Causes of clock drift? / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Causes of clock drift?

1 2 next

Move Thread

LAN_403

aonomus

Tue Feb 24 2009, 12:15PM

Registered Member #1497 Joined: Thu May 22 2008, 05:24AM
Location: Toronto, Ontario, Canada
Posts: 801

So the article on wikipedia doesn't seem to help much, but I've been having this odd issue, my 10-15 year old alarm clock, fine until this point has been continuously moving faster and faster through time. Every day I need to re-sync it to the actual time of day, and in the morning its 10-20 minutes ahead.... (A real pain cause you end up waking up early...).

This only happened since I recently moved into this townhouse, and the only things I suspect are the power frequency (steady at 59.9 Hz, same as the old apartment), or the temperature in my room. This apartment has the tendency to get cold overnight, could this be the cause? A temperature based change would be suspected, but would a lower temperature not cause the crystal to oscillate slower, causing time to pass more slowly for the alarm clock, than the rest of the world.

Proud Mary

Tue Feb 24 2009, 01:04PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

10 - 20 min/day is way outside the deviation you could expect from any kind of domestic temperature change. so far as crystal oscillator drift is concerned.

Clock oscillators will typically stay on frequency within 10 - 20 ppm over 0 - 50 deg. C.

Crystals age over time, perhaps 5 ppm per year, but this is an error that creeps up slowly, and not something you'd notice!

The crystal might have cracked, or come undone, I suppose, but I've no idea how to predict what effect that this might have on the operation of clock.

Supply voltage variation, or a variation in the oscillator load, might just be sufficient to cause your large error, but if the slowness be variable from day to day, I would suspect some kind of mechanical failure - a worn bearing or slipping sprocket in the stepper motor drive for example - but there is no way to tell by guess work.

GeordieBoy

Tue Feb 24 2009, 01:13PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

Do you have access to any of the following:

1. A variac
2. A variable frequency supply
3. An environmental chamber

You could test susceptibility to line voltage and frequency variations with the first two, and temperature changes with the last one. You could always try putting the clock in the fridge (not freezer!) and see if that has a more dramatic impact on the timing error.

It does sound excessive for a crystal oscillator so maybe there is a cheap ceramic resonator in there.

Or maybe it gets its timing information from the mains supply frequency but harmonics or noise are causing multiple clocking of the internal counters.

-Richie,

Bjørn

Tue Feb 24 2009, 02:26PM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

but would a lower temperature not cause the crystal to oscillate slower, causing time to pass more slowly for the alarm clock, than the rest of the world.

The crystal would shrink and oscillate faster, just like shrinking a pedulum makes it faster. Tim Hunkin demonstrated it in The Secret Life of Machines by baking a watch in an oven to make it slow down.

I had an alarmclock that would switch to a RC oscillator if power was lost, maybe yours have done the same becuase of some fault?

Dr. Dark Current

Tue Feb 24 2009, 02:48PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

I have here a very old LED alarm clock, and when I took it with me on holidays, it did show anything from 0 to 30 minutes more every day.
The cause was noise in the line, as this clock gets its timing signal from the line frequency.

aonomus

Tue Feb 24 2009, 03:21PM

Registered Member #1497 Joined: Thu May 22 2008, 05:24AM
Location: Toronto, Ontario, Canada
Posts: 801

I have a small fridge I use for storing samples and other things and a variac, however the line frequency has never deviated from 59.9Hz in the last week whatsoever (its not like we switch to 400Hz for a few hours per day :))

I'll give the fridge experiment a shot after I try cycling power to it, we have had a brownout recently (line voltage sag).

Dr. Dark Current

Tue Feb 24 2009, 03:30PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

aonomus wrote ...

however the line frequency has never deviated from 59.9Hz in the last week whatsoever (its not like we switch to 400Hz for a few hours per day :))

But I guess it still can have noise? Maybe the neighbor decided to control his washing machine with a dimmer...?

aonomus

Tue Feb 24 2009, 03:40PM

Registered Member #1497 Joined: Thu May 22 2008, 05:24AM
Location: Toronto, Ontario, Canada
Posts: 801

Dr. Kilovolt wrote ...

aonomus wrote ...

however the line frequency has never deviated from 59.9Hz in the last week whatsoever (its not like we switch to 400Hz for a few hours per day :))

But I guess it still can have noise? Maybe the neighbor decided to control his washing machine with a dimmer...?

I would be very *very* amused if he tried controlling a 240VAC washing machine with a 120VAC dimmer and had it explode.... I suppose line noise is a possibility but its way outside the range that line noise would make sense... I used to live in a 13 floor apartment complex which had constant power issues (main feed into building kept overheating and blowing out switchgear).

Bored Chemist

Tue Feb 24 2009, 06:02PM

Registered Member #193 Joined: Fri Feb 17 2006, 07:04AM
Location: sheffield
Posts: 1022

Just a quick note for our freinds across the pond; you can get 240V dimmers.
Anyway, the noise is the simplest explanation. A cheap filter might save you valuable sleeping time.

Does the clock have a battery?

Proud Mary

Tue Feb 24 2009, 09:03PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Perhaps the clock is just worn out!

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob