Counting the Swings of a Clock Pendulum

  • Be sure to checkout “Tips & Tricks”
    Dear Guest Visitor → Once you register and log-in:

    This forum does not automatically send notices of new content. So if, for example, you would like to be notified by mail when Steve posts an update to his blog (or of any other specific activity anywhere else), you need to tell the system what to “Watch” for you. Please checkout the “Tips & Tricks” page for details about that... and other tips!

    /Steve.

rfrazier

Well-known member
Sep 30, 2020
489
143
@JimWilliamson You bring up some interesting points which I'll have to reply to tomorrow. Past my bed time here at 03:23 AM. But in answer to your last question, I ERRONEOUSLY thought the clock's target beats per hour had to be a multiple of 3600 since the Clock Tuner software only had those values in its default menu. I found out you can customize that menu. I attempted to set the Grandmother clock to 3600 BPH which is too slow for it, and the wall clock to 7200 BPH which is too fast for it. That's what prompted me to go down this road to fix those errors. The reality is that while some clocks are multiples of 3600, many, maybe most, have their own beats / hour number that they're designed for. By counting the ticks as the minute hand goes fully around at least once, I hope to document the proper number for each clock. This may not be obvious at first. It wasn't to me. But, the number of ticks required to make the minute hand go around once is, BY DEFINITION, the number of beats per hour for that clock, based on the gear ratios in the power train. What we have to do when we adjust it is to make that number of ticks occur in exactly one hour of real time. Therefore, the hands represent real time. For the weight driven clock, setting the beat rate to its design rate number should put it right on time or really close, or back on time if the need arises later (assuming I still have the ability to measure beat rate at all). It won't matter if it's on time for the measurements, since the clock will think it's gone an hour if that number of ticks has occurred. For the spring driven clock, setting it to its design rate number would put it on time at first, but presumably it would slow down as the spring unwinds. More later.

May your bits be stable and your interfaces be fast. :cool: Ron
 
Last edited:

rfrazier

Well-known member
Sep 30, 2020
489
143
Pondering a SHTF event: If I wished to have reliable, on-time "time" after an event that might disable electronic clocks / electric devices, I might place into proper storage a few electronic clocks/devices that might survive an EMP, along with power supplies that would be needed to run them. I'll assume said electronic clock would survive such a disruption and could be activated in a week or two. At that point, I'd look to keep my mechanical clock running as it had been when any disruption happened. The time of the mechanical clock would be taken as true. When an electronic clock could be activated and set off of the mechanical's time, the electronic clock would then be taken as "true" time. The electronic clock would offer tighter tolerance for keeping time. The mechanical clock would then set its time to the electronic clock (though only after the electronic proves it had not been damaged). Should any disruption happen to the electronic, the mechanical would again be taken as true (and again be used to set an electronic device, should it come to that). This has me thinking I would wish to keep my over-the-air atomic clock in a protected storage device, plus one or two other electronic clocks. Over-the-air atomic time might not be operational, but if yes, it could be a trustworthy master time device. A GPS might be wise to place in protected storage, in case an EMP does not take out the satellites - they could be a good master time source.
@JimWilliamson Yes SHTF is what we're discussing. That acronym / phrase always perplexed me. Why would someone throw sh** at a fan? But I realized that is what world leaders are doing. All we can do is try to stay out of the way. Re: EMP. Blast, I had forgotten that little factor and was only thinking of long term power failure. You might say we'll have more important things to worry about than the time of day, but time can still be important to arrange meetings with people, gather at events, listen to news, etc. I'm going to coin a new phrase, instead of NSFW - not safe for work, I'll call this a NSFH - not safe for humans situation. You bring up good points about getting through an EMP. It's worthy of thinking about and takes the difficulty of planning to a whole new level. It depends on whether the EMP fries your stuff permanently or temporarily. I have a very small Faraday bag from https://www.satellitephonestore.com which came as part of a mini solar panel kit. I may have to think bigger.

I found this USB rechargeable clock which claims to run 100 days on a charge that might be handy.

