adjusting Timex quartz watch

[e***@efront.com]

I want to get an old Timex ladies quartz analog watch running again.

I know nothing about its history, but I find it consistently loses 2
minutes a day, with a new battery.

I see it has a screw on the movement that seems to be inviting itself
to be adjusted. It is brass, on a large plastic washer.

http://www.box.net/shared/fkur1iijcj

I think I remember correctly when I say that, in the old days, quartz
watches could be adjusted.

Is this such an adjusting screw?

If so, which way do I turn it to speed the watch up, and by roughly
how much of a turn?

[dAz]

Post by e***@efront.com
I want to get an old Timex ladies quartz analog watch running again.
I know nothing about its history, but I find it consistently loses 2
minutes a day, with a new battery.
I see it has a screw on the movement that seems to be inviting itself
to be adjusted. It is brass, on a large plastic washer.
http://www.box.net/shared/fkur1iijcj
I think I remember correctly when I say that, in the old days, quartz
watches could be adjusted.
Is this such an adjusting screw?
If so, which way do I turn it to speed the watch up, and by roughly
how much of a turn?

nope, the trimmer will only affect the rate in microseconds a day, your
watch needs a service, and if it turns out the fault is not mechanical
then the quartz oscillator may need to be replaced.

also keep in mind these timexs can be serviced but in the day even Timex
didn't service them, instead opting for a movement swap.

oh and back to that trimmer, if you don't believe me and decide to
fiddle, use no pressure in a downwards force, the trimmer is made of
ceramic, and will crack under light to moderate pressure.

[e***@efront.com]

Post by dAz

Post by e***@efront.com
I want to get an old Timex ladies quartz analog watch running again.
I know nothing about its history, but I find it consistently loses 2
minutes a day, with a new battery.
I see it has a screw on the movement that seems to be inviting itself
to be adjusted. It is brass, on a large plastic washer.
http://www.box.net/shared/fkur1iijcj
I think I remember correctly when I say that, in the old days, quartz
watches could be adjusted.
Is this such an adjusting screw?
If so, which way do I turn it to speed the watch up, and by roughly
how much of a turn?

nope, the trimmer will only affect the rate in microseconds a day, your
watch needs a service, and if it turns out the fault is not mechanical
then the quartz oscillator may need to be replaced.
also keep in mind these timexs can be serviced but in the day even Timex
didn't service them, instead opting for a movement swap.
oh and back to that trimmer, if you don't believe me and decide to
fiddle, use no pressure in a downwards force, the trimmer is made of
ceramic, and will crack under light to moderate pressure.

I believe you dAz. Thanks.

[d530]

Post by dAz
nope, the trimmer will only affect the rate in microseconds a day, your
watch needs a service, and if it turns out the fault is not mechanical
then the quartz oscillator may need to be replaced.

Oh no, condenser trimmer affects oscillator rate at least in seconds a day -
practically proven.
I doubt this Timex is an exception.

[Frank Adam]

Post by d530

Post by dAz
nope, the trimmer will only affect the rate in microseconds a day, your
watch needs a service, and if it turns out the fault is not mechanical
then the quartz oscillator may need to be replaced.

Oh no, condenser trimmer affects oscillator rate at least in seconds a day -
practically proven.
I doubt this Timex is an exception.

Yup, i can usually trim +-5-10 seconds on most watches that i've
bothered with. That is definitely per day, so Daz may have had a dim
moment when he said that. He is allowed to have that occasionally at
his age. ;-p

ps: It is a damn Timex after all. This is a great opportunity to buy a
decent watch. ;)

[dAz]

Post by Frank Adam

Post by d530

Post by dAz
nope, the trimmer will only affect the rate in microseconds a day, your
watch needs a service, and if it turns out the fault is not mechanical
then the quartz oscillator may need to be replaced.

Oh no, condenser trimmer affects oscillator rate at least in seconds a day -
practically proven.
I doubt this Timex is an exception.

Yup, i can usually trim +-5-10 seconds on most watches that i've
bothered with. That is definitely per day, so Daz may have had a dim
moment when he said that. He is allowed to have that occasionally at
his age. ;-p

I am not old enough to call you whippersnapper just yet ;)

Post by Frank Adam
ps: It is a damn Timex after all. This is a great opportunity to buy a
decent watch. ;)

hmm timex......... now where did I put my 4lb hammer...... if I am in a
good mood I'll ask the customer to take it off their arm first ;)

[dAz]

Post by d530

Post by dAz
nope, the trimmer will only affect the rate in microseconds a day, your
watch needs a service, and if it turns out the fault is not mechanical
then the quartz oscillator may need to be replaced.

