 1940s, when crystal clocks came in. This operated a master pendulum in a vacuum chamber, with the only influence on its movement was an tiny gravity weight controlled by a synchronised slave pendulum. This was so accurate that the biggest influence on it's timekeeping was tidal nutation.... https://en.wikipedia.org/wiki/Shortt%E2%80%93Synchronome_clock (https://en.wikipedia.org/wiki/Shortt%E2%80%93Synchronome_clock)
Here's a previous thread on the same subject that might be of some help... https://www.eevblog.com/forum/projects/electronic-pendulum-(self-optimizing)/msg3597889/#msg3597889 (https://www.eevblog.com/forum/projects/electronic-pendulum-(self-optimizing)/msg3597889/#msg3597889)
P.S. If you want to energise the pendulum more frequently, then the best thing is to sense and control the absolute amplitude of the swing - the period of its motion is very dependent on amplitude, a true pendulum would have a parabolic arc, not a circular one. Changes in the size of the impulse caused by small changes in supply voltage etc. have a big effect on accuracy.
-
By the words "eternal pendulum" in the title, I was assuming it's about some toy pendulum, where all it matters is that it doesn't stop. If the pendulum's destination is time keeping, that would be more complicated.
micha1102b, what is the use of the pendulum you want to build?
-
Looking back at the previous thread that I linked, I just realised that it is the same OP!
... and same schematic.
-
.
-
.
-
P.S. If you want to energise the pendulum more frequently, then the best thing is to sense and control the absolute amplitude of the swing - the period of its motion is very dependent on amplitude, a true pendulum would have a parabolic arc, not a circular one. Changes in the size of the impulse caused by small changes in supply voltage etc. have a big effect on accuracy.
There were, after Shortt clocks, others named Fedchenko (http://leapsecond.com/hsn2006/pendulum-tides-ch5.pdf) clocks that using a modified suspension achieved isochronism (equal period despite amplitude). The bob hanged from three different suspension springs, having different lengths and widths, this approximated a cycloidal (https://en.wikipedia.org/wiki/Cycloid) arc, which is the mathematical condition for isochronism. Some details can be found here (https://articles.adsabs.harvard.edu//full/1957SvA.....1..637F/0000637.000.html). AFAIK these Fedchenko clocks were the most accurate mechanical clocks ever made.
Back to the thread, I do still have a pair of old Kundos similar to that on my previous image. In my list of long awaiting projects there is one consisting on replacing the simple 2 transistor circuit by a more recent version fulfilling both targets:
- achieve constant amplitude
- compensate for temperature
Not even decided if I wanna use a μcontroller of prefer an analog circuit. Some day...
Yes you're right, it was cycloidal arc that I was trying to think of, not parabolic.
I've seen a photo of a mechanical clock that had an extra long pendulum suspension spring and a pair of curved cheeks that it wrapped around to approximate the correct cycloidal curve - by no means perfect, but sufficient to improve the accuracy with the remaining variation in spring energy, after the fusee.
-
.
-
Correct in every respect! :)
I've attached the photo, from one of my clock books, that I mentioned. As you say, wide swinging verge escapement (so obviously no fusee at that date!), with silk suspension passing between the cheeks. I noted from your very nice Fedchenko pdf link how tiny the swing is, minimising the difference between the circular and cycloidal arc that needs to be compensated.
As a curiosity I've also included a photo of one of Robert Hook's experiments in driving a long heavy pendulum with a watch escapement, both minimising interference by the driving force and presumably also the arc of swing (I wonder how many escapements he ruined with that experiment!). It's a nice example of how efficient a swinging pendulum is in conserving energy though.
-
@all:
Thanks so far for your interesting input.
I'll be back after digging through all the stuff. ASAP. ;)
Best regards
-
(Gyro); I read your comment, about CLOCK BOOKS.
A really good, engineering approach is seen, in my copy of 'Laurie Penman's book, on various clock escapements.
Recommended. I think the Penman book is called:
'Practical Clock Escapements'.
Lots of detail.
-
Years ago (around 20) this "solar pendulum" appeared in the book "Junkbots, Bugbots & Bots on Wheels":
http://www.accomplished.org/2010/06/02/solar-pendulum/ (http://www.accomplished.org/2010/06/02/solar-pendulum/)
That's about the same time frame in which those solar powered flip-flap flowers appeared on the market and as I recall early ones were made of discrete components -- later replaced by an epoxy blob.
The book suggests using the pancake coils you would find inside a 5 1/4" floppy drive motor.
[attach=2]
(Probably not hard to find in the early 2000's but more difficult today.)
An explanation of the circuit begins on page 201. A copy of the book may be borrowed from the Internet Archive library:
https://archive.org/details/junkbotsbugbotsb0000hryn (https://archive.org/details/junkbotsbugbotsb0000hryn)
-
https://nutsvolts.texterity.com/nutsvolts/201208/?folio=32&pg=32#pg32 (https://nutsvolts.texterity.com/nutsvolts/201208/?folio=32&pg=32#pg32)
I've seen one run for 40 years from a https://www.worthpoint.com/worthopedia/1960s-ever-ready-gas-lighter-battery-282104148 (https://www.worthpoint.com/worthopedia/1960s-ever-ready-gas-lighter-battery-282104148)