Timegrapher missing beats?

[spectre6000]

I'm trying to time a small movement, and I can see/hear that the time grapher is missing every other beat. It sounds like a galloping horse. It's ticking out little triplets most of the time. I put another watch on to check, and it was happy as a clam. I turned the gain all the way up. It's just not catching half the balance's movement about half the time. What's that about?

Edited December 20, 2020 by spectre6000

[jdm]

Sometime turning gain down is what helps. However, most likely it's making a noise that confuses the instrument. Listen to it, or record and post . That is an important diagnostic step often neglected.

[spectre6000]

There's a direct correlation between the beats I can see it make, and the beats that it's recording. Since it was missing a beat I could SEE it making, I turned the gain up, and it still missed them. Once that was happening, I could snap my fingers or what have you, and it would catch that as well. Is it possible something could be cushioning a beat or something?

[JohnR725]

5 hours ago, spectre6000 said:

It sounds like a galloping horse.

It would be nice if you would take a picture so we can see what the machine is showing?

Classically the sound of a galloping horse on a timing machine usually means one thing. Except it does not match your description? But I'm still going to give you a link to the video anyway.

https://youtu.be/Rcqrb3_vin8

[spectre6000]

I'm pretty sure it's not rebanking. It's losing beats rather than adding them. I can watch the balance oscillate and listen to the timegrapher fake-tick in response. Out of two full oscillations, only three of the impulses register. The fourth goes unnoticed. This is only an average. It's not entirely consistent, and will occasionally catch everything correctly for a few cycles.

The timegrapher plot is just a scatter plot, and it won't go long enough to get an entire screen's worth of data points without resetting and starting over, even if I manually set the rate.

I put another identical movement on the timegrapher, and it checked out just fine. 

Sorry it's sideways.

[JohnR725]

We need a ham radio operator as they understand Morse code as that's what it sounds like? One of the unfortunate things about the Chinese timing machines is it would be nice to have actual audio. In other words what does that really sound like is that they hairspring bumping into something it's really interesting if you can listen to the audio of the watch. Unfortunately that being a smaller watch you're going to have to have really good hearing to hear it if you put your ear.

But it definitely right that is not what I thought it was at all.

Let's go for a long shot here I assume your timing the watch where it's all wound up? Try letting all the power off and only winding it enough that the watch runs and see what happens I suspect it's going to make no difference at all but let's try. Usually it's much easier to find things that repeats like they hairspring opens up and hits the center wheel is a common problem but yours seems to be random that it's worse.

 

[spectre6000]

I usually leave the sound off on the timegrapher since it's not real and doesn't tell me anything.

In that video, it is wound at least most of the way (it was fully wound a handful of hours earlier in the day). I initially only wound it half way to be in the middle of the spring's curve, but it started doing this, so I wound it more to see if more energy in the spring would get it to hit hard enough to register.

It does have the center wheel overlapping the balance. I'll take a closer look at the hair spring next time I'm there and able to do so. I got seven identical NOS watches with this, or very similar movements (Lorsa 8F and 8FA so far, and there may be one other nearly identical movement in the batch). 

[JohnR725]

2 hours ago, spectre6000 said:

I usually leave the sound off on the timegrapher since it's not real and doesn't tell me anything.

Yes one of the minor downfalls the Chinese machines no real sound. It's one of my unfinished projects is to go into the Chinese 1000 machine and add in sound but I just haven't got there yet. Because sound is so really nice real sound for hearing problems.

2 hours ago, spectre6000 said:

I got seven identical NOS watches with this, or very similar movements (Lorsa 8F and 8FA so far,

One little reminder new old stock does not necessarily mean functional perfect running watch. This came up for a different watch another discussion but it was found that the new old stock really wasn't a very good watch at all. So we decided that maybe the reason why there is new old stock especially if it's old was because it had an issue. Except when you look up this movement it does seem to be readily available or was available so maybe it's not the case with this watch if you have running ones.

 

https://17jewels.info/movements/l/lorsa/lorsa-8f/

[spectre6000]

I was actually thinking about how to split out the signal from the piezo microphone, amplify it separately, and use the existing speaker to play the actual sound of the movement this morning. 

These watches mostly run to some degree, just not super well after sitting around for 40-50 years (no idea when they were actually made, but they look very 70s to my eye, and were sold as such). I mentioned it in another thread, but I actually got one out of the batch to run really well last night. This evening (assuming I can find the time), I hope to see just how good I can get it on the timegrapher. Right off the movement holder, it was 0 beat error and 0 seconds in two or three positions with very minor fiddling. Another position was something like -8, and that was as bad as it got. Part of the intent of the accuracy expectations thread was to get an idea of what I should expect for this exact movement. I figure it's a pretty average movement, which would suggest I could call my efforts a success if I can get the positions to within a minute of each other. So far, it's looking like it's in precision movement territory, even though I doubt it qualifies by any other metric. 

In addition to at least one of the batch being plenty serviceable, I recall this movement having not done this that last time I had it on the timegrapher. No idea what could have happened in the middle. Might be a clue.

[JohnR725]

11 minutes ago, spectre6000 said:

I was actually thinking about how to split out the signal from the piezo microphone, amplify it separately, and use the existing speaker to play the actual sound of the movement this morning. 

I was thinking about finding the signal in the machine itself. The microphone itself does have an op amp so you could tap into that output but it's still need something to drive the speaker. Then I was looking at eBay for amplifier boards because if I modified a timing machine conceivably others might want to do the same thing if it worked so I would need something, that people could get versus me building something of my own. I picked up a couple of the amplifier boards one of them I liked because it was nice and tiny and it basically had a enable switch. The reason I like that was on the witschi timing machines that do have a speaker with audio you just push a button the audio comes on. There is no volume adjustment and it works really well.

But like so many other projects it's not getting done. Then the insides of the 1000 and 1900 are similar. Okay sort of similar in that there is a little more inside the 1900 because of the graphical display. The signal processing is somewhat similar.

29 minutes ago, spectre6000 said:

Part of the intent of the accuracy expectations thread was to get an idea of what I should expect for this exact movement.

One of the problems with accuracy specifications are there misleading. For instance Seiko watches are known to keep bad time. But if you look at the specifications they aren't necessarily outstanding specifications. I suspect it's a way for the factory to adjust their watches close by automated process and its within specifications. But what if you take one of those crappy running Seiko watches put it on the timing machine and make the beat better And regulate better than what? Then you have a watch keeping better time. It's actually one of the services were I work my boss likes impressing the timing machine so for a fee He will will regulate your watches better than the factory did. It's one of his selling points is the comment that the factory didn't take the time to regulate as close as they could. Which is what happens when you watch company making thousands of watches by automated process they're not going to take the time to regulate super close. 

 

[spectre6000]

4 hours ago, JohnR725 said:

I was thinking about finding the signal in the machine itself. The microphone itself does have an op amp so you could tap into that output but it's still need something to drive the speaker. Then I was looking at eBay for amplifier boards because if I modified a timing machine conceivably others might want to do the same thing if it worked so I would need something, that people could get versus me building something of my own. I picked up a couple of the amplifier boards one of them I liked because it was nice and tiny and it basically had a enable switch. The reason I like that was on the witschi timing machines that do have a speaker with audio you just push a button the audio comes on. There is no volume adjustment and it works really well.

But like so many other projects it's not getting done. Then the insides of the 1000 and 1900 are similar. Okay sort of similar in that there is a little more inside the 1900 because of the graphical display. The signal processing is somewhat similar.

If that's your aim, maybe check out an eval board. A while back, I needed some ultra low noise amplifiers to buffer/boost the signal on some pH probes. I was able to find an eval board that was practically plug and play right out of the box. Way easier than designing, printing, populating, and soldering PCBs. Might be a cheap and effective way to get a nearly plug and play solution. 

I haven't been inside my 1900 yet, but here's where I'd start:

https://www.digikey.com/en/products/filter/evaluation-boards-audio-amplifiers/789?s=N4IgbCBcoJYCZRAWgIwAYUgDQgA4BdFsR8BPXAU0QEMBnAYxAF8mg

I don't know what the specs are on the speaker, what power is available or what the specs are on the signal, but there's bound to be something that's at least close to functional out of the box.

4 hours ago, JohnR725 said:

One of the problems with accuracy specifications are there misleading. For instance Seiko watches are known to keep bad time. But if you look at the specifications they aren't necessarily outstanding specifications. I suspect it's a way for the factory to adjust their watches close by automated process and its within specifications. But what if you take one of those crappy running Seiko watches put it on the timing machine and make the beat better And regulate better than what? Then you have a watch keeping better time. It's actually one of the services were I work my boss likes impressing the timing machine so for a fee He will will regulate your watches better than the factory did. It's one of his selling points is the comment that the factory didn't take the time to regulate as close as they could. Which is what happens when you watch company making thousands of watches by automated process they're not going to take the time to regulate super close. 

That's why I was just looking for rules of thumb. It doesn't seem like it's worth it a lot of the time to try to track down service sheets or manufacturers' specs. A good rule of thumb for, in this case, an average movement is good enough for me.

[JohnR725]

10 hours ago, spectre6000 said:

If that's your aim, maybe check out an eval board. A while back, I needed some ultra low noise amplifiers to buffer/boost the signal on some pH probes. I was able to find an eval board that was practically plug and play right out of the box. Way easier than designing, printing, populating, and soldering PCBs. Might be a cheap and effective way to get a nearly plug and play solution. 

Actually my intent is to put a Lot of restrictions on the project? For instance if I want to hear a watch ticking for me it's easy I have a witschi watch expert 2 timing machine Audio it's a built-in feature. But it's not a built-in feature for the Chinese machines which is what's recommended for anyone in watch repair. That means the restriction would be make use of the machine that everyone currently has.

This is one of the reasons I was looking at eBay for preamp/amplifier board's was to get something that everybody could work with.

Then I'm quite familiar with that company all the orders I've made over the years. Although I do forget that they have evaluation boards even though I've purchased them from their. So I have to see what I can find. The thing with eBay is usually people anywhere in the planet could get something from eBay.

Then normally you would be right that designing making a circuit board is a pain. Having over the years tried Quite a few of various methods and now you can just send a way it's insanely cheap except? I've accumulated a sizable pile of new copper circuit board material. Then several years ago I saw a milling machine for making circuit boards but way way out of my price range. Then cheap Chinese timing machines, one of them is sitting on my kitchen table and I can make circuit boards by milling them. I've already completed two of them and the third one is waiting for the Christmas holiday to be populated. I can't say it's a lifelong dream but it's been a long time dream to do this.

basically a project with restrictions to make use of what people on this discussion group have.

Then as far as being inside of the machines I have photographs of the inside of both the 1000 and the 1900 machine. I just haven't got around to digging out the oscilloscope and seeing if it still works. As it's been a very long time since I've turned it on last.

11 hours ago, spectre6000 said:

That's why I was just looking for rules of thumb. It doesn't seem like it's worth it a lot of the time to try to track down service sheets or manufacturers' specs. A good rule of thumb for, in this case, an average movement is good enough for me.

I was thinking maybe you should go back to the old days? The old days of the paper tape timing machine which you Chinese machine sort of simulates. In the days of the paper tape machine watch was placed on the microphone paper started feeding out the sound of the printing on the paper occurred. This is where the sound of your horse galloping was very very noticeable and in a classroom situation everyone knew what was going on. All you worried about was trying to get the lines close together they look reasonably straight and the usual recommendation is the watch should run fast. The modern newfangled digital machine the graphical display simulates the paper tape ignore everything else sort of. The only reason I added in sort of is if you go to extreme on the graphical display you get a rollover error C do need to pay attention to the beat number. In the words of it says some extreme high number of the graphical display looks perfect that it probably really is perfect. So basically just don't worry about everything that machine has to offer regulate the watch reasonably close let it run 24 hours if the hand seem to be keeping time you're fine.