It is currently 28 Mar 2024, 10:22


Suitable 32.768KHz oscillator for Pulsar Modules...read on..

For electronic related stuff like module repair, silver epoxy fixes etc.
  • Author
  • Message
Offline
User avatar

retroleds

Guru

Guru

  • Posts: 3636
  • Joined: 04 Feb 2006, 10:34
  • Location: Surrounded by hicks and sticks (farms and woods) - Michigan,USA

Re: : Suitable 32.768KHz oscillator for Pulsar Modules...rea

Post07 Jul 2010, 22:47

azimuth_pl wrote:any comments are welcome.
however... almost every component will change values in different temps.
It would be a rare part that changes resistance by a factor of 100, over a 6-7 degree c. temperature change. That would be, by human body temperature standards, a pretty good thermisitor. :!:
http://www.retroleds.com - Sales of vintage LED, LCD, analog watches, parts and gadgets - repair tutorials & tips
Nov. 2022 - back in business!! BItter divorce is in home stretch, come grabs some great deals, I had to open the safe . . . damn attorneys. piss.
Offline
User avatar

azimuth_pl

Guru

Guru

  • Posts: 848
  • Joined: 10 Aug 2004, 16:28
  • Location: Poland, Warsaw

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post07 Jul 2010, 23:08

thanks for the input.
I'm not inclined that much to gutting Pulsars to read values for all components as I have to many other watch technologies to explore. I'm still a LED-nut but only half as much as before.
my guru friend was curious so I gave him a couple of modules for testing and that was his analysis. he had plenty of time and also created an inhouse frequency generator that works the same as the oscillator chip but he added a trimmer for perfect specs ;))
I'll have to talk to the man in more detail to understand his reasoning.

as for values... I was not referring to a change of 100 :)
"The first and still only LED watch maniac in the East Block" - www.crazywatches.pl
Offline
User avatar

bruce wegmann

Pulsar Moderator

Pulsar Moderator

  • Posts: 1287
  • Joined: 02 Aug 2004, 02:13
  • Location: San Diego, CA

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post08 Jul 2010, 03:19

While these oscillators do work, it's worth noting the frequency tolerance specified on the data sheet. It's 50ppm [parts per million], which sounds pretty good until you consider that there is something over 31 million seconds in a year. 50ppm X 31 million = 1550 seconds a year, or about 4.25 seconds a day. Most decent mechanical watches will do that good. Original Time Computer guarantee was 60 seconds a year, or about 2 (yes, 2!) ppm. I'm amazed modern manufacturing techniques can't routinely produce something at least as good as that. Also, there's no simple way to make these work with the original trimmers, so, basically, you get what you get, as far as overall accuracy is concerned. Still, it's better than a dead watch, and further down the road, the great majority of modules are going to need a repair of this sort...
Offline
User avatar

Old Tom

Wizard

Wizard

  • Posts: 274
  • Joined: 15 Dec 2006, 10:48
  • Location: London, England

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post08 Jul 2010, 10:03

Time to lay the urban myth of the thick film resistors on the P2/P3 boards being "thermistors" to adjust the frequency to rest.They are not that, what they are are twofold- the ones just below the LED drive transistors are current limiters (no big mystery there), the big labyrinthine ones in the oscillator circuit (under and beside the main chip) are parasitic oscillation stoppers, RCA recommended adding these when (misusing ) digital inverter stages as oscillator stages in early A series CMOS. Think for a moment, if you could adjust the frequency of a crystal oscillator with something as simple as a resistor why would you fit two variable capacitors to do this to your circuit board (far less reliable). However the real reason for printing these resistors on the board with the shape they have has not a lot to do with the resistance value- they are there to add a few nano-Henries of inductance to the circuit to make sure the oscillator runs at the fundamental frequency and not an overtone (CMOS circuitry is voltage switched and a few hundred thousand ohms or less of resistance makes b*gger all difference to the actual operation, notice how the main bias resistors in digital watches are always in the megohm and higher region). And why print the labyrinthe resistors on the board rather than just laying down a trace in the same shape, well, you do need some resistance to limit the back voltages induced by the self-inductance frying the very delicate input gate in the oscillator, no resistance and you might get enough volts to puncture the junction.

