Re: [time-nuts] GPSDO Alternatives
Hi
The solution to the problem is well known in several forms. Cost is below $5 for pretty much all of them. No need to re-invent the wheel. The gotcha is that you can't do it 100% with internal CPU peripherals. You will need some glue.
Bob
On Dec 6, 2012, at 7:43 PM, Chris Albertson albertson.chris@gmail.com wrote:
What if the OCXO had a sine wave output and then you used the PPS leading
edge to gate the sine wave to a sample and hold. then the sample is
measured by the Arduino's ADC? I think(?) you get a 10-bit ADC or is it
8-bits. The problem is the required speed of the sample and hold, maybe
not easy to build
Chris Albertson
Redondo Beach, California
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
It is not a 16 bit capture and it does not run at 1/4 the clock rate.
The Shera uses a 12 bit counter to capture the phase difference
between the OCXO frequency divided by 16 and GPS pulse per second
output to a resolution of about 42ns.
What I suggested effectively captures the same phase difference but
divides the OCXO by 65536 and only has a resolution of 100ns. The
OCXO division is just conceptual though as it comes from the 16 bit
timer/counter overflow internal to the microcontroller so in practice,
the timer/counter is arbitrary length with an interrupt about every
6.5ms doing the housekeeping for the extra bits. It directly measures
the pulse per second period to a resolution of 100ns using the OCXO as
a clock. If a 20 MHz OCXO was used (limited by the maximum clock
frequency of the microcontroller), then the resolution would be 50ns.
That is not how I plan on designing my own GPSDO which has taken a
significant turn from what I posted about here some time ago but I
have stopped discussing that until I have some results to share.
On Thu, 6 Dec 2012 19:34:08 -0500, Bob Camp lists@rtty.us wrote:
Hi
Again, the math is pretty simple.
A 16 bit capture running at a 1/4 clock is not going to get you very near a Shera. It's even further from the more modern "enhanced Shera" designs.
Bob
On Dec 6, 2012, at 6:59 PM, David davidwhess@gmail.com wrote:
You can use the ATmega328 16 bit timer/counter in input capture mode
to count the number of 10 MHz OCXO cycles per pulse per second period
to a resolution of 100ns but there are some problems:
The ATmega328 16 bit timer/counter external clock is limited to 1/4 of
the CPU frequency with an asynchronous source so the 10 MHz OCXO would
need to be divided down which would further limit performance and
require an external divider. Modifying the Aruino board to use the 10
MHz OCXO in place of the CPU clock solves that problem.
Then operating the counter/timer in input capture mode with the GPS
pulse per second signal connected to the input capture pin would allow
almost Shera like performance. The timing resolution would be 2.4
times lower (and not asynchronous) limiting performance over short
time spans.
On Thu, 6 Dec 2012 14:57:19 -0800, Chris Albertson
albertson.chris@gmail.com wrote:
You'd have to seriously divide down the output from the 10MHz OCXO if you
were going to use it as an interrupt. Maybe to divide by 10,000? and even
at the higher clock rate you'd still have poor resolution.
I image each interrupt handler would sample some internal counter and the
background task would look at the delta between the two and adjust the DAC
to drive the OCXO to close the difference. The resolution would be
(maybe?) a "handful" of clock cycles. Given enough time, say a 1000
second period it might wrk well enough. I can't know without doing a more
detailed design
On Thu, Dec 6, 2012 at 1:59 AM, Don Latham djl@montana.com wrote:
Chris:
The question I have again is about a simple phase detector.
I did ask if the arduino interrupt ports could be used as a phase
detector; one on the GPS and one on the OXCO. Too much jitter? If the 12
MHz clock is too slow, would an 80 MHz clock ARM arduino style processor
work? I'm simply too new at this to decide.
Don
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
There are lots of sampling ADCs which will support that type of
operation directly or you can easily design and build a sampling phase
detector but that all involves significant extra circuitry outside of
the microcontroller.
Take a look at the Racal Dana 1992 reference frequency multiplier
option (the schematic is on page 7-33 of the service manual) for an
example of a sampling phase detector used in a similar application.
On Thu, 6 Dec 2012 16:43:26 -0800, Chris Albertson
albertson.chris@gmail.com wrote:
What if the OCXO had a sine wave output and then you used the PPS leading
edge to gate the sine wave to a sample and hold. then the sample is
measured by the Arduino's ADC? I think(?) you get a 10-bit ADC or is it
8-bits. The problem is the required speed of the sample and hold, maybe
not easy to build
Sorry. The Shera counter is 16 bits and not 12 bits but that does not
change what I posted.
On Thu, 06 Dec 2012 19:17:48 -0600, David davidwhess@gmail.com
wrote:
It is not a 16 bit capture and it does not run at 1/4 the clock rate.
The Shera uses a 12 bit counter to capture the phase difference
between the OCXO frequency divided by 16 and GPS pulse per second
output to a resolution of about 42ns.
What I suggested effectively captures the same phase difference but
divides the OCXO by 65536 and only has a resolution of 100ns. The
OCXO division is just conceptual though as it comes from the 16 bit
timer/counter overflow internal to the microcontroller so in practice,
the timer/counter is arbitrary length with an interrupt about every
6.5ms doing the housekeeping for the extra bits. It directly measures
the pulse per second period to a resolution of 100ns using the OCXO as
a clock. If a 20 MHz OCXO was used (limited by the maximum clock
frequency of the microcontroller), then the resolution would be 50ns.
That is not how I plan on designing my own GPSDO which has taken a
significant turn from what I posted about here some time ago but I
have stopped discussing that until I have some results to share.
Chris: yes, dividing would have to be done. Doesn't TVB have a simple
divider block?
You don't really have to "close" the difference, just maintain it?
Don L
Chris Albertson
You'd have to seriously divide down the output from the 10MHz OCXO if
you
were going to use it as an interrupt. Maybe to divide by 10,000? and
even
at the higher clock rate you'd still have poor resolution.
I image each interrupt handler would sample some internal counter and
the
background task would look at the delta between the two and adjust the
DAC
to drive the OCXO to close the difference. The resolution would be
(maybe?) a "handful" of clock cycles. Given enough time, say a 1000
second period it might wrk well enough. I can't know without doing a
more
detailed design
On Thu, Dec 6, 2012 at 1:59 AM, Don Latham djl@montana.com wrote:
Chris:
The question I have again is about a simple phase detector.
I did ask if the arduino interrupt ports could be used as a phase
detector; one on the GPS and one on the OXCO. Too much jitter? If the
12
MHz clock is too slow, would an 80 MHz clock ARM arduino style
processor
work? I'm simply too new at this to decide.
Don
--
"Neither the voice of authority nor the weight of reason and argument
are as significant as experiment, for thence comes quiet to the mind."
De Erroribus Medicorum, R. Bacon, 13th century.
"If you don't know what it is, don't poke it."
Ghost in the Shell
Dr. Don Latham AJ7LL
Six Mile Systems LLP
17850 Six Mile Road
POB 134
Huson, MT, 59846
VOX 406-626-4304
www.lightningforensics.com
www.sixmilesystems.com
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
--
Chris Albertson
Redondo Beach, California
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
--
"Neither the voice of authority nor the weight of reason and argument
are as significant as experiment, for thence comes quiet to the mind."
De Erroribus Medicorum, R. Bacon, 13th century.
"If you don't know what it is, don't poke it."
Ghost in the Shell
Dr. Don Latham AJ7LL
Six Mile Systems LLP
17850 Six Mile Road
POB 134
Huson, MT, 59846
VOX 406-626-4304
www.lightningforensics.com
www.sixmilesystems.com
Hi
The Shera counter is not running in the same fashion you would be running an input capture pin.
Bob
On Dec 6, 2012, at 8:45 PM, David davidwhess@gmail.com wrote:
Sorry. The Shera counter is 16 bits and not 12 bits but that does not
change what I posted.
On Thu, 06 Dec 2012 19:17:48 -0600, David davidwhess@gmail.com
wrote:
It is not a 16 bit capture and it does not run at 1/4 the clock rate.
The Shera uses a 12 bit counter to capture the phase difference
between the OCXO frequency divided by 16 and GPS pulse per second
output to a resolution of about 42ns.
What I suggested effectively captures the same phase difference but
divides the OCXO by 65536 and only has a resolution of 100ns. The
OCXO division is just conceptual though as it comes from the 16 bit
timer/counter overflow internal to the microcontroller so in practice,
the timer/counter is arbitrary length with an interrupt about every
6.5ms doing the housekeeping for the extra bits. It directly measures
the pulse per second period to a resolution of 100ns using the OCXO as
a clock. If a 20 MHz OCXO was used (limited by the maximum clock
frequency of the microcontroller), then the resolution would be 50ns.
That is not how I plan on designing my own GPSDO which has taken a
significant turn from what I posted about here some time ago but I
have stopped discussing that until I have some results to share.
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hi
Unless you really want to go crazy with measuring very long delays, you do indeed want to align the pps from your OCXO with the pps from your GPS.
Bob
On Dec 6, 2012, at 8:48 PM, "Don Latham" djl@montana.com wrote:
Chris: yes, dividing would have to be done. Doesn't TVB have a simple
divider block?
You don't really have to "close" the difference, just maintain it?
Don L
Chris Albertson
You'd have to seriously divide down the output from the 10MHz OCXO if
you
were going to use it as an interrupt. Maybe to divide by 10,000? and
even
at the higher clock rate you'd still have poor resolution.
I image each interrupt handler would sample some internal counter and
the
background task would look at the delta between the two and adjust the
DAC
to drive the OCXO to close the difference. The resolution would be
(maybe?) a "handful" of clock cycles. Given enough time, say a 1000
second period it might wrk well enough. I can't know without doing a
more
detailed design
On Thu, Dec 6, 2012 at 1:59 AM, Don Latham djl@montana.com wrote:
Chris:
The question I have again is about a simple phase detector.
