GPS 1PPS

Searching on "GPS 1PPS jitter" gets some interesting results. Here's the one I liked best, mostly because it has pictures:

--
Tim Wescott 
Control systems, embedded software and circuit design 
I'm looking for work!  See my website if you're interested 
http://www.wescottdesign.com

Good stuff, thanks.

I wish I had time to research this better; I'm still typing proposal.

Sounds like the main output of GPS is a fairly jittery 1 PPS, and the clock is derived from that, and there are a zillion ways to do that.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Google has time-nuts on it's radar so I use site search. Go to: and inscribe: site:

snipped-for-privacy@febo.com/ into the search box, followed by your key word collection. For example, searching for 1pps clock sync results in:

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

That sounds like what I know. If you're a well-funded lab with time on your hands you could purpose-design a GPS receiver and get a fairly jittery N PPS, with N > 1, up to the message rate of the satellites.

If that helps...

--
Tim Wescott 
Control systems, embedded software and circuit design 
I'm looking for work!  See my website if you're interested 
http://www.wescottdesign.com

The clock that I will get is LVDS, somewhere around 100 MHz, and I'm not allowed to say exactly what is the frequency. Yes, that's silly.

I once, when I was still an undergrad, had to get a security clearance to be told the frequency to design an LC filter for, still having no idea what the application was. They finally cleared me and revealed the top-secret value, which was 3.5 KHz or something like that. I can't reveal the exact frequency, or somebody would have to kill you all.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Maybe I'm reading too much into your description, but if the 100 MHz is "derived from GPS", and the 1 PPS is certainly "derived from GPS", then they must be coherent or at least in sync with each other. However, if the 1 PPS comes from the receive signal, and the 100 MHz comes from an onboard clock oscillator, unless the onboard clock comes from a GPSDO (GPS disciplined oscillator), the two signals are not going to be in sync. I suggest you re-read the spec, and look for any accidental ambiguity in what is meant by "derived from GPS". It might just mean it comes from the same box as the GPS receiver.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

A jittery 1 PPS reference requires a giant flywheel, and waiting a long time for it to spin up.

The best GPS frequency sources discipline an extreme OCXO or a rubidium clock, which generally has an OCXO inside. The phase noise is dominated by the XO, and all the rubidium does is provide good holdover.

It is sort of an obscure thing to get obsessed about. Things with motors or sparks or loud noises are more fun.

It's interesting to trigger a scope from one good clock and look at the rising edge of another one, at maybe 5 ns/cm. Check that every few hours and see if it's moved.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

It would be nice if they were.

I don't actually have a spec. My input has been emails and phone calls from far, far away. So my proposal is based on a lot of guesswork about an application that can't be revealed.

This sort of situation is not uncommon. My proposal becomes the system specification. I have to do all the typing. It's almost done!

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Your real question is how do you write the proposal. The answer is either

1) you don't and forget the job 2) write a worst case scenario proposal and include caveats. The caveats should evoke responses that will clarify the design requirements. Otherwise you may be stuck with a job with no positive financial outcome.

But I am sure you already understand all of that.

They have sent you on a snipe hunt down a rabbit hole populated by wild geese. Tell them you need more time.

-- john, KE5FX

At least 1 hobby magazine has published a 1Hz GPS time reference, although I think they had a 10MHz XO.

Can't remember if it was Elektor or EPE - maybe both.

You are right, he didn't say it *was* 100 MHz, he said, "If the clock were, say, 100 MHz". My bad, but so what? The frequency of the clock is not the issue.

Of course, but if he has no control over it, which it seems he doesn't, he has to design around that issue.

You have strayed too far into allegory... My point is if he needs to generate the 1 PPS synchronized to a clock, I have a friend who has done exactly that who should be able to assist him saving him time, effort and money. But the devil is in the details. Until the details are shared, no one will know what will help John.

--

Rick C

That can be hard to do when you have so little information to work with.

Can you discuss some of what you will do with the signals? So far I have not seen anyone offer any useful solutions which is largely because you have not fully explained the problem.

--

Rick C

Why not a VCOCXO? They exist.

Some of the later mechanical clocks were quite accurate. I believe some measurements on the Shortt clock place it to 1 second in about 12 years. This is hugely better than quartz watches everyone thinks are so good.

--

Rick C

Thanks for the info.

--

Rick C

Because noise on the control terminal introduces jitter.

A great rock running dead-center where it wants to will always beat pulling a great rock slightly off-kilter to where *you* want it.

There was a great Amateur Scientist article on that. In the '50's, IIRC.

Cheers, James Arthur

Except that the DDS has limits on noise too. How low can the noise on a VCOCXO get? Depending on frequency a DDS can get about 1 ppm pretty easily at 100 MHz, I expect a few more bits isn't hard. But that's staying in the digital domain. If you need an analog signal the limit is the DAC for sure.

Why can't the noise on the control terminal be reduced enough?

--

Rick C

One OXCO that we are considering has a good SC-cut crystal with a VCO pull range of +-0.5 PPM. Varicaps are nasty so the pull range has to be small to get good phase noise. It has a mechanical coarse frequency adjust, probably a sapphire piston cap or something exotic like that. It costs over $1000.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics

Yes. We were given a week to get something together.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics