Discussion:
1 pps Sync'ing
(too old to reply)
Martyn Smith
2018-06-24 10:37:44 UTC
Permalink
Hello,

I am a newbie question.

I have an application where I have a 10 MHz and 100 MHz squarewave output from my GPS frequency standard.

The frequency standard uses the PRS10 rubidium. The 100 MHz output is just a 100 MHz VCO locked to the 10 MHz.

These outputs are disciplined by the GPS's 1 pps (as far as frequency). But they are not in phase with it.

I need the rising edges of both the 10 MHz and 100 MHz squarewave outputs to be aligned with the GPS 1 pps (UTC) to within 1 ns.

Anyone already done this? I'm sure I've seen a distribution amplifier that does this at 10 MHz.

But the 100 MHz is actually the more important one that I need to align.

Regards

Steve
_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Bob kb8tq
2018-06-24 18:27:51 UTC
Permalink
Hi

You can indeed do this, the question is - do you *really* want to?

Depending a bit on your GPS module, the 1 pps output can jump around a few nanoseconds
on a second to second basis. Keeping the 100 MH edge locked implies modulating the 100 MHz
by at least a few ppb at some rate faster than 1 Hz. That will degrade the phase noise on the
outputs more than a little ….

A deeper issue is that the GPS module really isn’t reporting “GPS time” at the 1 second
level. It’s reporting GPS Time + atmospheric noise. What you would be tracking is more the bounce
in the atmosphere than anything that GPS actually is doing. Something like an L1 / L2 receiver
would help some with this.

Bob
Post by Martyn Smith
Hello,
I am a newbie question.
I have an application where I have a 10 MHz and 100 MHz squarewave output from my GPS frequency standard.
The frequency standard uses the PRS10 rubidium. The 100 MHz output is just a 100 MHz VCO locked to the 10 MHz.
These outputs are disciplined by the GPS's 1 pps (as far as frequency). But they are not in phase with it.
I need the rising edges of both the 10 MHz and 100 MHz squarewave outputs to be aligned with the GPS 1 pps (UTC) to within 1 ns.
Anyone already done this? I'm sure I've seen a distribution amplifier that does this at 10 MHz.
But the 100 MHz is actually the more important one that I need to align.
Regards
Steve
_______________________________________________
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hal Murray
2018-06-24 20:47:31 UTC
Permalink
Post by Martyn Smith
I need the rising edges of both the 10 MHz and 100 MHz squarewave outputs to
be aligned with the GPS 1 pps (UTC) to within 1 ns.
That will be interesting. You may have to get your cables and distribution
amplifier inside the PLL.

How good is your GPS? What's the night/day difference in GPS timing? (after
sawtooth correction)

What is your application? What sort of noise/jitter/wander are you sensitive
to and/or what can you tolerate?

One thing you can do is to derive the 10 MHz from the 100 MHz with a simple
divide by 10. Then run both through reclocking FFs. If they are in the same
package with the same clock, they will be closely aligned, but there will be
"noise" from the 10MHz on the 100 MHz due to different loads on the power
supply wiring. That's probably sub ns.

Do you need your PPS to be aligned to UTC? If not, divide down to make your
own PPS and run it through the same reclocking.

If you do need your PPS aligned to UTC, you can PLL your PPS with the PPS from
GPS. The S part of PPS means that things are slow so a software/DAC solution
will probably be appropriate.
--
These are my opinions. I hate spam.



_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Tom Van Baak
2018-06-25 14:38:19 UTC
Permalink
Steve,
Post by Martyn Smith
I need the rising edges of both the 10 MHz and 100 MHz squarewave outputs
to be aligned with the GPS 1 pps (UTC) to within 1 ns.
Two comments.

1) If you divide 100 MHz -> 10 MHz -> 1 Hz then (with careful design, calibrated layout / wiring) you can have all three rising edges within 1 ns *of each other*. I assume this is your goal. If you already have a 100 MHz GPSDO then it would surprise me if the outputs are not already synchronized with each other. I mean, almost all GPSDO give a 1PPS output that is derived from the LO, by design. If your goal is 1 ns, you obviously do not want to use the raw 1PPS from the receiver. What make/model GPSDO are you using?

2) On the other hand, if your goal is that all the outputs are within 1 ns *of UTC*, then you have a big problem. Not even national timing laboratories, with a pile of cesium and H-maser references, can get their UTC(k) within 1 ns of UTC. Perhaps you can explain a bit more what you are actually doing. It sounds like an interesting application. Relative timing at the 1 ns level is easy. Absolute UTC timing at the ns level is much, much harder.

/tvb

----- Original Message -----
From: "Martyn Smith" <***@live.co.uk>
To: <time-***@lists.febo.com>
Sent: Sunday, June 24, 2018 3:37 AM
Subject: [time-nuts] 1 pps Sync'ing
Post by Martyn Smith
Hello,
I am a newbie question.
I have an application where I have a 10 MHz and 100 MHz squarewave output from my GPS frequency standard.
The frequency standard uses the PRS10 rubidium. The 100 MHz output is just a 100 MHz VCO locked to the 10 MHz.
These outputs are disciplined by the GPS's 1 pps (as far as frequency). But they are not in phase with it.
I need the rising edges of both the 10 MHz and 100 MHz squarewave outputs to be aligned with the GPS 1 pps (UTC) to within 1 ns.
Anyone already done this? I'm sure I've seen a distribution amplifier that does this at 10 MHz.
But the 100 MHz is actually the more important one that I need to align.
Regards
Steve
_______________________________________________
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Bob kb8tq
2018-06-25 14:46:54 UTC
Permalink
Hi

On a Time Nut metaphysical note -
Post by Tom Van Baak
Steve,
Post by Martyn Smith
I need the rising edges of both the 10 MHz and 100 MHz squarewave outputs
to be aligned with the GPS 1 pps (UTC) to within 1 ns.
Two comments.
1) If you divide 100 MHz -> 10 MHz -> 1 Hz then (with careful design, calibrated layout / wiring) you can have all three rising edges within 1 ns *of each other*. I assume this is your goal. If you already have a 100 MHz GPSDO then it would surprise me if the outputs are not already synchronized with each other. I mean, almost all GPSDO give a 1PPS output that is derived from the LO, by design. If your goal is 1 ns, you obviously do not want to use the raw 1PPS from the receiver. What make/model GPSDO are you using?
2) On the other hand, if your goal is that all the outputs are within 1 ns *of UTC*, then you have a big problem. Not even national timing laboratories, with a pile of cesium and H-maser references, can get their UTC(k) within 1 ns of UTC. Perhaps you can explain a bit more what you are actually doing. It sounds like an interesting application. Relative timing at the 1 ns level is easy. Absolute UTC timing at the ns level is much, much harder.
Does UTC exist ( real time ) at the 1 ns level?

Bob
Post by Tom Van Baak
/tvb
----- Original Message -----
Sent: Sunday, June 24, 2018 3:37 AM
Subject: [time-nuts] 1 pps Sync'ing
Post by Martyn Smith
Hello,
I am a newbie question.
I have an application where I have a 10 MHz and 100 MHz squarewave output from my GPS frequency standard.
The frequency standard uses the PRS10 rubidium. The 100 MHz output is just a 100 MHz VCO locked to the 10 MHz.
These outputs are disciplined by the GPS's 1 pps (as far as frequency). But they are not in phase with it.
I need the rising edges of both the 10 MHz and 100 MHz squarewave outputs to be aligned with the GPS 1 pps (UTC) to within 1 ns.
Anyone already done this? I'm sure I've seen a distribution amplifier that does this at 10 MHz.
But the 100 MHz is actually the more important one that I need to align.
Regards
Steve
_______________________________________________
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
_______________________________________________
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Attila Kinali
2018-06-25 16:37:20 UTC
Permalink
On Mon, 25 Jun 2018 10:46:54 -0400
Post by Bob kb8tq
Does UTC exist ( real time ) at the 1 ns level?
Currently, it does not. The clock comparison uncertainties are
still in the order of 1-3ns for short distances (several 100km)
and grow quite a bit once you go over something like 2000-3000km.
Unless you are using a dedicated fiber. But very few labs can
afford to have a fiber connecting them to another lab.

Also consider that, even within a lab, doing sub-ns accurate
time comparisons is challenging. All delays and their drifts
need to be precisely measured, characterized and factored in.
There is a reason why the SR-620 is still the main workhorse
in metrology labs, even though we have equipment available
that beats the SR-620 in terms of resolution and stability.
It's just not necessary to go below 20ps as the other factors
cannot be eliminated with that high precision/accuracy anyways.

If only precision time comparison is required, IIRC the current
limit is around 50-100fs. And even that requires a lot of work
and characterization.

Attila Kinali
--
It is upon moral qualities that a society is ultimately founded. All
the prosperity and technological sophistication in the world is of no
use without that foundation.
-- Miss Matheson, The Diamond Age, Neil Stephenson
_______________________________________________
time-nuts mailing list -- time-***@lists.febo.com
To unsubscribe, go to https://lists.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Loading...