D. I. Y. Watch Timing Machine.

[svorkoetter 138]

I'm a watch DIYer, having serviced two watches myself, partially using tools of my own making. One thing I've been working on over the last few months is my own PC based timing machine. I've finally got something working reliably, and thought I'd preview it here. My plan is to eventually write up a detailed article on my web site describing how to build the hardware, and providing a download for the software. I may also produce a "pro" version of the software with more features, to be sold for a reasonable fee.

 

Here's the hardware, all home made of course:

 

 

The next picture shows the timing trace from my "Black Lagoon", a modified Invicta 8926OB with a Seiko NH35A movement (yes, the amplitude is quite low and tends to fluctuate, topping out around 220 degrees - apparently this is par for the course for Seiko movements):

 

 

In this picture, the software is in scope mode, showing directly the ticks and tocks. Notice how with this movement, they alternate in loudness. Since the scope mode was run after the timing mode, the scope also shows the threshold that was used by the timing mode to detect the start of each tick.

 

 

And this is the settings window:

 

 

Here's a brief list of features so far (I have to resist creeping featuritis):

• Numerical display of daily rate error, balance amplitude, and beat error.
• Horizontal "paper tape" mode with sub-pixel vertical resolution that can show fluctuations that are too small to see on a typical stand-alone timing machine.
• User selectable averaging period.
• Rate determination using linear least squares, which gives a more meaningful reading than just naively averaging the rate.
• Automatic signal and noise level determination. There's nothing to adjust.
• Aggressive noise filtering in the software to keep the hardware simple.
• Automatic rate detection for common rates.
• Simple, uncluttered, fixed-size display (1024x600).
• Scope mode to aid in diagnosis of various watch faults.

Edited July 7, 2014 by svorkoetter

[bobm12 828]

And I just spend good money in a timegrapher machine!

 

Excellent work! Keep it up!.

 

I've heard of some software that can be downloaded (for a price) that converts the pc into a timegrapher. What you are doing is certainly friendlier than what I saw. I'll try and find a link so you can "compare notes". Good to have more of this in development, I hope someday prices can be more accessible with competition...plus the flexibility of common place home hardware: a pc.

 

By the way, are you referring to the Seiko class by Nick Hacko?

 

Robert

Edited July 7, 2014 by bobm12

[Geo 1630]

Hi svorkoetter, that's excellent work you've been doing of late. I really appreciate what you have done, but the computing technicalities are way over my head. Have you done a comparative test with a professional timegrapher? It would be interesting to see how the results compare. I look forward to further reports on the development.

Now I know why you've not been on the forum for a while!

[WillFly 622]

That's very clever - way over my head technically, but I can appreciate the work involved.

[anilv 756]

I like this very much... sign me up!

 

Anil.

[autowind 116]

I want the free version please as I am a senior citizen and I deserve it.

 

Wish I had the brains to figure it out,  well done.

[newguy 0]

Where can I get the complete kit? You should sell the whole thing as one. If you do, I'm a buyer

[svorkoetter 138]

There's no kit as of yet, but I may consider it if there's enough interest. But I'm not sure if it will be economically viable. It's the sort of thing you can build for well under $50 yourself by scrounging parts, but a low volume kit might end up costing as much as a Chinese timing machine.

[Geo 1630]

It's the sort of thing you can build for well under $50 yourself by scrounging parts, but a low volume kit might end up costing as much as a Chinese timing machine.

That's just what I was thinking Stefan, fine tuning for production and setting up costs would be very expensive and end up costing more than a Chinese one.

Edited July 7, 2014 by Geo

[Lee 59]

I'd pay for both a kit (to be constructed at home, rather than prebuilt, we all like to say how we made it!) & I'd pay for the software. I have a timing machine, but if this could be supplied cheap enough it would be worth having for the cool factor.

[Mark 1412]

This is excellent work.

 

I use Graham Baxter's software at the moment and it is very comprehensive, works with co-axials which is mainly why I purchased it. Graham will convert most old timing machine microphones for use with the software. What annoyed me before purchasing the software is that there was no alternatives to consider - and so I wish you luck in your development efforts.

 

If you haven't seen it yet then check it out and maybe you can glean some ideas: http://www.delphelectronics.co.uk/products.html

 

Also - feel free to send me a copy to compare against the timegrapher and Grahams software if you want.

[svorkoetter 138]

Construction of the preamplifier is simple enough that anyone with a few electronics assembly skills and tools could build it from scratch. I built it on stripboard ("Veroboard") so there's not even a circuit board to etch. A kit might be overkill.

