Timegrapher Issue

[mikepilk]

18 minutes ago, praezis said:

In other words: it is able to count to two  - not more.

LOL. Reminded me of the Holy Hand Grenade of Antioch

''First shalt thou take out the Holy Pin. Then shalt thou count to three, no more, no less. Three shall be the number thou shalt count, and the number of the counting shall be three. Four shalt thou not count, neither count thou two, excepting that thou then proceed to three. Five is right out. Once the number three, being the third number, be reached, then lobbest thou thy Holy Hand Grenade of Antioch towards thy foe, who, being naughty in My sight, shall snuff it.'

[Folkvisor]

I think I'm going to case it now. I can always mess with the beat issue by simply taking the back off. It's only gaining 4 seconds a day now which is pretty decent.

Plus that, I'm thinking it's best to quit when I'm ahead...lol

Again, thank you all for your help.

[JohnR725]

3 hours ago, mikepilk said:

Unfortunately the Weishi 1000 does not have this, both lines are the same colour. Once you get about 5-6ms out, they can wrap round and look like you are in beat. That's why I always do a visual check when assembling to make sure it's not too far out.

it's not just your 1000 machine that has this feature it is almost every other timing machine in the existence. As far as I know all of the witschi's use one color for the graphical display. or basically wildcard displays are now relatively common even for a witschi the graphical display is still black for both lines.

then personally I do think it's a good idea to be visually in beat before you're even on a timing machine. Especially with a movable stud because it's so easy to go past and hopelessly become lost as to where you are.  then if you throw in people's rapid movements of going back and forth things get even more fun. At that point you definitely have to stop and visually go into beat as you'll never find it.

 

 

1 hour ago, Folkvisor said:

Plus that, I'm thinking it's best to quit when I'm ahead...lol

that's an attitude at work we would never be able to tolerate. On a regular basis my boss gets obsessed with and on a regular basis we go too far. Yes the golden rule of don't please the timing machine is typically not listen to and yes you can really dig a deep hole if you're not careful and waste a heck of a lot of time trying to fix things when there were just fine the way they are even if they or not perfect there just fine.

[Folkvisor]

I was told that there should be about 90 degrees between the regulator 

12 hours ago, JohnR725 said:

one of the minor issues with having a 6497 is variations. So for instance if it's a modern 6497 putting it in beat is easy because it has the etachron system as you'll notice in the attached technical sheet for this watch. I think this is the last one they printed from 2016. So putting it in beat is easy just nudge the stud holder.

To be really careful and go really slow though because it's really easy to go past and then get hopelessly lost as to where you are.

CT_6497-1_FDE_482160_11.pdf 5.46 MB · 1 download

So, as I understand it, the stud holder puts more or less tension on the hairspring while the regulating lever adjusts its length. Is that sort of how it works?

2 hours ago, JohnR725 said:

that's an attitude at work we would never be able to tolerate. On a regular basis my boss gets obsessed with and on a regular basis we go too far. Yes the golden rule of don't please the timing machine is typically not listen to and yes you can really dig a deep hole if you're not careful and waste a heck of a lot of time trying to fix things when there were just fine the way they are even if they or not perfect there just fine.

A friend of mine was the engineer in charge of the shuttle cooling systems. He told me their motto was: IT ONLY HAS TO WORK ONCE

[JohnR725]

2 hours ago, Folkvisor said:

So, as I understand it, the stud holder puts more or less tension on the hairspring while the regulating lever adjusts its length. Is that sort of how it works?

no that's not how it works at all. The stud holder holds the end of the hairspring. The stud holder allows you to rotate the entire balance wheel and put it in beat it shouldn't put any tension on anything. otherwise without a movable stand holder if you Move that end of the hairspring you have to move the other end by rotating the collet which is why they have a movable study holder makes so much easier.

How much some pictures this is what you're trying to do. No hairspring is shown unfortunately but then that would've made the picture harder to see. In beat is when you have the eight alignment of the roller jewel in the middle of the pallet fork between the banking pins and escape wheel on the other side of the position it is in. Oh and ideally it's not an American pocket watch with movable banking pins because that changes things a little bit you can end up with people shifting both the overall the lead then that messes up things so ideally it's an invisible alignment between the roller jewel and having the pallet fork perfectly centered is what you're trying to do

 

then even those is for a newer watch using the etachron system which I do find amusing because Seiko uses it and they get to do that because they hold the patent for I believe the regulator pins in the start patent is held by the Swiss. In a case here's a picture of how all that works. Then the spacing between the stud holder and the regulator of courses determined by regulating the watch. 

 

 

[Folkvisor]

A fabulous explanation and really clear diagrams! Thank you.

[VWatchie]

On 8/25/2023 at 2:12 PM, VWatchie said:

On my Weishi 1900, one line is blue and the other line is yellow. If/when I go past these lines switch palaces which is a bit helpful. So, it would seem the machine can differentiate between the two beats.

On 8/25/2023 at 4:05 PM, praezis said:

In other words: it is able to count to two  - not more. It cannot tell which is left or right swing, in or out pallet jewel.

Not sure what you're trying to say, but I guess I wasn't very clear about what I was trying to explain, so perhaps there was a misunderstanding 

Anyway, I'll try to illustrate the feature of my Weishi 1900 which is helpful when adjusting the beat. At least when the balance has a moveable stud and you're adjusting while measuring.

In this picture, we're out of beat. But as you can see the two beats have different colours (blue and yellow). 

Now, after having moved the stud we can see that the direction was correct but shot over the target. If the lines had been the same colour, it would have been a little more difficult to tell. Admittedly, you can sometimes see the lines crossing each other while adjusting, but the colours still make it easier. Not critical, but nice to have.

So, perhaps a feature you could implement in the PCTM app!?

[xyzzy]

Newest version of tg color codes the two beats in the paper strip graph.  The sounds are different for the entry pallet vs the exit pallet, as one can see in the two waveforms it shows.  It uses the sounds to tell them apart, rather than just assuming each detected tick will alternate.  There is some risk of making a different choice of which sound should be the "first" if the app is restarted.  The loudest is first, but there is some variation between the teeth of the escape wheel and for other reasons, so it might always be the case that the entry or exit pallet is loudest.

 

 

Edited August 27, 2023 by xyzzy

[praezis]

On 8/26/2023 at 8:25 PM, VWatchie said:

So, perhaps a feature you could implement in the PCTM app!?

Hi VW,

it is there already, but iirr only if you select single noise.

What you showed will only work if you test continuously. If you remove the watch, turn the collet and put it on the mic again, colors will be arbitrary.

Frank

[JohnR725]

21 hours ago, xyzzy said:

The sounds are different for the entry pallet vs the exit pallet, as one can see in the two waveforms it shows.  It uses the sounds to tell them apart, rather than just assuming each detected tick will alternate. 

 

5 hours ago, praezis said:

What you showed will only work if you test continuously. If you remove the watch, turn the collet and put it on the mic again, colors will be arbitrary.

I'm having a confusion with the two separate things I've quoted. We have the top quote which is an absolute you can tell the difference versus the second quote by @praezis which indicates what I always thought that you can't actually tell one from the other.

 

[VWatchie]

7 hours ago, praezis said:

it is there already, but iirr only if you select single noise.

Good, I didn't know! Excellent software BTW! 

7 hours ago, praezis said:

What you showed will only work if you test continuously. If you remove the watch, turn the collet and put it on the mic again, colors will be arbitrary.

Yes, and that's why I wrote...

On 8/26/2023 at 8:25 PM, VWatchie said:

At least when the balance has a moveable stud and you're adjusting while measuring.

 

2 hours ago, JohnR725 said:

the second quote by @praezis which indicates what I always thought that you can't actually tell one from the other.

I'm in that camp as well.

2 hours ago, JohnR725 said:

We have the top quote which is an absolute you can tell the difference

I must have read that very differently from you!? Not an absolute!

[JohnR725]

25 minutes ago, VWatchie said:

Not an absolute!

I'm well aware of that. I once had an argument in another group for somebody absolutely insisted that you could tell the difference the top line was always and the bottom line was always and some arguments you just walk away from

but let me quote the whole thing and rip it apart to make sure all on the same page as I'm still confused by what is quoted below.