Oh no, condenser trimmer affects oscillator rate at least in seconds a day -
practically proven.
I doubt this Timex is an exception.

irrelevant, I should have said milliseconds, but since the average range
is 5-6seconds for trimmers a 10 seconds a day +/- is out of range, mind
you the last time I touched a trimmer is so long ago that I cannot
remember it.

I did have a quartz timer once, and the times I actually needed to tweak
the trimmer was so little that it was a big waste of time in having a
quartz timer, I got rid of it and never replaced it, the manufacturers
must have thought so too since they removed the trimmer from the base
line of movements.

[d530]

Post by dAz
I did have a quartz timer once, and the times I actually needed to tweak
the trimmer was so little that it was a big waste of time in having a
quartz timer, I got rid of it and never replaced it, the manufacturers
must have thought so too since they removed the trimmer from the base line
of movements.

Yes, however I miss condenser trimmers in quartz movements.
I think the main reason of removing trimmers from them is cutting the costs
of production. Heard somewhere that trimmer considerably increases cost of
movement.
So without trimmers we have:
A) good movements due to better tolerancies of electronic components - well
known manufacturers
B) cheap no name movements with bad accuracy (loosing or gaining more than 1
s/day is not uncommon)


In case A) I suspect that there is a form of capacitancies on circuit boards
of movements. According to frequency of oscillations manufacturers cut
certain copper paths on circuit boards thus changing accuracy of movements.
Not long ago I saw such cut conducting paths in very basic Casio movement.
Such regulation (as irreversible) is conducted only once by manufacturer.

In case B) we have watches or stopwatches sold for 1 Euro but very
inaccurate.
As an example: once I bought in Lidl market nice stopwatch (worth 4 Euro).
It was gaining more than 1 s/day in watch mode. It was too much for me
bearing in mind that except poor accuracy this stopwatch had very reliable
pushbuttons (you probably know what I mean - cheap 1 Euro stopwatches are so
unreliable. When you press start or stop pushbuttons they very often behave
like you pressed them twice or dont react at all - depends on quality of
contacts under pushbuttons.
I decided to improve it's accuracy. Certainly there was no trimmer inside,
but there was some room to take ceramic condenser.
And because I haven't had small trimmer handy I started to soldering in
constant capacity condensers.
With 3,3 pF stopwatch was loosing 1 s per 3 days. Next tried 2,2 pF and
achieved satisfying accuracy - loosing 1 s/eleven days at room temperature.
For any hobbist wanting to change frequency of quartz movements lacking
trimmer: movement has to be fast not slow because adding capacity slows
frequency of oscillations. Trimmer condenser or constant capacity ceramic
condenser should be soldered between one of pins of quartz oscillator and
"+" pole of supply.

Hope my English was undertood - it's not my native language.

[Jack Denver]

Yes, your English is fine, much better than my Polish, dziekuje .

Eliminating the trimmer did not mean that all adjustment facility was lost.
Modern movements (even cheap ones) usually have a facility programmed into
the chip to skip X number of cycles PER SECOND (where X is say 0 to 255).
The crystals are all trimmed to run slightly fast and at the factory the #
of skipped beats is set, usually permanently. This can be thru some kind of
programming device (the chip has a small area of PROM to hold this value) or
the circuit board sometimes has 8 traces that can be physically cut or
burned with a laser in different combinations to set the value. Sometimes
the value can be reset later with a special programming machine (which
watchmakers never have - a factory service center might), or the lines can
be resoldered and re-cut. Other times the value is burned once and cannot be
changed later.

If you look at modern movements on an oscilloscope you would see, for
example , wave running at 32,900 HZ for most of a second, then wave going
idle for 132 milliseconds, then back to the 32,900Hz wave, then idle for 132
ms, etc.

A 1 Euro stopwatch is another story - when your whole budget is 25 cents
(retail is usually 400% of manufacturing cost) every fraction of a penny
counts so they use the cheapest crystals they can find - usually you are
timing say a foot race that lasts a few minutes or maybe a couple of hours
at most, so losing 1 sec/ in 3 days is nothing.

Post by d530

Post by dAz
I did have a quartz timer once, and the times I actually needed to tweak
the trimmer was so little that it was a big waste of time in having a
quartz timer, I got rid of it and never replaced it, the manufacturers
must have thought so too since they removed the trimmer from the base
line of movements.