 

The microphone is just simple woodworking, and it can be simplified even further by using a Korg clip-on piezo mic instead of building a mic stand.

 

I've tried Graham's software (in demo mode only, where it can only read a pre-recorded file) and it's quite impressive and full featured. I wanted to make something a little simpler to use though.

 

WRT, I'd be glad to send you a copy of the software as it is so far, since I'd like to get some feedback (and testing). If you already have a working mic, it should just work "out of the box". I just have a few loose ends to finish up (such as printing and PNG export support). Send me a message and we'll work out how to get you a copy.

[Mark 1412]

This is an exciting project. Couple of ideas for you:

 

Simple is good. It would be nice if you do not need to use a mouse (or have the choice not to have to use a mouse) - all functions controllable via keyboard shortcuts. That is one thing that annoys me about Grahams software, there are some keyboard shortcuts but the most basic of them has been omitted - the ability to start and stop timing with a keypress (like the space bar or something). Even better would be to enable the shortcuts to be configurable.

 

It sounds like a picky thing but really - messing with a mouse at your workbench is annoying :)

 

If you are programming this in C++ then I would recommend QT as this is multi-platform and you could compile it for PC, Mac or Linux as long as you don'y statically link the QT libs then you can redistribute without license.

[svorkoetter 138]

Yes, keyboard shortcuts were already on my to-do list. I don't want to bother with a configuration screen for them, so they'll probably be fixed. I like the idea of the Space bar for start/pause/resume. Perhaps Backspace for stop, Enter for settings, and the more-or-less standard Ctrl+P for printing?

 

It's written in Delphi for Windoze, but I could look into switching to Lazarus to get cross-platform portability. The problem is no matter what I use, the audio API will not be portable between platforms, so I'll still have to develop that separately for each (so it may never happen). Another possibility is Java, which does have a cross-platform audio interface (but that also probably isn't going to happen unless I find more pressing reasons to learn Java).

[Mark 1412]

I guess you are right.

 

Java does not seem desirable at all to me.

[bobm12 828]

svorkoetter, any possibility to integrate the audio using a java interface (like an external call, I don't know Delphi but maybe it can be done) and open source that part to get help inside that community (only for the audio)? That would make it portable and probably you won't have to worry about coding the audio itself. It's been a long time I used any programming language including java, so correct me if I'm wrong.

[bobm12 828]

I guess I was posting when you were Mark. I don't like java either but the practicality of just this interface may be worth looking into. I do believe that using java to process the audio might probably slow down the whole thing. (?)

[svorkoetter 138]

I think for now I'll stick to Windows. :-)

[Mark 1412]

I think for now I'll stick to Windows. :-)

Probably the wisest route

Sent from my iPhone using Tapatalk

[bobm12 828]

Yeah, it will simplify matters and java won't be an issue. In any case, good luck with the project and let us know of the outcome. I'll be a customer too when -- and if -- you say the word.

 

Robert

[svorkoetter 138]

Last night, I finished servicing my Poljot 3133-powered Sturmanskie chronograph, and had a chance to use my timing system "for real" for the first time:

 

 

Note that the annotations you see on the screenshot were added by hand, although I plan to add a feature to do that automatically.

 

I first used it to test the performance in four different positions, then used it while adjusting the watch, and then tested once more. Until now, I'd been recording the watch and then measuring ticks by hand in Audacity (an audio editing program). It was really nice to be able to have the performance displayed in real time.

 

This is the watch:

 

Edited July 9, 2014 by svorkoetter

[Mark 1412]

That is very awesome.

 

I was just thinking - if this was an iPad/Android app you would sell thousands of copies.

 

e.g. what a great use for a iPad mini or a nexus.

[svorkoetter 138]

I doubt I'd sell a lot either way, since one still has to build the preamp and microphone (the built in mics just aren't good enough for this sort of thing). I'm not sure how many people would be willing to do that, even though I think that anyone with the skills to need a timing machine would be able to put together a simple circuit.

Edited July 9, 2014 by svorkoetter

[hokavan 14]

Interesting works. I've paid 10$ for a software, named Kello on iPhone. It has similar principle however it takes long time to get the tick tock signal due to limited iPhone sound amplification.

Here is the result of my watches

[svorkoetter 138]

Kello is nice in principle, but many people have reported trouble getting it to "hear" the watch, and it also doesn't report anything other than the watch's daily rate.