Car clock

[Scott]

How does my car clock work? It cannot be quartz or radio controlled as it runs slightly fast. It cannot be synchronous because the supply is DC. Is there any form of adjustment to get it to run to time?

[charles]

quartz can run fast - or slow,

[Scott]

In what circumstances? I thought it was supposed to be accurate to one second in so many thousand years. Can the crystal be replaced?

[Chris Hogg]

I have always thought that some quartz crystal oscillator circuits can be fine-tuned a few cps either side of the principal frequency, by way of a simple variable capacitor, and observing the result using an accurate oscilloscope. I may be wrong, and even if I'm not, I wouldn't recommend trying it unless you're going to scrap the oscillator anyway and making a complete balls-up won't matter.

[Tim+]

When cheap crystals and other components are used. Crystals vibrate at a regular frequency, not an absolute frequency. Which frequency and how narrow the frequency spread is is what manufacturers pay for.

Nothing like that good but some meet the makers specs better than others.

I dare say in theory. How easy that will be to do in any particular car and whether it?s worth the trouble is harder to say.

Tim

[Ian Jackson]

In message [e-mail], Scott [e-mail]> writes

I think you're thinking of rubidium - as in rubidium atomic clock.

I'm sure it can - but replacing it with a rubidium frequency standard might prove quite a challenge!

[charles]

That depends on the quality of the crystal and the control of its operating temperature. Really high accuracy crystals probably cost as mucha a car.

[Ian Jackson]

In message [e-mail], Chris Hogg [e-mail] writes

I doubt if many present-day watches actually have a trimmer capacitor, However, the first one I had (in the 1970s, and bought from a guy at work who seemed to be able to get his hands on a supply) DID have a trimmer. With a bit of trial and error, I got the time accurate to a few seconds a month.

[invalid unparseable]

All of them.

Not with commercial grade clocks.

Yes, but that wont fix the problem.

[Harry Bloomfield, Esq.]

charles brought next idea :

Or use a normal crystal, but in a temperature controlled oven.

The usual way, is have the clock/watch correct itself once per day, by receiving MSF.

[invalid unparseable]

Yes. But they still vary with temperature.

Nope, not even a frequency counter because they arent that accurate. You adjust it by seeing if the clock gains or loses.

It wont work anyway, it will still be temperature sensitive with commercial grade clocks. You need a temperature controlled micro oven for the xtal and that's not economic.

You can get clocks that use the gps system to keep accurate time now and ones that use the mobile phone system.

[Harry Bloomfield, Esq.]

Actually they do calibrate them, against a very accurate and stable crystal timebase in watch repairers.

[Theo]

The first 32.768kHz watch crystal that shows up on Farnell has a tolerance of +/-20ppm:

There are 86400 seconds in a day, therefore:

86400*(20/1000000) = 1.728 seconds/day it could be either fast or slow by that amount. That's 10.5 hours/year.

The frequency stabilty is also affected by temperature - for this one that would be another -20ppm across the kinds of temperature ranges a car might be exposed to, so the same time skew again if the tolerance happened to start at -20ppm.

I today dug out an old digital camera that I haven't used for ~5 years. Its clock was half an hour fast. That's not bad going, all things considered.

Theo

[alan_m]

What car - how old?

[Martin Brown]

It is almost certainly quartz crystal and not quite trimmed right. Such mistakes were very common. One telescope maker fixed it by adding GPS functionality to reset their RTC to local time. Mine as built from a previous generation lost 15s/month reliably. They didn't install the loading capacitors around the crystal to save money ( about 10c ).

It isn't hard to get a quartz clock good to 1ppm or about 30s/year but only the better manufacturers actually bother to calibrate them.

It is slightly harder for a clock in a car than a watch on someone's wrist which tends to be temperature compensated by the wearer.

[alan_m]

They drift by x parts per million per degree celcius hence why with a crystal if you wanted repeatable accuracy you put it in temperature controlled environment.

[Vir Campestris]

Maybe it can. But they trim it until it's "close enough". That's all.

(My watch runs fast)

Andy

[Clive Arthur]

Yebbut a watch runs at a comparatively constant temperature.

[Fredxx]

Being a car clock it will be processor based. Some are smart enough to add or take clock cycles to calibrate the crystal in an automated fashion.

How old is the car? Crystals do age to a certain extent. Cheap ones are

100ppm, ~10s per day. How much is the drift?

For the past decade, most recent radios rely on broadcast stations encoding the time.

[polygonum_on_google]

My car seems to use GPS is some form - which is both good and frustrating as it doesn't have built-in satnav.

However, it does not switch to/from summer time. I have switch that on or off manually.