Yes, however I miss condenser trimmers in quartz movements.
I think the main reason of removing trimmers from them is cutting the
costs of production. Heard somewhere that trimmer considerably increases
cost of movement.
A) good movements due to better tolerancies of electronic components -
well known manufacturers
B) cheap no name movements with bad accuracy (loosing or gaining more than
1 s/day is not uncommon)
In case A) I suspect that there is a form of capacitancies on circuit
boards of movements. According to frequency of oscillations manufacturers
cut certain copper paths on circuit boards thus changing accuracy of
movements. Not long ago I saw such cut conducting paths in very basic
Casio movement. Such regulation (as irreversible) is conducted only once
by manufacturer.
In case B) we have watches or stopwatches sold for 1 Euro but very
inaccurate.
As an example: once I bought in Lidl market nice stopwatch (worth 4 Euro).
It was gaining more than 1 s/day in watch mode. It was too much for me
bearing in mind that except poor accuracy this stopwatch had very reliable
pushbuttons (you probably know what I mean - cheap 1 Euro stopwatches are
so unreliable. When you press start or stop pushbuttons they very often
behave like you pressed them twice or dont react at all - depends on
quality of contacts under pushbuttons.
I decided to improve it's accuracy. Certainly there was no trimmer inside,
but there was some room to take ceramic condenser.
And because I haven't had small trimmer handy I started to soldering in
constant capacity condensers.
With 3,3 pF stopwatch was loosing 1 s per 3 days. Next tried 2,2 pF and
achieved satisfying accuracy - loosing 1 s/eleven days at room
temperature.
For any hobbist wanting to change frequency of quartz movements lacking
trimmer: movement has to be fast not slow because adding capacity slows
frequency of oscillations. Trimmer condenser or constant capacity ceramic
condenser should be soldered between one of pins of quartz oscillator and
"+" pole of supply.
Hope my English was undertood - it's not my native language.

[Frank Adam]

On Mon, 20 Apr 2009 15:05:14 -0400, "Jack Denver"

Post by Jack Denver
Yes, your English is fine, much better than my Polish, dziekuje .
Eliminating the trimmer did not mean that all adjustment facility was lost.
Modern movements (even cheap ones) usually have a facility programmed into
the chip to skip X number of cycles PER SECOND (where X is say 0 to 255).
The crystals are all trimmed to run slightly fast and at the factory the #
of skipped beats is set, usually permanently. This can be thru some kind of
programming device (the chip has a small area of PROM to hold this value) or
the circuit board sometimes has 8 traces that can be physically cut or
burned with a laser in different combinations to set the value. Sometimes
the value can be reset later with a special programming machine (which
watchmakers never have - a factory service center might), or the lines can
be resoldered and re-cut. Other times the value is burned once and cannot be
changed later.
If you look at modern movements on an oscilloscope you would see, for
example , wave running at 32,900 HZ for most of a second, then wave going
idle for 132 milliseconds, then back to the 32,900Hz wave, then idle for 132
ms, etc.

My timer picks up most ETA circuits running at up to +5 secs/day. Be
it on the 1, 4 or even 10 second gate, this does not change, which one
would expect should, since when timing analogs, we pick up the pulses,
not the crystal itself. And since the pulses are connected to the
seconds hand...hm.. see my problem ? :-)

The mainstream ETAs of the past that i've tested(Swiss or other),
pretty much all shown a similar base rate. Those would be the 955s,
956s and similar, the castrated ones of the same models with trimmers,
which could be brought to +0 and acted just like it on bench testing.
Admittedly, i haven't bothered to bench test a trimmer-less ETA,
mainly because there is no reason to.
The newer el cheapo ETAs(FOx.xxx) show a much better base rate, down
to a second or better. Seikos and other TMIs will be religiously below
0.3 and Miyotas(even the A$4.00 2035s) hover in that region too.
Yes, some movements have had the factory adjustments, but i'm not
buying that these are an *intentional fine adjustment* in the factory.
Rather they are necessary ones, so the chips can compensate for the
particular crystal fitted. Crystals are not equal. The cheaper the
crystals, the wider tolerances will shift, which has to be addressed
at the chip. That's my theory and i'm sticking by it. :)

[Jack Denver]

Post by Frank Adam
On Mon, 20 Apr 2009 15:05:14 -0400, "Jack Denver"
My timer picks up most ETA circuits running at up to +5 secs/day. Be
it on the 1, 4 or even 10 second gate, this does not change, which one
would expect should, since when timing analogs, we pick up the pulses,
not the crystal itself. And since the pulses are connected to the
seconds hand...hm.. see my problem ? :-)

The problem is in your equipment - unless something is drastically wrong no
quartz watch runs +5 on the wrist - .5 sec is considered the outside
permitted variation.

