time-nuts@lists.febo.com

Discussion of precise time and frequency measurement

View all threads

News from NPL

RK
Rob Kimberley
Fri, Dec 15, 2006 5:15 PM

Just put two recent releases from NPL on my web site -

Autumn 2006 T&F News

http://www.timing-consultants.com/images/Web%20Site%20App%20Notes/NPL%20TF%2
0News%20Autumn%2006.pdf

MSF Update (UK's 60 KHz Time Tx)

http://www.timing-consultants.com/images/Web%20Site%20App%20Notes/MSF%20leaf
let.pdf

Rob K

Just put two recent releases from NPL on my web site - Autumn 2006 T&F News http://www.timing-consultants.com/images/Web%20Site%20App%20Notes/NPL%20TF%2 0News%20Autumn%2006.pdf MSF Update (UK's 60 KHz Time Tx) http://www.timing-consultants.com/images/Web%20Site%20App%20Notes/MSF%20leaf let.pdf Rob K
RK
Rob Kimberley
Fri, Dec 15, 2006 5:28 PM

I see that the links didn't get through the system correctly.

If anyone has a problem, just go to

http://www.timing-consultants.com and hit the Application Notes page.

And yes, I know I need to take out the blank spaces in the file names which
get translated to "%20"

:-)

Rob K

-----Original Message-----
From: time-nuts-bounces@febo.com [mailto:time-nuts-bounces@febo.com] On
Behalf Of Rob Kimberley
Sent: 15 December 2006 17:16
To: 'Discussion of precise time and frequency measurement'
Subject: [time-nuts] News from NPL

Just put two recent releases from NPL on my web site -

Autumn 2006 T&F News

http://www.timing-consultants.com/images/Web%20Site%20App%20Notes/NPL%20TF%2
0News%20Autumn%2006.pdf

MSF Update (UK's 60 KHz Time Tx)

http://www.timing-consultants.com/images/Web%20Site%20App%20Notes/MSF%20leaf
let.pdf

Rob K


time-nuts mailing list
time-nuts@febo.com
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts

I see that the links didn't get through the system correctly. If anyone has a problem, just go to http://www.timing-consultants.com and hit the Application Notes page. And yes, I know I need to take out the blank spaces in the file names which get translated to "%20" :-) Rob K -----Original Message----- From: time-nuts-bounces@febo.com [mailto:time-nuts-bounces@febo.com] On Behalf Of Rob Kimberley Sent: 15 December 2006 17:16 To: 'Discussion of precise time and frequency measurement' Subject: [time-nuts] News from NPL Just put two recent releases from NPL on my web site - Autumn 2006 T&F News http://www.timing-consultants.com/images/Web%20Site%20App%20Notes/NPL%20TF%2 0News%20Autumn%2006.pdf MSF Update (UK's 60 KHz Time Tx) http://www.timing-consultants.com/images/Web%20Site%20App%20Notes/MSF%20leaf let.pdf Rob K _______________________________________________ time-nuts mailing list time-nuts@febo.com https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
PS
Pete Slagle
Fri, Dec 15, 2006 8:40 PM

Rob Kimberley wrote:

I see that the links didn't get through the system correctly.

http://tinyurl.com reliably tames unwieldy links.

Rob Kimberley wrote: > I see that the links didn't get through the system correctly. http://tinyurl.com reliably tames unwieldy links.
DF
David Forbes
Wed, Dec 20, 2006 5:22 PM

A discussion of driving computer clocks from accurate sources recently
came up. While it's true that PCs generally use the CPU clock or the
baud rate clock for the timekeeping function, the observation was made
that it might be nice to run a standard RTC that uses a 32768 Hz crystal
from a 10 MHz source. The TAPR Clock Block is not designed to do this,
but it is easy to do using a non-PLL approach.

Tom Van Baak has programmed a PIC microcontroller to make all decades of
frequencies from a 10 MHz source. In a similar way, the PIC can be
programmed to produce any frequency desired, such as 32768 Hz, using the
method of a dual-modulus prescaler.

If you're not familiar with this technique, read the Wikipedia entry:
http://en.wikipedia.org/wiki/Dual-modulus_prescaler

Since a PIC executes 2.500 MIPS when clocked at 10 MHz, its effective
input frequency is 2.5000 MHz. The output frequency is 32768 Hz. Thus
the modulus M is 76, and the remainder A is 9632.

This means that you would execute 9632 loops that take 77 clock cycles,
followed by (32768-9632) = 23136 loops that take 76 clock cycles, to
make exactly 32768 output cycles per second given a 10 MHz CPU clock.

The code for this should fill considerably less than a page of text.

Bear in mind that you would be making a very ugly 32768 Hz signal if
viewed in the frequency domain, so only use the above method for
timekeeping.

A discussion of driving computer clocks from accurate sources recently came up. While it's true that PCs generally use the CPU clock or the baud rate clock for the timekeeping function, the observation was made that it might be nice to run a standard RTC that uses a 32768 Hz crystal from a 10 MHz source. The TAPR Clock Block is not designed to do this, but it is easy to do using a non-PLL approach. Tom Van Baak has programmed a PIC microcontroller to make all decades of frequencies from a 10 MHz source. In a similar way, the PIC can be programmed to produce any frequency desired, such as 32768 Hz, using the method of a dual-modulus prescaler. If you're not familiar with this technique, read the Wikipedia entry: http://en.wikipedia.org/wiki/Dual-modulus_prescaler Since a PIC executes 2.500 MIPS when clocked at 10 MHz, its effective input frequency is 2.5000 MHz. The output frequency is 32768 Hz. Thus the modulus M is 76, and the remainder A is 9632. This means that you would execute 9632 loops that take 77 clock cycles, followed by (32768-9632) = 23136 loops that take 76 clock cycles, to make exactly 32768 output cycles per second given a 10 MHz CPU clock. The code for this should fill considerably less than a page of text. Bear in mind that you would be making a very ugly 32768 Hz signal if viewed in the frequency domain, so only use the above method for timekeeping.