[time-nuts] Discussion of precise time and frequency measurement

Will Matney xformer at citynet.net
Thu Jun 30 00:45:19 UTC 2011


Yea, it's similar to the way you track a voltage reference or standard over
time between calibrations, but here, it would be frequency instead. I don't
know if it would really need to track it all the time, maybe check it so
many hours a day, or so many times during peak, and night periods to see
how it may shift. I would say you could track the phase the same way, just
use a different circuit to pick up any shift. I don't know if there might
be some software already written by one of the calibration companies for
something like this.

I imagine if the app could write to a file, each reading, and then use that
info on the graph. If one did monitor it the whole time, they could make
the software log any changes from 60 Hz the same way, while time stamping
them, and adding them to the file. It wouldn't have to write to it unless
it was off by so may specified parts per million, or say microsends, etc.
Once done, and at the periods end, a quick look at the graph would show the
plot points where it deviated from 60 Hz over a month, etc.

One would also want the file in case the computer would crash, or need a
reboot, etc, as you sure wouldn't want to lose all the recordings. I've
never programed anything using graphs, so I'm not sure how hard it would



*********** REPLY SEPARATOR  ***********

On 6/29/2011 at 5:15 PM Chris Albertson wrote:

>OK, now I understand why you wanted a sound card with a wel regulated
>clock.   Yes they make those.  The common cheap cards use a square can
>oscillator good to about 100PPM.      Better cards can use a better
>and way more expensive clock source.  I think knowing the sample rate
>is accurate is good enough.   Then you simly feed the 60Hz signal to
>the sound card and record it.  Latter plot a spectra or whatever.   I
>thought of that too but then rememebered that the audio card's best
>sample rate is only about 96K samples per second so the time
>resolution is only on order it 1/48,000 second.
>Then I though I could square the 60Hz wave and feed it to pin one on a
>serial port the same way you conect the PPS signal from a GPS.  I
>could build a device to sense the 60Hz power by aiming a photo
>transistor at the florescent lights on the ceiling.   I'm only half
>joking.  I think a neon pilot lamp and a photo cell both inside a
>light proof metal box might work.
>The device driver that reads the PPS from a GPS will not care if the
>signal is comming from the power line it will be hapy to time tage
>every cycle with a nano second counter value.
>I think your plot is simply the delta between cycles.  Or the period
>of the 60Hz input.  If you plot the period it might go up or down over
>On Wed, Jun 29, 2011 at 4:30 PM, Will Matney <xformer at citynet.net> wrote:
>> Chirs,
>> What I would think that would be needed, software wise, is an app that's
>> similar to the long term plots used in calibration. I would say that you
>> know the ones I'm speaking of, as it is a graph that stretches over a
>> (or longer), and lets say the horizontal center line would represent 60
>> and above or below it would be how much it's off in which ever
>> you want, either in frequency, or in PPM. Long term time is on the
>> horizontal axis, and frequency+/- is the vertical. That was my thinking
>> other day, so it would log any shifts over a long timespan, really
>> to the old paper plots off a frequency standard. Now whether you could
>> this with a sound card, I don't know, as one would need a frequency
>> standard, along with the line frequency, as inputs so they could be
>> compared.
>> Best,
>> Will
>> *********** REPLY SEPARATOR  ***********
>> On 6/29/2011 at 4:11 PM Chris Albertson wrote:
>>>On Wed, Jun 29, 2011 at 3:00 PM, Bill Dailey <docdailey at gmail.com>
>>>> The FireWire thing makes sense but is there some way to make the
>> computer itself a slave to an external source also?
>>>Why?   The computer has many clocks in it and they are not all derived
>>>all from one master.  So to answer  we'd need to know which clock
>>>care about.  Obviously the clock to keep the current time of day can
>>>be sync'd from s master source, like GPS or an Internet NTP server.
>>>Maybe it's best you describe the big picture. What are you trying to do?
>>>But if you mean the audio interface.  Then all you really need is a
>>>stable clock on that.  There is an "elastic" buffer between the audio
>>>interface and the software so the computer can go off and do
>>>"whatever" for 1/10th of a second and the interface wil buffer the
>>>data, none gets lost as long as the buffer does not overflow.
>>>Modern CPUs do not run at a nice fixed rate. they are
>>>non-deterministic  and might execute many instructions at once or even
>>>out of order.   The newer Intel chips adjust their internal clock
>>>rates to control power use and core temperature.    If you need to
>>>sync to the external world then software techniques are used
>>>Chris Albertson
>>>Redondo Beach, California
>>>time-nuts mailing list -- time-nuts at febo.com
>>>To unsubscribe, go to
>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>>and follow the instructions there.
>>>__________ Information from ESET Smart Security, version of virus
>> signature database 5851 (20110206) __________
>>>The message was checked by ESET Smart Security.
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at febo.com
>> To unsubscribe, go to
>> and follow the instructions there.
>Chris Albertson
>Redondo Beach, California
>__________ Information from ESET Smart Security, version of virus
signature database 5851 (20110206) __________
>The message was checked by ESET Smart Security.

More information about the time-nuts mailing list