> Date:    Wed, 31 May 2000 08:23:27 -0500
>
> i have the same problem
>
> please reply to me if you find an answer
>
> been told by the manufacturer of our pcs
> that it is not possible

Is it the problem or the solution that "is not possible"?  ;-)

> >>> "Anthony Spee (EPA)" <[log in to unmask]> 05/30/00 08:52PM =
> >>>
> I have a HP Vectra VE (Celeron 366 Mhz with 32 MB of RAM) running =
> Windows95
> loosing about
> 2 minutes daily on its PC clock.

I see two possible causes:

1.  There is some software that interrupts repeatedly at regular intervals,
    and miscalculates, or somehow "steals" a tiny time slice every time it
    interrupts.  Try running the system for a day, after booting from a
    simple DOS boot floppy - no TSRs or applications.  I suspect that it
    is a software problem, and the bare system will keep much better time.
    Try adding one application at a time to locate the problem.  If you
    are a hot programmer, you may be able to locate the faulty interrupt
    loop in a similar fashion, and adjust it accordingly.

2.  There could be a problem with the real-time clock crystal circuit.
    If you have access to an oscilloscope, or digital frequency meter,
    you will probably find that the frequency is 32768 Hz.  It is possible
    to adjust the frequency of oscillation (even the capacitance of the
    oscilloscope or meter probes can have an effect) by varying the
    capacitances in the circuit.  The problem is that you are probably
    stuck with a range of roughly 5 to 25 pF, or the oscillation stops,
    and this might be good for +/- 10 seconds per day at best.  Odds
    are, the crystal is running within 50 ppm of real time.


> Is there a way I can calibrate the PC's clock to run faster or slower?

Yes, but probably not enough.  Some early PCs actually had variable
capacitors in the circuit.  They were more trouble than they were worth,
so now everybody uses fixed capacitors, and most systems should be good
to within a minute per month, until the battery dies.

> Or
> Is there a way the PC's time can daily reference to the BIOS time and be
> adjusted accordingly?

Find the faulty interrupt loop.  My guess is that the program was
developed for a faster processor than yours.  To exaggerate, suppose the
loop took 10 seconds to execute on that machine.  The last thing it does
is update the clock by 10 seconds - although it takes 12 seconds
on your computer.  You lose 2 seconds every loop.

> Or
> Can I set something up to automatically adjust the time by X seconds =
> daily?

Best solution is to fix the faulty loop.  Failing that, you could have
your programmer set up a TSR that adjusts the time by, say, +15 seconds
every three hours.  Just make sure that you can change this to suit
changing conditions by yourself.  Odds are, someone on this list knows of
such a program that is readily available.

> Please note: the PC is not networked or connected to the internet.
>
> Whats the PC used for you ask?
> The PC is running a program based on Foxpro which collects real time data
> from a call centre
> switch and displays it on overhead message boards for call centre staff to
> view.
> These overhead displays also show the current time of day (derived from =
> the
> PC's time).
> The PC time is loosing about 2 minutes every day - very annoying since =
> this
> incorrect time
> is displayed throughout the call centre.
> I replaced the PC thinking the clocking on the mother board was faulty but
> the new PC also
> looses time. And guess what? neither PC looses time when running back at =
> my
> office.
> I can't run the Foxpro program properly back at he office since I dont =
> have
> a call centre
> there so I ran a video (no not one of those videos) overnight to keep the =
> PC
> busy and it only
> lost 2 seconds.

That sounds like the culprit.  Perhaps it's the time it takes to answer a
call that is the problem.  Look for something that is interrupt-driven.
It could be a result of something that takes place outside your computer
- perhaps in the telephone exchange.  It would take some sophisticated
time comparisons to see if the time loss is the same for each interrupt,
or if it varies depending upon circumstances.

>
> Regards,
>
> Tony Spee
> Email:  [log in to unmask]

Boyd Ramsay

[log in to unmask]

            Do you want to signoff PCBUILD or just change to
                    Digest mode - visit our web site:
                    http://nospin.com/pc/pcbuild.html