Digital Alarm Clocks for Teens, Upgraded 5.7" Electronic LCD Clocks for Bedrooms with USB Charger

I'm going to keep thread related to clocks and start a new thread about powering small devices in general.
That thread is here:

May your bits be stable and your interfaces be fast. :cool: Ron
 
Last edited:

rfrazier

Well-known member
Sep 30, 2020
489
143
WHOO HOO. I got the photoelectric sensor counter and it WORKS ... with some caveats. If you have any interest in playing with photo sensors, whether related to clocks or not, I'd recommend tinkering with this kit.

I was able to get what I think is a pretty accurate BPH rate for my Grandmother clock, which is close to the number I was approaching just by continually tuning the rate. But, it's good to have confirmation of it. But, the setup is very touchy and non obvious and hard to replicate. I cannot tell you exactly how to do it, as yours will be different. But, I can tell you how the equipment works.

Since people may be interested in the photoelectric sensor counter kit that are not interested in clocks, I've put the information on that in a separate thread.


May your bits be stable and your interfaces be fast. :cool: Ron
 
  • Like
Reactions: SeanBZA

Ralph

Well-known member
Sep 24, 2020
121
20
I assumed all those adjustments were correct and left them as they were. My clock varies in accuracy week to week. I fine regulated it pretty close. Some weeks it's slow, others fast, and on occasion dead on. I got it mostly because I like mechanical clocks, not precision time keeping. Now days there are enough things around that have accurate time. In the days before atomic clocks I used to use WWV and CHU (Canada) on shortwave to get the correct time. I think WWV has dropped a frequency or two since then. It used to be 2.5, 5, 10, 15, 20, and 25MHz which made it a quick way to check reception across the HF band.

Wild Spectra Mobile worked pretty good on balance wheels, I still have it on my Android tablet. On the spring winding down effecting the accuracy, I have a Russian made clock where they actually put a pin in it to prevent the mainspring from winding down too far. The clock runs for a week so I guess they took the 'lost windings' into account.
 
Last edited:

rfrazier

Well-known member
Sep 30, 2020
489
143
Hi @Ralph It's true that mechanical's aren't incredibly accurate nor consistent. My Grandmother clock was running fine before I munged up the timing. If I get it back under +/- 1-2 minutes per week, I'll be happy. I have noted that I can only get to about 1-2 ticks per hour accuracy with my counting method which amounts to about 1 minute per day. The fine tuning will have to be the set it, adjust it, watch it method. The other wall clock has been rarely used, so I'm still seeing what it can do. I've got a vintage Chinese travel alarm clock on the way and an automatic watch. Those 4 will probably keep me plenty busy setting them every week. But, they'll be running if I need a backup to the electronic ones. Interesting point you made about a restricted winding spring.

May your bits be stable and your interfaces be fast. :cool: Ron
 

Ralph

Well-known member
Sep 24, 2020
121
20
I have a clock with a similar issue regulating. It is what I believe is called an anniversary clock. It runs about 3, maybe 6 months on a wind and is running pretty fast. It takes roughly 5 seconds to rotate the weights in one direction. Regulation is by turning a disk above the weights which changes the diameter of the weights. For now I just keep it running until it stops and I take a wild guess turning the regulator and see what happens.

For anyone not too familiar, a fairly common balance wheel watch (not a pendulum, rather a wrist or pocket watch) runs at 2.5Hz, that's 5 ticks per second. Each tick is a swing in one direction of the escape lever and rotation of the balance wheel. That works out to 18.000 times per hour, or roughly 158 million times per year. Even a mediocre mechanical watch can run for decades with little to no maintenance. Billions of mechanical movements while keeping pretty accurate time and not wearing itself out is pretty amazing.

I saw the pinned mainspring in a Russian submarine clock I bought back in the 1970s or 80s. I removed the movement from it's case to oil it and spotted the pin. The glass cover is on a hinged door with a cam to lock it shut using the winding key. It runs about a week, probably 8 days. Anyway, when I first got it I recall looking at the gasket on the cover and wondering just how water tight the seal was. It wasn't long until I realized if there was enough water in a submarine to stop that clock there were more serious issues going on than what time it was!
 
Last edited:
  • Like
Reactions: SeanBZA

rfrazier

Well-known member
Sep 30, 2020
489
143
Hi @Ralph . It is pretty amazing that the mechanical clocks can run so long. My pendulum clock was up over 35,000 ticks just in a 9 hour (I think) test that I did. You do have to wind and set every week, or every 1-2 days for my travel alarm, which is why I use my electronic clocks as a reference and don't have TOO many mechanical's. But, should the worst happen and the electronics go South, the mechanical's still run.

I'm afraid I don't know about anniversary clocks. I THINK that one should run at about 30 seconds per tick. However, the forums over on the National Association of Watch and Clock Collectors website are a great resource. You could post a question there. Review the forum rules. If you're a newbie, say so. Give them an idea what you do and don't know. There are lots of knowledgeable people over there and many are willing to help. They like to see pictures of clocks people ask about. If you can't actually work on the clock beyond simple adjustments, let them know that too.

Anniversary and 400 Day and Atmos Forum

General Clock Repair Forum

General Clock Forum

Here's my photoelectric sensor thread over there, which has slightly different content from the one here, and of course, different replies.

I've slightly modified my reference chart that I'm filling out as I tune each clock into the best timekeeping zone. This was originally described in post # 17 in this thread.

I have a blank line at the top to identify which clock the card is for.
The columns I now have are: Date, Time (new time I set to), Days (since last check), Variance (from expected time) / Adjustment (made to the nut / lever) (I have variance and adjustment in one column to save space.), Beat (rate after adjustment - if measured), Beat Error (tick tock symmetry after adjustment).

So a sample data line might look like:
Grandmother Clock
09/21/22 1210 PM 3 days + 2 M (means 2 minutes fast) / - N (means small negative nudge to the adjuster) 3963 (BPH) OK (means new beat is OK)

Objective is to get to + / - 2 minutes per week for me.

Good luck with your clock.

May your bits be stable and your interfaces be fast. :cool: Ron
 
Last edited:
  • Like
Reactions: JimWilliamson

rfrazier

Well-known member
Sep 30, 2020
489
143
For what it's worth, here's an image of the reference cards I'm using to set the timing on my clocks. Anyone who wishes may use these or make up your own. See attached PDF.

May your bits be stable and your interfaces be fast. :cool: Ron
 

Attachments

  • Clock Timing Cards09212022.pdf
    139.4 KB · Views: 45

Aldo

Member
Sep 18, 2020
17
7
Here is an updated and cleaned up version, zipped spreadsheets and PDF of the same thing, complete with detail adapted from comment #28.
 

Attachments

  • Clock Timing Cards 2022-09-22.ZIP
    63.1 KB · Views: 38

rfrazier

Well-known member
Sep 30, 2020
489
143
@Aldo Thanks for doing that. I didn't know a quick way to do the chart so I just drew some and photocopied them. Yours are much nicer. Good job. With the caveat that I'm not a clock expert, I have a few MINOR suggestions for alterations or corrections. These are not criticisms, just slight improvements. Let me know if you want me to share those suggestions in a forum post or a private message. I had originally shared the hand written cards on the National Association of Watch and Clock Collectors forum as well. I may wish to share yours there once we tweak them a bit, if that's ok.

May your bits be stable and your interfaces be fast. :cool: Ron
 

Aldo

Member
Sep 18, 2020
17
7
Sure, feel free to reply-comment (this thread is already in the weeds Haha) or PM.
 

rfrazier

Well-known member
Sep 30, 2020
489
143
@Aldo Here are a few thoughts for minor corrections. I'm using your PDF as a reference:

* If you make changes to the files, I suggest appending -v2 to the filename for version 2.