On 8/27/2023 at 1:57 PM, xyzzy said:

Newest version of tg color codes the two beats in the paper strip graph.  The sounds are different for the entry pallet vs the exit pallet, as one can see in the two waveforms it shows.  It uses the sounds to tell them apart, rather than just assuming each detected tick will alternate.  There is some risk of making a different choice of which sound should be the "first" if the app is restarted.  The loudest is first, but there is some variation between the teeth of the escape wheel and for other reasons, so it might always be the case that the entry or exit pallet is loudest.

okay let's look at the quote which I'm having a problem with. Then I highlighted one of the lines in bold so we can see the problem. Notice it says the sounds are different and if we had the waveform which we don't it claims you can visually tell the difference. This by the way I was under the impression was impossible but it is quoted above as yes it's true.

Then notice it goes on to say it uses the sound to tell them apart rather than just assuming which once again implies an absolute.

Then we get to the confusion the risk of if you take the watch often put it back on again. Which is the reality of the situation if you take a watch often you put it back on the machine as far as I know there is no way of telling one from the other by listening. It may be possible with  witschi and their newest or one other machines that uses an optical method it would obviously know which direction the balance wheel swinging in so that it would no but only because of that not because of the waveforms being different because as far as I know they are not accept I will comment on the waveforms in a moment

then continuing on and I've made that bold the loudest is first. Then the last line is a contradiction to what I've already commented about in the same paragraph which is why have a confusion

as far as I know there is no difference in the sound from one to the other except. Yes this is watch repair that always seems to be exceptions. If you have access to something with our oscilloscope and you can look at the waveform like the witschi we have at work I can see 4 waveforms unless of course you're running a higher frequency watch which I typically never do that I've noticed we can get six. So what I've noticed when you look at the oscilloscope is it's not an exact science there is variations for a whole variety of reasons. Then I have noticed were sometimes some of them the  loudness does seem to very but that's only occasionally I was interpreted to something else.

So as far as I know there should be no difference between exit and entry as far as audio sound goes for analysis but the above paragraph indicates that I'm wrong. But also confusingly indicates that I might be right which is why I'm confused

 

 

 

[nickelsilver]

VWatchie, I have to agree Frank's software is pretty badass. I'm using it with a Vibrograf mic and his preamp (simple, cheap, effective in-line thing) and calibrated with his again, simple, cheap, effective calibration stick. It tracks spot on with my 7500 buck witschi. But will also do odd escapements (cylinder, duplex, clocks).

 

My Witschi is on my bench but I often have an oddball piece in my computer that would have formerly been on a paper tape machine, as the PCTM just picks up everything. Not trying to make an advertisement but it's really good stuff. I don't know what version you have but the full one will tell you the heavy spot when doing dynamic poise- brilliant.

 

 

[xyzzy]

14 hours ago, JohnR725 said:

Notice it says the sounds are different and if we had the waveform which we don't it claims you can visually tell the difference. This by the way I was under the impression was impossible but it is quoted above as yes it's true.

Then notice it goes on to say it uses the sound to tell them apart rather than just assuming which once again implies an absolute.

Here's an image of the two beats.  It's pretty clear that they look different.

Suppose while messing with the beat adjustment the signal is lost for a bit and we lose many beats.  If we just keep track by assuming the beats alternate it might be out of sync when the signal comes back.  Perhaps the pallet that was getting white dots will now be assigned the yellow dots?  But by looking at the sound it should be possible tell the white beat from the yellow beat.

Now suppose we restart the app.  It will not have yet acquired the waveforms above and assigned them colors.  Starting fresh, it will acquire the above waves again and has the two clearly different shapes.  But which should be white and which yellow?  In this case, the louder peak is used as the distinguishing factor.  Above, the yellow is clearly louder than the white, so it should be reasonably consistent.  But often the difference is not so great, and the size of the peaks varies based on the escape tooth making the sound, the position of the rest of the train, amount of wind, and just randomly.  So it might not be consistent.

[JohnR725]

5 hours ago, xyzzy said:

It's pretty clear that they look different.

You really need to promote this as a feature of your software.

 

18 hours ago, nickelsilver said:

Frank's software is pretty badass

For such a bad ass software that I believe also comes in a free version where would I find it? For instance if I go to the search engine I use and I search for Watch Timing machine software The first thing I get is this link

http://www.watchoscope.com/

Continuing down the page couple a discussion groups hopelessly out of date making them almost worthless. Couple links to phone apps equally worthless.

We even get a link to the TG software that needs to promote its super interesting feature of discerning one sound from another as an absolute.

Then we get this video which is interesting I assume it's a review?

 

 

No matter what the video is interesting because it gets me to the thing that I was looking for as somebody asked me in a private discussion where to find the free software yes it does come in a free version and here's the link I don't suppose Frank has a website?

https://c.web.de/@337134913998293880/YuEh_TobSjaCfyBDix_1gg

[VWatchie]

12 hours ago, JohnR725 said:

I don't suppose Frank has a website?

https://c.web.de/@337134913998293880/YuEh_TobSjaCfyBDix_1gg

As soon as you reach something on the Internet that starts with "http" (Hyper Text Transfer Protocol) or "https" (s for Secure) it is a website, but perhaps you mean a website that would be more promotional and informative!?

Anyway, for those of you who would want to get hold of the pro version, get in touch with @praezis.

Oh, I really enjoyed the music by Chris Zabriskie (Readers! Do You Read?) in that video which shows version 2 of the PCTM. I believe the most recent version is 3.2.2.

Edited August 30, 2023 by VWatchie

[nickelsilver]

It's been a while since I got the PCTM software but I contacted Frank directly here by personal message. I don't know why he's not more visible on the net with such a cool product (and he does optical platforms for the Luthy hairspring vibrating tool too!). Just send him a message.

[praezis]

Thank you both for your kind words! 

Esp. from you VWatchie, as I was not sure if you like it.

Re the two beats: If there was always a higher and a lower beat, it would be in fact a way to tell both apart. But I doubt that things are so simple. 

John mentioned it: noise can be different after removing and mounting the watch on the mic again. 

Also there are watches with about the same loudness of both beats. Here the loudest beat will be just arbitrarily changing, and the testing person would be most confused (I can see the future desperate forum posts ).

Frank

N.B. VW already posted the link for download of free version PCTM and all info, thank you.

[VWatchie]

2 hours ago, praezis said:

I was not sure if you like it

It's a complete product and there are many features simply not found in other affordable devices/software and I'm still discovering its many features. So yes, I really like it!  Especially as a Swiss-made Witschi is out of my reach. The Chinese Weishi 1900 machine is OK I guess, but PCTM really takes it to another level.

I just tried the single noise (which I hadn't not noticed before) and got my "two colours"! Also, since I changed the OS language from Swedish to English I can now select the 50x magnification for the diagram by clicking the 5x magnification button while holding the Ctrl-key. Not sure it was the language change that made the difference, but anyway, it now works, although it's not a critical but more of a "nice-to-have" feature to me.

Edited August 30, 2023 by VWatchie

[JohnR725]

11 hours ago, VWatchie said:

As soon as you reach something on the Internet that starts with "http" (Hyper Text Transfer Protocol) or "https" (s for Secure) it is a website, but perhaps you mean a website that would be more promotional and informative!?

Anyway, for those of you who would want to get hold of the pro version, get in touch with @praezis.

Actually there's a whole bunch of problems.

Seeing as how my stories of personal experience do not go well I'll skip over the story aspect. from time to time I will scour the Internet for timing machines to see what's available. Using a variety of terms in foreign languages all sorts of interesting things can be dug up but like my experience the other day demonstrates finding Frank software is a bit difficult. So as you point out he has a website but his website lacks content for findability which means it can't be found. Yes I know it has PDFs and technically the search engine will find those but still without the proper keywords it's basically invisible.

Then often times on discussion groups or the other day a private message. Somebody noticed by reference to free timing machine software and they want to know where they are.  so usually because I'm lazy I used to an Internet search and typically I always find this which works out fine as it's one of the ones I recommend.

http://www.watchoscope.com/

So then there's the problem where is the link to Franks software? Yes I could go to the laptop which has it installed and see if I can find a reference to where the heck to find it or I can look at my bookmarks except that would've been a waste of time. The reason to be a waste of time is if I did a search in my bookmarks what exactly what search for if I can't find it online because I can't remember the initials how would I even find it in my bookmark folder for instance the bookmark title is this "Online-Speicher - PCTM free"  Oh but don't worry it's not a problem anymore. I have folders for watch stuff now so it's in the folder with other timing machine stuff so I won't have to go through this silly exercise again

but it still brings up my question of the program does look really nice and everybody gives a great reviews but how the heck would you ever find it if you didn't know it existed? What am I missing here?

Oh and it's not just Franks software that it's hard to find a lot of other programs the people doing their websites don't really understand findability.