So, comments please :-D
Offline
User avatar

rewolf

Guru

Guru

  • Posts: 1863
  • Joined: 11 Jul 2004, 15:32
  • Location: Ravensburg, Southern Germany

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post08 Jul 2010, 14:07

I'd like to see a diagram of the oscillator circuit including the 2 resistors in question. That would make it easier to judge whether the resistors (if they are resistors at all) could be used for temperature compensation. I have never seen any crystal oscillator that is temp-compensated simply by thermistors - but that doesn't mean it can't work, it could be an unconventional solution that did its job. See http://iopscience.iop.org/0022-3735/15/7/008.
Here is a thermistor compensation with a varactor diode acting as temperature-dependent "trimmer capacitor": http://www.4timing.com/techoscillator.htm.
IMHO it could be true that the "meanders" are indeed inductors - but to have any effect the'd have to be rather in the µH than the nH range.

For easier reference:
a) the original thread where Phil Bowler (leddwatch) 'disclosed the secret': click here.
b) thread with Ed's findings: click here


@bruce: the point is adjustment (trimming).
You can get 2ppm with modern crystals as well if you adjust them with a trimmer.
But all those oscillator chips are unadjusted and thus depend on initial production tolerance of the crystal they contain - usually 20ppm (10ppm for good ones).
I doubt that in 1973 any 32kHz crystal had less than 20ppm initial deviation.
But on top of the possible 2ppm is the temperature dependent resonant frequency of the crystal.
Quartz watches are adjusted for "wrist temperature".
Quartz crystals are specified for 25°C. The 32kHz tuning fork crystals have a negative tempco in both directions, i.e. they run slower at temps above or below 25°C.
Last edited by rewolf on 08 Jul 2010, 17:27, edited 1 time in total.
Offline
User avatar

retroleds

Guru

Guru

  • Posts: 3636
  • Joined: 04 Feb 2006, 10:34
  • Location: Surrounded by hicks and sticks (farms and woods) - Michigan,USA

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post08 Jul 2010, 17:18

I have at no time said those resistors are adjusting the frequency(although, isn't stabilizing a frequency an adjustment?). I merely have commented on their wide change in resistance when subjected to a fairly small change in temperature(difference between room temp and me blowing on them via a drinking straw). I've seen enough modules with a small can type resistor tapped across the course trimmer, which looked totally factory, to believe the factory intended for those to to contribute to the accuracy of the movement. A temperature compensating unit would need to be big enough to not be effected by some small spot heat - their size and configuration below the IC, which does generate some heat, seems well planned. To me.

Of course, if you start with a module where that section is damaged already, you are not going to get much of a change in readings.

Normal wrist temperature of a wrist watch at "room temp" is 86 f.(30c), so it seems their function was more geared toward compensating for cold temps, rather than warmer.
http://www.retroleds.com - Sales of vintage LED, LCD, analog watches, parts and gadgets - repair tutorials & tips
Nov. 2022 - back in business!! BItter divorce is in home stretch, come grabs some great deals, I had to open the safe . . . damn attorneys. piss.
Offline
User avatar

rewolf

Guru

Guru

  • Posts: 1863
  • Joined: 11 Jul 2004, 15:32
  • Location: Ravensburg, Southern Germany

Re: : Suitable 32.768KHz oscillator for Pulsar Modules...rea

Post08 Jul 2010, 19:21

retroleds wrote:....A temperature compensating unit would need to be big enough to not be effected by some small spot heat - their size and configuration below the IC, which does generate some heat, seems well planned. To me.
The correct place for a temperature sensor would be as close to the quartz crystal as possible, NOT influenced by heat from the chip (which does BTW only warm up VERY little, if at all, when the LEDs are on).