Post by Frank Adam
Yes, some movements have had the factory adjustments, but i'm not
buying that these are an *intentional fine adjustment* in the factory.
Rather they are necessary ones, so the chips can compensate for the
particular crystal fitted. Crystals are not equal. The cheaper the
crystals, the wider tolerances will shift, which has to be addressed
at the chip. That's my theory and i'm sticking by it. :)

The nature of the fine adjustment feature of the chips is that they can
subtract (skip) cycles but never add, so the circuit is intentionally set up
to resonate slightly fast (this is a function of not just the crystal but
the load capacitor) and then the fine adjustment is added to bring the
effective rate down by skipping cycles . I'm not making this up - this is
how modern trimmerless movements are designed. I think that somehow your
machine is somehow picking up the unadjusted frequency. Does the tape show
any jitter? As I said before a scope would show this very clearly . I
believe that in some chips the adjustment occurs only 1x/ MINUTE so you'd
have 59 regular (fast) pulse on the stepper and one slow pulse - if the
watch is running 5 sec/day fast, then the one "slow" plus needs to be only
.003 secs slower than the rest .

[Frank Adam]

On Mon, 20 Apr 2009 20:51:52 -0400, "Jack Denver"

Post by Jack Denver

Post by Frank Adam
On Mon, 20 Apr 2009 15:05:14 -0400, "Jack Denver"
My timer picks up most ETA circuits running at up to +5 secs/day. Be
it on the 1, 4 or even 10 second gate, this does not change, which one
would expect should, since when timing analogs, we pick up the pulses,
not the crystal itself. And since the pulses are connected to the
seconds hand...hm.. see my problem ? :-)

The problem is in your equipment - unless something is drastically wrong no
quartz watch runs +5 on the wrist - .5 sec is considered the outside
permitted variation.

Equipment is fine, i was shooting from the hip without much thinking.
Ever wished you could pull part of a post back ? :)

Post by Jack Denver

Post by Frank Adam
Yes, some movements have had the factory adjustments, but i'm not
buying that these are an *intentional fine adjustment* in the factory.
Rather they are necessary ones, so the chips can compensate for the
particular crystal fitted. Crystals are not equal. The cheaper the
crystals, the wider tolerances will shift, which has to be addressed
at the chip. That's my theory and i'm sticking by it. :)

The nature of the fine adjustment feature of the chips is that they can
subtract (skip) cycles but never add, so the circuit is intentionally set up
to resonate slightly fast (this is a function of not just the crystal but
the load capacitor) and then the fine adjustment is added to bring the
effective rate down by skipping cycles . I'm not making this up - this is

I know you're not. I was referring to the chops on the circuit board
though, which has to do with the frequency as prescribed by the quartz
and i don't see that as a fine tune, rather as a required adjustment.

Post by Jack Denver
how modern trimmerless movements are designed. I think that somehow your
machine is somehow picking up the unadjusted frequency. Does the tape show
any jitter? As I said before a scope would show this very clearly . I
believe that in some chips the adjustment occurs only 1x/ MINUTE so you'd
have 59 regular (fast) pulse on the stepper and one slow pulse - if the
watch is running 5 sec/day fast, then the one "slow" plus needs to be only
.003 secs slower than the rest .

That explains why i would've missed it, i don't actually time the
watches, just use the thing to check for pulses..

Just doing a test with an ETA 956.112 (Longines actually)
Stuck on the 2 second gate so i can see all changes. It sits on
+5.78-79 with an "o/f" at the 60th cycle. O/f means that the timer
can't decypher the step into any of it's patterns so it pulls it's
display.
That would be the correction cycle then, but it is easily missed,
because a weak or missed pulse will also give the o/f.

[dAz]

Post by Frank Adam
On Mon, 20 Apr 2009 15:05:14 -0400, "Jack Denver"

Post by Jack Denver
Yes, your English is fine, much better than my Polish, dziekuje .
Eliminating the trimmer did not mean that all adjustment facility was lost.
Modern movements (even cheap ones) usually have a facility programmed into
the chip to skip X number of cycles PER SECOND (where X is say 0 to 255).
The crystals are all trimmed to run slightly fast and at the factory the #
of skipped beats is set, usually permanently. This can be thru some kind of
programming device (the chip has a small area of PROM to hold this value) or
the circuit board sometimes has 8 traces that can be physically cut or
burned with a laser in different combinations to set the value. Sometimes
the value can be reset later with a special programming machine (which
watchmakers never have - a factory service center might), or the lines can
be resoldered and re-cut. Other times the value is burned once and cannot be
changed later.
If you look at modern movements on an oscilloscope you would see, for
example , wave running at 32,900 HZ for most of a second, then wave going
idle for 132 milliseconds, then back to the 32,900Hz wave, then idle for 132
ms, etc.

My timer picks up most ETA circuits running at up to +5 secs/day. Be
it on the 1, 4 or even 10 second gate, this does not change, which one
would expect should, since when timing analogs, we pick up the pulses,
not the crystal itself. And since the pulses are connected to the
seconds hand...hm.. see my problem ? :-)

maybe your machine needs to be recalibrated ;)

ETA service agent: "yes see!, all our vatches keep perfek time"

you know when quartz first came out, you had to adjust them to somewhere
in the 0.1 0.15 fast per day range to compensate for when the owner
places the watch on their arm, and when the watch warms up to body temp
it will lose 0.1seconds per day, so in theory the watch would in the 0
to 0.05seconds per day fast, doesn't work out that way of course.

actually I think 90% of my customers would not notice or care if the
watch was out 5 seconds a day, ladies watches with their 20sec pulse
motors and no seconds hand, fashion watches even with seconds hand and
no marks on the dial except for a dot at the 12.....

Post by Frank Adam
The mainstream ETAs of the past that i've tested(Swiss or other),
pretty much all shown a similar base rate. Those would be the 955s,
956s and similar, the castrated ones of the same models with trimmers,
which could be brought to +0 and acted just like it on bench testing.
Admittedly, i haven't bothered to bench test a trimmer-less ETA,
mainly because there is no reason to.
The newer el cheapo ETAs(FOx.xxx) show a much better base rate, down
to a second or better. Seikos and other TMIs will be religiously below
0.3 and Miyotas(even the A$4.00 2035s) hover in that region too.
Yes, some movements have had the factory adjustments, but i'm not
buying that these are an *intentional fine adjustment* in the factory.
Rather they are necessary ones, so the chips can compensate for the
particular crystal fitted. Crystals are not equal. The cheaper the
crystals, the wider tolerances will shift, which has to be addressed
at the chip. That's my theory and i'm sticking by it. :)

I don't own a quartz timer since I got rid of the citizen timer 25years
ago, never missed it because if the movements they make today cannot be
adjusted then there is no point in having one.

I test the circuits, put a battery in, check over the next day or two,
if they show the right time, out the door, if they don't then it's
service time or movement swap.

[d530]

Post by Jack Denver
Yes, your English is fine, much better than my Polish, dziekuje .

It's very kind of you, dziekuje :)

And thanks for explanation of modern, digital regulation of accuracy of
quartz movements.
I've stopped at former, analog regulation.

[dAz]

Post by d530

Post by dAz
I did have a quartz timer once, and the times I actually needed to tweak
the trimmer was so little that it was a big waste of time in having a
quartz timer, I got rid of it and never replaced it, the manufacturers
must have thought so too since they removed the trimmer from the base line
of movements.

Yes, however I miss condenser trimmers in quartz movements.
I think the main reason of removing trimmers from them is cutting the costs
of production. Heard somewhere that trimmer considerably increases cost of
movement.
A) good movements due to better tolerancies of electronic components - well
known manufacturers
B) cheap no name movements with bad accuracy (loosing or gaining more than 1
s/day is not uncommon)
In case A) I suspect that there is a form of capacitancies on circuit boards
of movements. According to frequency of oscillations manufacturers cut
certain copper paths on circuit boards thus changing accuracy of movements.
Not long ago I saw such cut conducting paths in very basic Casio movement.
Such regulation (as irreversible) is conducted only once by manufacturer.

as to what Jack was saying there are no capacitors used to regulate the
movements, those cut tracks lead directly to the divider circuit, the
circuit is measured in the factory and the closes rate is determining
according to a preset tolerance and one or more track are cut, in theory
you could solder up the tracks and recut the tracks, but it is pretty
clunky it can only be adjusted in 255 steps so you may have it gaining
by 0.2seconds per day and the next step is either -0.1 sec or 0.5+ secs
per day nothing in between.

the other method is an eeprom which can be reprogramed by a very
expensive Witchi timer and it gets more complex, some of the programable
circuits actually adjust the speed or rate of the stepping motor from
minute to minute, so over the minute or so the pulse to the motor maybe
be spread apart or closer together to achieve a very accurate rate like
10seconds a year, with such a movement you cannot simply measure the
quartz frequency and tweak a trimmer, it has to be measured over several
minutes and then wipe and reprogram the eeprom.

ok the elimination of the ceramic trimmer was done for costs because it
was a pricey component, it was easily damaged by ham fisted watchmakers
who treated it like a normal watch screw (actually it didn't take that
much to break it), it might coast the maker 5cents each, but over a
million movements that is a big saving.

the other main problem with these trimmers is they were affected by
humidity, sure using solid state caps would solve that problem but the
rate adjustment is too crude to be practical, hence the cut tracks or
better the once only programmed chip or an eeprom on the top grade quartz.

you can see an example here
Loading Image...
the older movement on the left is sort of a transitional one, there is a
space for a trimmer but it never had one fitted, so one of the other
methods would have been used to adjust the rate at the factory, earlier
versions of this movement did have trimmers, later versions the circuit
was updated so there was no space for a trimmer.

the circuit on the right shows a modern type with the 5 cutable tracks
to the left of the microchip.