* Beat can be measured in BPM or BPH (or even BPS if you want to get technical). Suggest changing the BPM column to BPM / BPH. Suggest changing the legend for that column to BPM / BPH Beats per min / hr after adjustment, if measured

* In the legend, suggest capitalizing the first letter of New time setting and Tick-tock symmetry after adjustment

* In giving it more thought, I realize my original coding for the adjustments is ambiguous and a bit confusing. That's my fault. Anyone reading this may want to note that this changes what I put in post 28. However, anyone can fill out a chart any way they want. I used the letter N to mean nudge, but it could also mean negative. Also, in your sample data, having N next to + or P next to - seems a little confusing. I propose getting rid of the word nudge and using the word change instead. N can mean negative (slowing the clock) and P can mean positive (speeding up) the clock. So, N, NN, NNN means tiny, medium, large negative change and P, PP, PPP means tiny, medium, large positive change. There are probably a hundred ways to do this and each person may be different. But, I suggest changing the sample data as follows:

*** (at the top of pg 3)
Adj N tiny negative change to the adjuster
BPM / BPH 66 BPM

*** (at the bottom of pg 3)
Adjustments
NNN large negative change to the adjuster
NN medium negative change to the adjuster
N tiny negative change to the adjuster
0 No Adjustment made
P tiny positive change to the adjuster
PP medium positive change to the adjuster
PPP large positive change to the adjuster

I hope this is understandable. I think this will make things a bit clearer. Thanks for your help.

May your bits be stable and your interfaces be fast. :cool: Ron
 
Last edited:
  • Like
Reactions: SeanBZA

Ralph

Well-known member
Sep 24, 2020
121
20
I have to check out those links and sites. I was thinking, before digital days there were people who actually made mechanical clocks. I wonder if they had a forgotten technique to regulate things or they tinkered around for days or weeks trying to get a clock accurate.

Here's a picture of (not sure of the actual name) what I will wrongly call a pendulum. It cannot be seen, but it is attached to a thin metal strip that allows it to rotate. The top of the metal strip has the escape lever. The thin horizontal wheel can be rotated which changes the diameter of the 'ball weights' and in turn the timing. The cone at the bottom center of the 'pendulum' keeps the weight assembly from swinging too much if the clock is moved without locking the mechanism. Since winding, it has picked up about 35 minutes- I need to do some serious regulating!

unnamed.jpg
 

rfrazier

Well-known member
Sep 30, 2020
489
143
@Ralph Interesting picture. I haven't worked on nor owned that type of clock. But, I have a few thoughts based on general principles. Just as an ice skater pulls her arms in to spin faster, it should be same with the clock. I'm assuming the clock is basically running OK. Also, it needs to be totally level. If it's running too fast, you want to slow it down by swinging the balls outward. Gently stop the clock and wait for the pendulum to stabilize in center position, assuming it doesn't restart. Rotate the adjuster so the threaded collar goes up to say half the distance between it and the nut of what's currently visible. This should move the balls outward. Wind it and set the time and record what you did on one of @Aldo s charts if you wish. Gently start the clock by rotating the pendulum no further than it normally swings. I don't know what accuracy to expect from those clocks. One thing I read somewhere said that, notwithstanding that it's a one year clock (presumably), accuracy suffers if you don't wind it monthly. I would think certainly those clocks should be as good at keeping time as pendulum or balance wheel clocks. See if it's gaining or losing more than 4 min / week. If it goes from gaining to losing, you went too far. If you have to adjust it again, move the collar half as far next time. Once you get it within 4 min / week, only move the collar a tiny amount, say 1mm on the nut at a time. You might want to go to the 400 day forum on NAWCC and post a question along with photos. You'll have to set up a free login to post there. It may be possible to use the Clock Tuner app on Android to test the beats if it can accommodate beats that slow and if it can hear the beats. I think there's one called ClockMaster on iPhone. But, you'd have to know the target beats per hour for the clock. You might be able to use a variation of the photoelectric sensor process from this thread as well. You could also do an advanced search on NAWCC on the 400 day forum for the word accuracy. That might turn up some interesting info. Good luck!

May your bits be stable and your interfaces be fast. :cool: Ron
 

Ralph

Well-known member
Sep 24, 2020
121
20
I have to check for the apps. You can hear it tick, but it's pretty low,- taking the movement out of it's case may help that. I wonder if you could play the target beats from the app and listen to them along with the clock. If you hit the right one they should stay in sync a while and slowly drift apart. Maybe I'll stop it tomorrow and play a bit. I think it's now around 40 minutes fast, I'll have to keep some paper by it to write notes instead of trying to remember everything.
 
  • Like
Reactions: rfrazier