retroleds wrote:Normal wrist temperature of a wrist watch at "room temp" is 86 f.(30c), so it seems their function was more geared toward compensating for cold temps, rather than warmer.
The usual frequency deviation for 32kHz tunig fork crystals (modern type) is -0.04ppm * (25°C - T)², that is, the frequency deviation is always negative with respect to 25°C.
Iif you adjust the oscillator for 0ppm @ 25°C, then the deviation is -1ppm (0.1s/day slow, 2.6s/month) @ 30°C resp. 20°C, and -4ppm @ 35°C resp. 15°C, and -9ppm (0.8s/day) @ 40°C resp. 10°C.
If it's adjusted for 30°C (wrist temperature), there is NO deviation at 20°C, but +1ppm (2.6s/month fast) @ 25°C, -3ppm @ 15°C, -8ppm @10°C. (I think a diagram would make this clearer...)
Conclusion: it seems wiser to adjust the oscillator for 30°C because this leads to less overall deviation over the interesting temperature range (15°C ~ 35°C).

Edit: here's the diagram:
Image
Offline
User avatar

retroleds

Guru

Guru

  • Posts: 3636
  • Joined: 04 Feb 2006, 10:34
  • Location: Surrounded by hicks and sticks (farms and woods) - Michigan,USA

Re: : Suitable 32.768KHz oscillator for Pulsar Modules...rea

Post09 Jul 2010, 00:30

rewolf wrote:.The correct place for a temperature sensor would be as close to the quartz crystal as possible, NOT influenced by heat from the chip (which does BTW only warm up VERY little, if at all, when the LEDs are on).
Thought this too, ; however, a temperature sensor embedded on the realitvely heavy ceramic circuit board would be more moderated in it's adjustments(possibly) since that would not react as quickly to every small temperature change. Find us a Pulsar engineer so we can ask what they were thinking. :lol:
Last edited by retroleds on 09 Jul 2010, 16:25, edited 1 time in total.
http://www.retroleds.com - Sales of vintage LED, LCD, analog watches, parts and gadgets - repair tutorials & tips
Nov. 2022 - back in business!! BItter divorce is in home stretch, come grabs some great deals, I had to open the safe . . . damn attorneys. piss.
Offline
User avatar

rewolf

Guru

Guru

  • Posts: 1863
  • Joined: 11 Jul 2004, 15:32
  • Location: Ravensburg, Southern Germany

Re: : Suitable 32.768KHz oscillator for Pulsar Modules...rea

Post09 Jul 2010, 13:58

retroleds wrote:... Find us a Pulsar engineer so we can ask what they were thinking. :lol:
I'd first ask IF it's a "sensor" (of whatever kind) at all...
Offline
User avatar

retroleds

Guru

Guru

  • Posts: 3636
  • Joined: 04 Feb 2006, 10:34
  • Location: Surrounded by hicks and sticks (farms and woods) - Michigan,USA

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post09 Jul 2010, 17:55

Well, I've contributed what I can to this...those who are inclined should be able to figure out the rest of this. A module that reacts well to the " heat and run/cool and quit " would be a good candidate...buy a few resistors and give them a try(I've already thrown out some good ranges to start at, follow the threads above). I just looked at a few this morning again...some of those resistors are flaked up and have very high resistance or completely open circuit, others are nice and smooth and show normal ranges. Looks like the aluminum traces failed badly...we all know how easily that oxidizes and goes to hell. You don't need to access the pins of the IC - the trimmer connections are very accessible. I'd take some nice pictures of the whole deal stripped down, but I need a new bulb for the microscope light.
http://www.retroleds.com - Sales of vintage LED, LCD, analog watches, parts and gadgets - repair tutorials & tips
Nov. 2022 - back in business!! BItter divorce is in home stretch, come grabs some great deals, I had to open the safe . . . damn attorneys. piss.
Offline
User avatar

azimuth_pl

Guru

Guru

  • Posts: 848
  • Joined: 10 Aug 2004, 16:28
  • Location: Poland, Warsaw

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post10 Jul 2010, 12:19

here is a detailed analysis of my expert friend performed on a handful of dead and working modules.
tested at 30ºC and 80ºC giving a variance of 1% which is in spec for a precision resistor = not a thermistor.
fell free to discuss and correct my technical translation if required.