Post by d530
In case B) we have watches or stopwatches sold for 1 Euro but very
inaccurate.
As an example: once I bought in Lidl market nice stopwatch (worth 4 Euro).
It was gaining more than 1 s/day in watch mode. It was too much for me
bearing in mind that except poor accuracy this stopwatch had very reliable
pushbuttons (you probably know what I mean - cheap 1 Euro stopwatches are so
unreliable. When you press start or stop pushbuttons they very often behave
like you pressed them twice or dont react at all - depends on quality of
contacts under pushbuttons.
I decided to improve it's accuracy. Certainly there was no trimmer inside,
but there was some room to take ceramic condenser.
And because I haven't had small trimmer handy I started to soldering in
constant capacity condensers.
With 3,3 pF stopwatch was loosing 1 s per 3 days. Next tried 2,2 pF and
achieved satisfying accuracy - loosing 1 s/eleven days at room temperature.
For any hobbist wanting to change frequency of quartz movements lacking
trimmer: movement has to be fast not slow because adding capacity slows
frequency of oscillations. Trimmer condenser or constant capacity ceramic
condenser should be soldered between one of pins of quartz oscillator and
"+" pole of supply.

well in practical terms the stopwatch simply doesn't need to be more
accurate than that, I have a mechanical stopwatch with a 3 seconds dial
and 2 minutes on the minute recorder, so in use it is for timing
something up to 2 minutes, this stopwatch could be 30seconds a day out
and it will make no difference to the result, I cannot put it on the
timer anyway since the 360,000bph train is so noisy it overloads the timer.

Post by d530
Hope my English was undertood - it's not my native language.

yes, no probs :)

[Jack Denver]

AFAIK, the adjustment is usually applied 1x/minute (or maybe in some chips
there is a flag where you can set it to either 1 or 2x /minute, not per
second - if you remove even 1 cycle/second out of 32,768 that would add up
to 2.63 seconds/day which is too coarse an increment. Rather the skip cycles
(1 to 255) are applied only 1x/ MINUTE which gives you a .04 sec granularity
and a maximum 11 secs trim if it is an 8 bit register. 59 regular steps, 1
slow step where the feed to the dividers goes idle for X beats, 59 regular
steps, 1 slow step. This explains why Frank's timing machine is confused. I
don't know of any that vary the stepper from minute to minute.

Post by dAz

Post by d530

Post by dAz
I did have a quartz timer once, and the times I actually needed to tweak
the trimmer was so little that it was a big waste of time in having a
quartz timer, I got rid of it and never replaced it, the manufacturers
must have thought so too since they removed the trimmer from the base
line of movements.

Yes, however I miss condenser trimmers in quartz movements.
I think the main reason of removing trimmers from them is cutting the
costs of production. Heard somewhere that trimmer considerably increases
cost of movement.
A) good movements due to better tolerancies of electronic components -
well known manufacturers
B) cheap no name movements with bad accuracy (loosing or gaining more
than 1 s/day is not uncommon)
In case A) I suspect that there is a form of capacitancies on circuit
boards of movements. According to frequency of oscillations manufacturers
cut certain copper paths on circuit boards thus changing accuracy of
movements. Not long ago I saw such cut conducting paths in very basic
Casio movement. Such regulation (as irreversible) is conducted only once
by manufacturer.

as to what Jack was saying there are no capacitors used to regulate the
movements, those cut tracks lead directly to the divider circuit, the
circuit is measured in the factory and the closes rate is determining
according to a preset tolerance and one or more track are cut, in theory
you could solder up the tracks and recut the tracks, but it is pretty
clunky it can only be adjusted in 255 steps so you may have it gaining by
0.2seconds per day and the next step is either -0.1 sec or 0.5+ secs per
day nothing in between.
the other method is an eeprom which can be reprogramed by a very expensive
Witchi timer and it gets more complex, some of the programable circuits
actually adjust the speed or rate of the stepping motor from minute to
minute, so over the minute or so the pulse to the motor maybe be spread
apart or closer together to achieve a very accurate rate like 10seconds a
year, with such a movement you cannot simply measure the quartz frequency
and tweak a trimmer, it has to be measured over several minutes and then
wipe and reprogram the eeprom.
ok the elimination of the ceramic trimmer was done for costs because it
was a pricey component, it was easily damaged by ham fisted watchmakers
who treated it like a normal watch screw (actually it didn't take that
much to break it), it might coast the maker 5cents each, but over a
million movements that is a big saving.
the other main problem with these trimmers is they were affected by
humidity, sure using solid state caps would solve that problem but the
rate adjustment is too crude to be practical, hence the cut tracks or
better the once only programmed chip or an eeprom on the top grade quartz.
you can see an example here
http://members.optushome.com.au/daz7/horological/WatchCircuits.jpg
the older movement on the left is sort of a transitional one, there is a
space for a trimmer but it never had one fitted, so one of the other
methods would have been used to adjust the rate at the factory, earlier
versions of this movement did have trimmers, later versions the circuit
was updated so there was no space for a trimmer.
the circuit on the right shows a modern type with the 5 cutable tracks to
the left of the microchip.

Post by d530
In case B) we have watches or stopwatches sold for 1 Euro but very
inaccurate.
As an example: once I bought in Lidl market nice stopwatch (worth 4
Euro). It was gaining more than 1 s/day in watch mode. It was too much
for me bearing in mind that except poor accuracy this stopwatch had very
reliable pushbuttons (you probably know what I mean - cheap 1 Euro
stopwatches are so unreliable. When you press start or stop pushbuttons
they very often behave like you pressed them twice or dont react at all -
depends on quality of contacts under pushbuttons.
I decided to improve it's accuracy. Certainly there was no trimmer
inside, but there was some room to take ceramic condenser.
And because I haven't had small trimmer handy I started to soldering in
constant capacity condensers.
With 3,3 pF stopwatch was loosing 1 s per 3 days. Next tried 2,2 pF and
achieved satisfying accuracy - loosing 1 s/eleven days at room temperature.
For any hobbist wanting to change frequency of quartz movements lacking
trimmer: movement has to be fast not slow because adding capacity slows
frequency of oscillations. Trimmer condenser or constant capacity ceramic
condenser should be soldered between one of pins of quartz oscillator and
"+" pole of supply.

well in practical terms the stopwatch simply doesn't need to be more
accurate than that, I have a mechanical stopwatch with a 3 seconds dial
and 2 minutes on the minute recorder, so in use it is for timing something
up to 2 minutes, this stopwatch could be 30seconds a day out and it will
make no difference to the result, I cannot put it on the timer anyway
since the 360,000bph train is so noisy it overloads the timer.

Post by d530
Hope my English was undertood - it's not my native language.

yes, no probs :)

[dAz]

Post by Jack Denver
AFAIK, the adjustment is usually applied 1x/minute (or maybe in some
chips there is a flag where you can set it to either 1 or 2x /minute,
not per second - if you remove even 1 cycle/second out of 32,768 that
would add up to 2.63 seconds/day which is too coarse an increment.
Rather the skip cycles (1 to 255) are applied only 1x/ MINUTE which
gives you a .04 sec granularity and a maximum 11 secs trim if it is an 8
bit register. 59 regular steps, 1 slow step where the feed to the
dividers goes idle for X beats, 59 regular steps, 1 slow step. This
explains why Frank's timing machine is confused. I don't know of any
that vary the stepper from minute to minute.

I know those temperature compensation movements needed to be tested
over a 6 minute period or something like that to get an average of the
overall rate because of the variable pulse rate motors, apart from the
fact the eeprom can only programmed by one machine that I know of these
movements will look to be out of whack on a normal timer, specially if
you pick a 10second gate.

not quite the thing I was talking about but it does show how they can
vary the motor pulse and duty cycles, this is more for when the movement
is driving a date wheel around or when the oil is starting to thicken

http://www.nxp.com/#/pip/pip=[pip=PCA2003_1]|pp=[t=pip,i=PCA2003_1]
http://www.nxp.com/acrobat_download/datasheets/PCA2003_1.pdf

[Jack Denver]

Section 7.3 explains the time correction method for this chip, which is more
or less what I said before.

Post by Jack Denver
AFAIK, the adjustment is usually applied 1x/minute (or maybe in some
chips there is a flag where you can set it to either 1 or 2x /minute, not
per second - if you remove even 1 cycle/second out of 32,768 that would
add up to 2.63 seconds/day which is too coarse an increment. Rather the
skip cycles (1 to 255) are applied only 1x/ MINUTE which gives you a .04
sec granularity and a maximum 11 secs trim if it is an 8 bit register. 59
regular steps, 1 slow step where the feed to the dividers goes idle for X
beats, 59 regular steps, 1 slow step. This explains why Frank's timing
machine is confused. I don't know of any that vary the stepper from
minute to minute.

I know those temperature compensation movements needed to be tested over
a 6 minute period or something like that to get an average of the overall
rate because of the variable pulse rate motors, apart from the fact the
eeprom can only programmed by one machine that I know of these movements
will look to be out of whack on a normal timer, specially if you pick a
10second gate.
not quite the thing I was talking about but it does show how they can vary
the motor pulse and duty cycles, this is more for when the movement is
driving a date wheel around or when the oil is starting to thicken
http://www.nxp.com/#/pip/pip=[pip=PCA2003_1]|pp=[t=pip,i=PCA2003_1]
http://www.nxp.com/acrobat_download/datasheets/PCA2003_1.pdf

[d530]

Post by dAz
well in practical terms the stopwatch simply doesn't need to be more
accurate than that, I have a mechanical stopwatch with a 3 seconds dial

Sure. As a stopwatch such accuracy is sufficient.

But the stopwatch I improved, as a quartz stopwatch, has additional modes of
operation: timer, alarms and of course digital watch.

Let's consider using such device as alarm - reminding you of something
rather important. In such case the accuracy gains importance.

Besides, as analytical chemist I'm rather accurate and precise, and that's
what I expect of measuring devices.

BTW - There is an old sentence: if you have two watches you don't know what
time it is - still applicable to average people.

A coworker from my lab deliberately sets her watch a minute faster, I don't
think she cares of watch accuracy.
So cheap mass produced watches don't need to be very accurate.

[dAz]

Post by d530
A coworker from my lab deliberately sets her watch a minute faster, I don't
think she cares of watch accuracy.
So cheap mass produced watches don't need to be very accurate.

I have a friend who runs his watch 10mins fast, and he is a bus
driver... figure that one out :)

I can never make any sense of people that do that, I mean if you set
your watch 10mins fast, you know it is 10mins fast and visually subtract
10mins from the time displayed, ok, so in theory you are suppose to meet
someone at ten, and your watch shows ten o'clock you know you are 10mins
early or have 10mins to get to where you are supposed to be.

why can't you look at your watch and see you have 10mins to go and just
get a move on, that's why I like analogue watches because you can
visually see how much time you have, how much is passed and what the
current time is.

my main problem with people who set their watch fast is if ever that
watch gets set to the correct time, then they will have a problem of
being late all the time :)

[Jerry G.]

Post by e***@efront.com
I want to get an old Timex ladies quartz analog watch running again.
I know nothing about its history, but I find it consistently loses 2
minutes a day, with a new battery.
I see it has a screw on the movement that seems to be inviting itself
to be adjusted. It is brass, on a large plastic washer.
http://www.box.net/shared/fkur1iijcj
I think I remember correctly when I say that, in the old days, quartz
watches could be adjusted.
Is this such an adjusting screw?
If so, which way do I turn it to speed the watch up, and by roughly
how much of a turn?

If you have a quartz watch loosing two minutes per day this is a big problem. The original spec for most Timex quartz watches is +- 15 seconds per month. Most are keeping better than +-6 to +-10 seconds per month when worn daily.

With the older type better quality quartz watches there was a trimmer capacitor to allow for adjusting the regulation. The newer ones are using software calibration for regulation during the time of manufacture. The quartz oscillator is using what is called a VCO (Voltage Control Oscillator) system to generate the frequency that is integrated as part of a PLL (Phase Lock Loop) system. The PLL system is able to self verify itself by determining its output in relation to the programmed standard. This makes for better stability. This is why the newer quality quartz watches are much more stable and average out with better accuracy over the long term.

With the trimmer type watches, I found with most watches turning the trimmer CW slowed it down, and CCW sped it up. I also found some to be the opposite. This trimmer adjustment can be critical. It is best to use the proper calibration instrument. The instrument is expensive. How I do it is open the watch, and give the trimmer a crack of a turn. I then set the watch to an exact time standard. I use http://www.time.gov/ as an exact time reference.

I come back a day later and see if the watch is still fast or slow. I then crack the trimmer again in the proper direction to increase or decrease the speed of the watch. Eventually after a number of trials I get the watch to where I want it.

When I have the watch sitting for the next check I leave the lit loosely over the watch. The lid not being on at all may have some effect (stray capacitance).

The trimmer capacitor is in a resonant circuit that is in parallel to the feedback circuit with the crystal. The trimmer is a variable capacitor. It is shifting the phase of the feedback in the resonant circuit between the output and input of the oscillator circuit. This feedback is causing a phase shift in the circuit causing the crystal to be pulled slightly off its resonant point. This is called pulling the crystal.

Normally, the oscillator circuit is designed that when the trimmer is at 50% of its range the frequency should be ideal. In the real world it may not be because of inherent tolerances of the components involved. The range of this trimmer if I recall correctly is about 5pf to about 120pf. The pf is pico-Farads which is 1 X -12F (Farad).

From the manufacture of trimmer capacitors (they can come in many ranges of value depending on the requirement from their client) are set so that CW increases the pf value, and CCW lowers the pf value. The client assembling the device can set the trimmer to be opposite if the want to. The trimmer can be rotated continuously 360 degrees.

There are two types of resonance circuit design that can be used for any oscillator. Parallel or series resonant. The direction of trimmer movement to change the frequency can be determined by where it is being used in the feedback circuit for the oscillator, and if it is a parallel or series resonant.

Anyway... I can go on and on about oscillator design. I thought I would write down some insight as a contribution to the groups.