Jump to content
AndyHull

What makes a quartz watch "tick"

Recommended Posts

If you have ever wondered what is going on inside one of those infernal kwarts thingamabobs, this article might be of interest.

https://zeptobars.com/en/read/Luch-quartz-wristwatch-IC-at-sc-tuning-fork

quartz-5x-50.jpg.913c0dc671e23c259f6fedd114deaab1.jpg

The full article is here -> https://3.14.by/en/read/quartz-wristwatch-watch-ic-decapsulation#fun

One point he skips over is the "16-stage divider", this doesn't divide by 16, it divides by 2^16 (or 32768) to give us our one second time reference.

I still prefer a well engineered mechanical watch, but you have got to admit, there is something quite captivating about that die shot.

 

Share this post


Link to post
Share on other sites
7 minutes ago, AndyHull said:

One point he skips over is the "16-stage divider", this doesn't divide by 16, it divides by 2^16 (or 32768) to give us our one second time reference.


I was thinking about that.
2^16 = 65536Hz ---> 32768Hz/65536Hz = 0.5Hz as time signal reference ,tick tock twice = 1s.

Share this post


Link to post
Share on other sites

Magic smoke, and angry pixies. That's what makes all electrickery work. 


BTW If you want to see one of these IC's yourself, you don't actually need fuming nitric acid to decap them (assuming you have no intention of ever getting the thing working again, and are not intent on probing signals or whatever).

The epoxy blob they sit under burns quite nicely, and silicon dies are pretty resilient to high temperature (within reason), so why not grab your microscope and indulge your inner pyromaniac. After all, who secretly doesn't desire to consign at least one of those annoying faulty cheap Chinese kwarts movements to a fiery end.
 

 

Share this post


Link to post
Share on other sites
3 minutes ago, HSL said:


I was thinking about that.
2^16 = 65536Hz ---> 32768Hz/65536Hz = 0.5Hz as time signal reference ,tick tock twice = 1s.

Drat, you spotted my error. I think you are correct, the movement may well both tick, and tock. I guess that is what all that other circuitry is for. You need to drive the stepper motor, and presumably that needs phase shifted signals, so my statement is slightly misleading.

Edited by AndyHull

Share this post


Link to post
Share on other sites
7 minutes ago, AndyHull said:

Magic smoke, and angry pixies. That's what makes all electrickery work. 


BTW If you want to see one of these IC's yourself, you don't actually need fuming nitric acid to decap them (assuming you have no intention of ever getting the thing working again, and are not intent on probing signals or whatever).

The epoxy blob they sit under burns quite nicely, and silicon dies are pretty resilient to high temperature (within reason), so why not grab your microscope and indulge your inner pyromaniac. After all, who secretly doesn't desire to consign at least one of those annoying faulty cheap Chinese kwarts movements to a fiery end.
 

 

  "electrickery" i like that.  how about digittrickery?  vin

Share this post


Link to post
Share on other sites

If you look at the die, things are slightly more complicated than just divide by 2, 16 times, there seem to be more than 16 division stages, and a tap off the ladder part way down... or maybe I am over thinking the problem. :D

Share this post


Link to post
Share on other sites

Externally, of course things are pretty simple. The chip drives a small magnetic coil, the oscillating field of which in turn causes a rotor to turn, (normally in the correct direction, and very precisely at the correct speed, except when subjected to the kind of torture shown above) which turns a gear.. and then the hip bone is then connected to the thigh bone etc in the manner of a conventional mechanical watch.

The trick is getting it all to work while sipping microscopic amounts of energy from the battery. If the battery voltage drops, either because it is getting exhausted and near the end of its useful life or because of corrosion on the contacts, or high resistance shorts on the PCB, or if the capacitance of the legs of the crystal oscillator are disturbed, again possibly due to dirt, corrosion, cracked PCB tracks, or other environmental screwups, then the precision of the oscillator will be affected, and the watch will run slow, fast, or not at all.

The low voltages involved, mean that relatively small amounts of corrosion, damp and dirt can have a relatively large effect.

So in summary, if it aint running, clean everything, and tidy up any dodgy looking solder joints and try again.

Share this post


Link to post
Share on other sites

.. oh.. one final point. In your enthusiasm to fiddle, poke and generally clean things, don't break the wires to the coil, they are a right pain the tail to fix back on, assuming there is enough wire left to actually re-connect them.

If you have managed to break a family heirloom in this way, I wish you good luck trying to repair. You can possibly get away with sacrificing a turn or two of the wire on the coil since there are quite a lot of turns, and the precision comes from the quartz oscillator, not the coil. The wire is enamel coated, so scrape the ends gently with fine emery paper or a very sharp blade, then flux them before attempting to  re-attach to the PCB. On the other hand, a lot of quartz movements are not worth the angst of even attempting to repair, so replace it, get out the blow torch, and proceed as shown above, to assign the offending item to the flames (for educational purposes of course).

Edited by AndyHull

Share this post


Link to post
Share on other sites

I’m not really a quartz fan, but respect their place in horological history.

A couple of years back I restored a 1980 ish Citizen / Bulova movement, which was wirth it because of the 9ct case it came in. The movement had a small potentiometer adjuste to tweak the calibration. Now I understand that any calibration errors are adjusted by changing data in an EEPROM - presumably those horrifically expensive Witchi machines have that capability built in.

The clever bit though is that whilst the crystal oscillates at 32767 Hz, you would think that calibration could be achieved by stepping the motor after one more or one less count of the crystal, but assuming the watch steps once per second, multiplying up, this would mean the adjustment could only be done in steps of 2.6 seconds per day.  So the calibration adjustment is typically set up to only affect one second in each minute - giving an adjustability of 1.3 seconds per month. More accurate watches use other tricks to give finer adjustment.

I’m writing this from memory, so if I have any details wrong, please forgive, but the concept surpiised me when I first read about it, and it is for this reason that a seemingly very accurate quartz watch must be used with geat care and averaged over a long period if you want to use the tick from it to calibrate a timing machine.

Share this post


Link to post
Share on other sites

Not to put too fine a point on the operation but actually, the 'divider' is a frequency counter or as I was taught when I received my US Navy electronics training back when knights still wore armor, a ring counter. Why was it called that? Because it literally counts every every frequency vibration and then overflows to the next counter in a ring pattern. It is a 16 position counter but the 16th position is the output so it's actually 2 ^15  which equals 32678. When the 15th register fills up, it sends the next pulse to the 16th position which sets a 1 and feeds it to an amplifier that drives the stepping motor one notch and zeroes the counter for the next cycle to begin all over again.

One of the inventions which made modern miniaturized quartz watch circuits possible was the invention of SMT (surface mount technology) 16 position ring counters. You are correct in that it is much more common (read less costly to manufacture) in a modern watch to find a circuit without a trimmer capacitor. The crystal instead is usually set to operate slightly higher than 32678 hz, then inhibit pulses are programmed from the cpu to delay the stepping motor thus correcting the accuracy. 

Quartz watches can get a bit technical where the electronics are concerned but they are truly marvels of engineering and inventive designs every bit as much as mechanical watches. Constant temperature helps immensely when it comes to stabilizing the frequency of a quartz oscillator. That's the main reason why the more you wear a quartz watch, the more accurate it becomes. Your body heat is the stabilizing, temperature compensation.

Share this post


Link to post
Share on other sites
On 1/13/2019 at 4:12 AM, jdrichard said:


Hey thanks. Figured with the amount of travel I do, I need to track at least 2 time zones at once.


Sent from my iPhone using Tapatalk Pro

Good excuse...I will try it out on the wife

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×