Re: Napco clock drift
- From: mpm <mpmillard@xxxxxxx>
- Date: Tue, 18 Nov 2008 21:27:33 -0800 (PST)
On Nov 10, 7:56�pm, "Eric B." <ebe...@xxxxxxxxxxx> wrote:
Hi,
I've had a Napco Magnum 1000 for about 16yrs already. �Overall, the system
works perfectly well - everything it was setup/designed to do is precisely
the way it works. Except for one thing.
The daily phone verification has a problem and has always had this problem
since day one. �Ever since the system has been installed, the clock seems to
drift a second or so every day. �If the system is programmed to dial out at
4am, every few months (never really checked it carefully, but might be a
question of 6-8 months or so), it drifts back to 9pm or 10pm where it
becomes a nuissance to the people living in the house. �At which point, the
dial out timer has to be reset back to 4am, and that works for another
series of months before the same issue repeats itself.
I was wondering if anyone else has ever encountered that kind of issue in
the past, and if so, what can be done to rectifiy it, instead of having to
manually reset it every time it drifts too much.
Can a faulty/flaky transformer be responsible for something like this? �I
know in theory it shouldn't, if the crystal on the board responsible for the
digital clock is accurate, but I'm out of ideas. �The whole alarm system is
on a UPS, so I am certain that the input current is glitch free.
Any thoughts?
Thanks!
Eric
I'm not familiar with this panel, but I'm well-versed on electronics.
A couple things could be going on here.
First, there is the distinct possibility that the panel cannot do two
things at once (in software).
It may be designed to NOT update the realtime clock because the
resources that would usually keep that clock accurate (i.e., the
microprocessor) are busy doing other things (like communicating with
central station). The end result is the realtime clock looses a few
seconds here and there, and eventually they all add up.
If so, you're probably out of luck.
If the panel uses dedicated clock hardware (like a Dallas/Maxim
DS1305, 1307 chip etc..) it could be the time reference crystal is
wearing. Crystals do drift with time - it is a well established
phenomenon. But you say it's been doing since day-1, which (unless
it's getting worse with passing years) may not be the problem.
The only other possibility is the timebase reference itself is not
very accurate. Typically, a 32kHz crystal is used, and this clock
gets divided down to provide minutes and seconds. This crystal would
be MUCH slower than the main processor clock (which is another crystal
- probably in the 2MHz, 4MHz, or even higher speed clock range).
Technical note: Usually crystals drift most during their first year
of operation. They continue to drift thereafter, but generally at a
much slower rate. Crystals generally don't drift "that far" from
whatever's stamped on them... but like everything else.. timing errors
accumulate. This is why you even have to reset quartz digital
wristwatches sometimes...
Regardless, (either) crystal can be warped (corrected) by adding, or
removing, the appropriate amount of capacitance. This is an advanced
topic, and requires an oscilliscope and soldering skills. If the
realtime clock has a separate timebase crystal, correcting that clock
would solve the problem. If however, the RTC is derived off the
processor clock, you have to be careful how far you correct it, as all
other panel functions that depend on a time base (i.e., touchtone
dialing, central station reporting, etc..) would be affected as well.
I somewhat doubt the two clocks are shared, but like I said, I'm not
familiar with your panel.
Finally, all crystals drift with temperture. Most real-time clock
crystals are characterized for "best" accuracy at or around room
temperature. The further you get from this value, the worse the error
frequency becomes. (A panel in the un-air-conditioned laundry room
will stuggle more with keeping time than one inside the house, for
example.)
For completeness, I will also say that a TCXO or OCXO device can be
used in place of a 32kHz clock crystal. TCXO = Temperature Controlled
Crystal Oscillator, or OCXO = Oven Controlled Crystal Oscillator.
(This is not common practice in alarm panels.)
As either name implies, the alternate devices attempt to flatten out
the temperature response curves of the crystal. The timing errors
associated with TCXO / OCXO are MUCH better than a plain old crystal,
but a crystal (in this application) is usually sufficient. As you
would expect, these devices are usually more expensive, so they don't
often find their way into consumer electronics.
Finally, the timebase(s) could be interferred with by external
electrical noise.
Again, not very likely, but it does happen. It almost always
manifests as a fast clock (extra clock ticks electrically), so again,
not very likely given your description of the trouble.
Now that we've covered clocks, the sad conclusion is (as others have
posted) that you're probably just going to have to live with it.... :-
(
-mpm
.
- References:
- Napco clock drift
- From: Eric B.
- Napco clock drift
- Prev by Date: Re: Napco clock drift
- Next by Date: Re: Lets try something .... again.
- Previous by thread: Re: Napco clock drift
- Next by thread: Re: Installer Manual needed for Honeywell 1500e
- Index(es):
Relevant Pages
|