I did ask if the arduino interrupt ports could be used as a phase
detector; one on the GPS and one on the OXCO. Too much jitter? If the
12
MHz clock is too slow, would an 80 MHz clock ARM arduino style
processor
work? I'm simply too new at this to decide.
Don
--
"Neither the voice of authority nor the weight of reason and argument
are as significant as experiment, for thence comes quiet to the mind."
De Erroribus Medicorum, R. Bacon, 13th century.
"If you don't know what it is, don't poke it."
Ghost in the Shell
Dr. Don Latham AJ7LL
Six Mile Systems LLP
17850 Six Mile Road
POB 134
Huson, MT, 59846
VOX 406-626-4304
www.lightningforensics.com
www.sixmilesystems.com
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
--
Chris Albertson
Redondo Beach, California
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
--
"Neither the voice of authority nor the weight of reason and argument
are as significant as experiment, for thence comes quiet to the mind."
De Erroribus Medicorum, R. Bacon, 13th century.
"If you don't know what it is, don't poke it."
Ghost in the Shell
Dr. Don Latham AJ7LL
Six Mile Systems LLP
17850 Six Mile Road
POB 134
Huson, MT, 59846
VOX 406-626-4304
www.lightningforensics.com
www.sixmilesystems.com
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hi
If all you want is a "something" locked to a GPS:
Take the pps from the GPS and hook it to an AD9548. You probably will need a 50 cent CPU to set up the registers. No muss, no fuss, nothing to invent or design.
Weather it does what you need to do is an entirely different question. Without a defined objective / need / performance goal this could go on for a couple hundred years…..
Bob
On Dec 6, 2012, at 7:43 PM, Chris Albertson albertson.chris@gmail.com wrote:
What if the OCXO had a sine wave output and then you used the PPS leading
edge to gate the sine wave to a sample and hold. then the sample is
measured by the Arduino's ADC? I think(?) you get a 10-bit ADC or is it
8-bits. The problem is the required speed of the sample and hold, maybe
not easy to build
Chris Albertson
Redondo Beach, California
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Good thought, Bob. AD9548 $27, eval board a whopping $250, get a
thunderbolt :-). The eval board has a lot of SMA's on it...
Don L
Bob Camp
Hi
If all you want is a "something" locked to a GPS:
Take the pps from the GPS and hook it to an AD9548. You probably will
need a 50 cent CPU to set up the registers. No muss, no fuss, nothing
to invent or design.
Weather it does what you need to do is an entirely different question.
Without a defined objective / need / performance goal this could go on
for a couple hundred years
..
Bob
On Dec 6, 2012, at 7:43 PM, Chris Albertson albertson.chris@gmail.com
wrote:
What if the OCXO had a sine wave output and then you used the PPS
leading
edge to gate the sine wave to a sample and hold. then the sample is
measured by the Arduino's ADC? I think(?) you get a 10-bit ADC or is
it
8-bits. The problem is the required speed of the sample and hold,
maybe
not easy to build
Chris Albertson
Redondo Beach, California
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
--
"Neither the voice of authority nor the weight of reason and argument
are as significant as experiment, for thence comes quiet to the mind."
De Erroribus Medicorum, R. Bacon, 13th century.
"If you don't know what it is, don't poke it."
Ghost in the Shell
Dr. Don Latham AJ7LL
Six Mile Systems LLP
17850 Six Mile Road
POB 134
Huson, MT, 59846
VOX 406-626-4304
www.lightningforensics.com
www.sixmilesystems.com
Hi
In this case it's very much a "you get what you pay for" sort of thing. You are indeed comparing an hourglass to a cesium standard.
Bob
On Dec 7, 2012, at 2:20 AM, "Don Latham" djl@montana.com wrote:
Good thought, Bob. AD9548 $27, eval board a whopping $250, get a
thunderbolt :-). The eval board has a lot of SMA's on it...
Don L
Bob Camp
Hi
If all you want is a "something" locked to a GPS:
Take the pps from the GPS and hook it to an AD9548. You probably will
need a 50 cent CPU to set up the registers. No muss, no fuss, nothing
to invent or design.
Weather it does what you need to do is an entirely different question.
Without a defined objective / need / performance goal this could go on
for a couple hundred years…..
Bob
On Dec 6, 2012, at 7:43 PM, Chris Albertson albertson.chris@gmail.com
wrote:
What if the OCXO had a sine wave output and then you used the PPS
leading
edge to gate the sine wave to a sample and hold. then the sample is
measured by the Arduino's ADC? I think(?) you get a 10-bit ADC or is
it
8-bits. The problem is the required speed of the sample and hold,
maybe
not easy to build
Chris Albertson
Redondo Beach, California
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
--
"Neither the voice of authority nor the weight of reason and argument
are as significant as experiment, for thence comes quiet to the mind."
De Erroribus Medicorum, R. Bacon, 13th century.
"If you don't know what it is, don't poke it."
Ghost in the Shell
Dr. Don Latham AJ7LL
Six Mile Systems LLP
17850 Six Mile Road
POB 134
Huson, MT, 59846
VOX 406-626-4304
www.lightningforensics.com
www.sixmilesystems.com
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.