symbol / type / value at 30ºC / value at 80ºC / resistor designation
R1 / SMD / 22,1MΩ / 22,2MΩ / parallel for oscillator
R2 / sputtered / 470Ω / 473Ω / serial for oscillator
R3 / sputtered / 439Ω / 441Ω / serial for emitter
R4 / SMD / 1,0MΩ / 1,07MΩ / parallel for lightsensor
Rb / sputtered / 2,22KΩ / 2,23KΩ / serial for base
Rc / sputtered / 75Ω / 75Ω / serial for collector
Rx / sputtered / 88,4KΩ / 89,1KΩ / pull-up -Uz
Ry / sputtered / 86,9KΩ / 87,4KΩ / pull-up +Uz
C1 / SMD / 27pF / not tested / fixed for oscillator
C2 / SMD / 2,7nF / not tested / lightsensor circuit

Image
"The first and still only LED watch maniac in the East Block" - www.crazywatches.pl
Offline
User avatar

rewolf

Guru

Guru

  • Posts: 1863
  • Joined: 11 Jul 2004, 15:32
  • Location: Ravensburg, Southern Germany

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post10 Jul 2010, 14:09

Great - now we're getting somwhere :-D
Did your friend happen to draw a circuit diagram of the oscillator part? I didn't succeed in following the traces on the photo, too many points where it's unclear whether there's a connection or not - and as a Pulsar ignorant I don't even know where the crystal is soldered :roll: .
Offline
User avatar

azimuth_pl

Guru

Guru

  • Posts: 848
  • Joined: 10 Aug 2004, 16:28
  • Location: Poland, Warsaw

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post10 Jul 2010, 15:50

no diagram I'm afraid.
hope this helps:
Image
"The first and still only LED watch maniac in the East Block" - www.crazywatches.pl
Offline

Handy

Techie

Techie

  • Posts: 124
  • Joined: 14 Jan 2005, 14:17
  • Location: Stoke-on-Trent,England.

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post10 Jul 2010, 20:09

Are these tantalum nitride resistors?
Offline
User avatar

azimuth_pl

Guru

Guru

  • Posts: 848
  • Joined: 10 Aug 2004, 16:28
  • Location: Poland, Warsaw

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post10 Jul 2010, 20:12

I'm not sure if such a detailed analysis is possible. they don't taste like tantalum ;)
"The first and still only LED watch maniac in the East Block" - www.crazywatches.pl
Offline

Handy

Techie

Techie

  • Posts: 124
  • Joined: 14 Jan 2005, 14:17
  • Location: Stoke-on-Trent,England.

Re: : Suitable 32.768KHz oscillator for Pulsar Modules...rea

Post10 Jul 2010, 20:18

azimuth_pl wrote:here is a detailed analysis of my expert friend performed on a handful of dead and working modules.
tested at 30ºC and 80ºC giving a variance of 1% which is in spec for a precision resistor = not a thermistor.
fell free to discuss and correct my technical translation if required.

symbol / type / value at 30ºC / value at 80ºC / resistor designation
R1 / SMD / 22,1MΩ / 22,2MΩ / parallel for oscillator
R2 / sputtered / 470Ω / 473Ω / serial for oscillator
R3 / sputtered / 439Ω / 441Ω / serial for emitter
R4 / SMD / 1,0MΩ / 1,07MΩ / parallel for lightsensor
Rb / sputtered / 2,22KΩ / 2,23KΩ / serial for base
Rc / sputtered / 75Ω / 75Ω / serial for collector
Rx / sputtered / 88,4KΩ / 89,1KΩ / pull-up -Uz
Ry / sputtered / 86,9KΩ / 87,4KΩ / pull-up +Uz
C1 / SMD / 27pF / not tested / fixed for oscillator
C2 / SMD / 2,7nF / not tested / lightsensor circuit

Image


A very interesting and informative post. Thank you.
Offline

Handy

Techie

Techie

  • Posts: 124
  • Joined: 14 Jan 2005, 14:17
  • Location: Stoke-on-Trent,England.

Re: : Suitable 32.768KHz oscillator for Pulsar Modules...rea

Post10 Jul 2010, 20:21

azimuth_pl wrote:I'm not sure if such a detailed analysis is possible. they don't taste like tantalum ;)

Tantalyzing! :-D
Offline
User avatar

rewolf

Guru

Guru

  • Posts: 1863
  • Joined: 11 Jul 2004, 15:32
  • Location: Ravensburg, Southern Germany

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post10 Jul 2010, 21:22

The video linked in Handy's post says at 2:57:
... Laser beams are cutting apart the resistance networks that will be used to regulate each crystal - each one a network of fifty (of fifteen?) tantalum nitride resistors, the most stable resistance material known. Each network evaporated onto a ceramic substrate ... and connected with gold conductors...

There are 15 printed resistors in azimutp_pl's picture, plus 7 under the display (eays to spot in the video), makes 22 - neither 15 nor 50.

Resistance networks to regulate (adjust?) the crystal? Or does "regulate" here rather mean "operate"?. In the next scene they describe the trimmer condensers.

Image
Offline
User avatar

LEDluvr

Guru

Guru

  • Posts: 1001
  • Joined: 06 Jun 2007, 22:49
  • Location: Los Angeles

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post10 Jul 2010, 21:26

Fascinating!

Image

Uploaded with ImageShack.us
Offline
User avatar

azimuth_pl

Guru

Guru

  • Posts: 848
  • Joined: 10 Aug 2004, 16:28
  • Location: Poland, Warsaw

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post10 Jul 2010, 22:22

now I recall :)
that print screen is from the Pulsar promo video I digitilized years ago on disc.
so it has to be tantallum nitrade resistors evaporated onto the ceramic substrate.
I used the word "sputtered" in my description as it is quite equal in sense and found in electronics terminology.

there are exactly 15 if you only count the spiral ones, consider Ry as two separate and include R3.
those 15 are clearly marked with a darker color on the diagram in the video.
that means that the rest can also be sputtered but not tantallum.

I'd say that "regulate" means "operate".
the whole concept is based on elementary knowledge:
http://en.wikipedia.org/wiki/Pierce_oscillator

this is a more detailed pic of the ones under the display:
Image
Last edited by azimuth_pl on 11 Jul 2010, 14:17, edited 1 time in total.
"The first and still only LED watch maniac in the East Block" - www.crazywatches.pl
Offline
User avatar

retroleds

Guru

Guru

  • Posts: 3636
  • Joined: 04 Feb 2006, 10:34
  • Location: Surrounded by hicks and sticks (farms and woods) - Michigan,USA

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post11 Jul 2010, 01:47

Does this vindicate my assertion those are temperature sensors or not? Great pictures and information you guys found - damned impressive!! 8-)

On the testing, it seemed to me the greatest change was between room temperature(22c/72f.) and body temperature(36c/98F). I can't imagine many watches being exposed to 80c. Unless left on the dashboard of a car in summer, which would probably lead the batteries to leak. Most electronics kind of crap out at that temperature range, not sure the Pulsar engineers tried to cover that range.

Tantalum, sounds like some cool Russian vodka. Or maybe a death metal band. NO - it is this band of hot chicks from Australia!!
http://www.myspace.com/tantalum

RWAR - talking about wanting to visit down under! :O`~
http://www.retroleds.com - Sales of vintage LED, LCD, analog watches, parts and gadgets - repair tutorials & tips
Nov. 2022 - back in business!! BItter divorce is in home stretch, come grabs some great deals, I had to open the safe . . . damn attorneys. piss.
Offline
User avatar

azimuth_pl

Guru

Guru

  • Posts: 848
  • Joined: 10 Aug 2004, 16:28
  • Location: Poland, Warsaw

: Suitable 32.768KHz oscillator for Pulsar Modules...read on

Post11 Jul 2010, 01:56

not sensitive to temperature in technical terms thus not a thermistor.
usually extreme variables are taken into consideration for testing purposes but as shown the high temperature did not impress the patient :)

btw. the largest quantities of this rare metal is in.... Oz!
"The first and still only LED watch maniac in the East Block" - www.crazywatches.pl
PreviousNext

Return to Electronical

Who is online

Users browsing this forum: No registered users and 17 guests