Info on BENZING

[omblau]

I have recently acquired a wonderful clock by Benzing , in the picture, I have two questions:
Since I am in the process of regulating the clock with clock tuner which kind of regularity should I expect from such a clock in seconds per day for example?
I would like to change the suspension spring since it's somewhat damaged, does anybody know where to find a replacement? I found one on ebay
https://www.ebay.it/itm/185855406271 do you think it's a good replacement?

Untitled.jpg

[Grassi]

Dear, the ebayseller looks absolutely trustful, just measure the dimensions of your needed spring and replace it.
Depending on length you will have to adjust the timekeeping precision of the movement.
Anyway expecting a few seconds deviation a day will be a good value depending on the pendulum rod material.

[omblau]

Thanks Grassi, I'll purchase what advertised on ebay.
A further question: does anybody know how to move the clock hands to adjust these kind of clocks? I tried to push the minute hand and stopped because it felt too much effort. First of all should I push clockwise or counter clockwise or it doesn't matter? Here is a picture of the movement for reference.

IMG_20231220_152810_136.jpg

I have already put the clock in beat with the help of Clock Tuner and regulated the pendulum around 4800 BPH. The 10 period reading seems to fluctuate ± 2 BPH on average, if I am right it amounts to about a fluctuation of one minute per day, .75 * 2/4800 * SQRT( 84000/10 ), quite a bit more than a few seconds per day. Should I stop here or do you think I could do better?
Here is a picture of the anchor and wheel after some compressed air blowing, it looks dirty but I wonder if I could improve the situation by cleaning the wheels or should better leave everything as it is. The anchor seems to have really minimal play in its axis, none on one side and barely perceptible on the the other (hundreds of a mm?). No signs of wear on the pallets.

IMG_20231220_163531_627.jpg

IMG_20231220_162916_470.jpg

[Bernhard J]

Hi,

A rate of a minute per day is imo clearly too much.

I would suggest a complete disassembly of the movement and thorough cleaning of all components. And then (after cleaning) check for wear and attend to any repairs needed. Then the clock should achieve a rate of a few seconds and hold this without fluctuation.

Best, Bernhard

[omblau]

Thanks Bernard, I was checking if an app is available which would allow me to make a recording for days rather than seconds so I downloaded 4 of them on my Android cell moto E7 power: Clock tuner, Watch accuracy, Tickoprint, Timegrapher.
Clock Tuner worked right away, none of the other three produced any data even after a few minutes of testing. Have any of you had a better experience?
I quickly reduced the beat asymmetry (the distance between the two dot lines) to less than 5% of the full scale of around 33000 (it doesn't say in which unit, what would the total of 33075 come from?), I don't know if it's a good result or not, anyway, looking at the graph, much better than my starting point.
I aimed at the average value of the pendulum period, 4800 BPH, and verified that the beat error shown just above the frequency is simply the reciprocal of the frequency, i.e. the variation of the beat, in s per day.
After some adjustments I was able to get down to less than 10 s/d average over 10 beats, the number of measurement samples set on Clock Tuner.
Is it right to evaluate the short-term-noise daily random-error in s/d as approximately one hundred times smaller than that of the fluctuation occurring in the samples measured by the timer? intuitively I would calculate this coefficient as the inverse of the square root of the number of samples per day, in my case 10 samples, i.e. 1/SQRT(24*3600/number_samples) = .01
In this case the error of about 10 s/d over ten periods would be reduced to about .1 s/d in one day, i.e. much less than what occurs in practice. Does this mean that there are other more important sources of noise at low frequency? Like temperature variations for example? should I stop here with the short time calibration ? Can I find out if this pendulum has an invar rod?
I could produce a timer to check the long term stability of this clock using arduino but If an app is available it would be a lot easier. By the way do you know if somebody has already published how to build an arduino timer for acoustic clock measurements?

[omblau]

Bernard, I forgot, I don't want to take the clock apart since I never done it before but I checked the play of the anchor arbor and is unmeasurable on the long side and below 1/10 of a mm on the short side. Is this too much?

[Bernhard J]

Hi,

I am old and my mobile is a 20 year old Nokia, no apps or the like . I adjust my clocks traditionally by monitoring the daily rates and correcting the pendulum, if necessary. The beat adjustment is done acustically by experience.

With old-fashioned greetings , Bernhard