[driverdan]

(2017)

I've looked at doing this a few times. I don't have references handy but there are cheaper atomic oscillators available now, under $1000. Still to expensive for me to justify it but one of these years I'll find one cheap.

[gh02t]

I have an NTP clock that uses GPS PPS and has a local TXDO. For any reasonable amount of time that you'd be out of GPS (barring nuclear war) a TXDO should be plenty sufficient for any sane time-related needs at a tiny fraction of the cost. Serving as a frequency reference for radio or precision counters or other semi-exotic (for at home) is the only reason I can see to actually have CSAC other than cool factor. Which, fair.

[jacquesm]

That 'local TXDO' is most likely being governed by some kind of phase lock on the PPS or some other divisor and may well lead to terrible clock jitter in the short term while giving you insane precision over the longer term. Allan deviation plots made by a device clocked by a standard that is better than the one that you are checking is the only way to be sure how good (or bad) things really are.

I have a couple of GPSDOs here and it is fun to play them against each other, the differences in the short term can be substantial, but over anything more than a few days they are extremely accurate.

[SAI_Peregrinus]

Also how much short-term inaccuracy there is depends on the type of oscillator used. An OCXO will (usually) have less short-term jitter and better stability with ambient temperature changes than a chip-scale atomic clock module, but will drift more than the atomic clock long-term. The really expensive setups combine various clocks to maximize stability over all the durations of interest & minimize influence from various external factors (temperature, humidity, barometric pressure, external magnetic fields, other EMI, gravitational bias, etc.).

Gravitational bias is a really fun one since you can easily demonstrate it & its' not immediately obvious. Take a quartz oscillator, and a reference oscillator. Measure the frequency of both. Turn the quartz oscillator on its side. The phase (and usually also frequency) will change: the physical quartz has inertia & gravitational mass, so changing its orientation changes how it oscillates. In one orientation the movement of the quartz atoms is perpendicular to the force of gravity, so both directions of oscillation are biased the same amount. In another orientation one the movement of the quartz atoms is parallel to the force of gravity, so the part of the crystal moving down accelerates a bit more than the part of the crystal moving up.

[jacquesm]

> An OCXO will (usually) have less short-term jitter and better stability with ambient temperature changes than a chip-scale atomic clock module, but will drift more than the atomic clock long-term.

Indeed! And that short term stability of a (good) OCXO is most impressive, you can fairly easily outperform a GPSDO and with a little work and a lot of attention to supply voltage and other external influences you can even (short term) outperform an atomic clock. But aging is a thing, you'll need to re-calibrate almost every week if you care about that kind of precision. I have a bunch of different standards to compare with each other and the first time I got lucky with a very good OCXO I was wondering if I had made some kind of mistake, it was that stable.

That gravitational bias thing sounds very interesting, I knew it existed but did not realize you could trivially demonstrate it like that, I'll have to try this.

[SAI_Peregrinus]

EEVBlog #646[1] demonstrates the gravitational bias.

[1] https://www.youtube.com/watch?v=zILwgQhjC_Q

[jacquesm]

Amazing... I have some other stuff on the go right now that takes priority but as soon as I have my hands free I'm on this.

Thank you.

[gh02t]

I'm aware, and yeah GPS is disciplining the TXCO, but I'm saying it's stable enough for non specialty uses outside of stuff like serving as a frequency reference. If you just want a clock source for NTP without the internet then all of this is already overkill.

You're also not likely to be out of GPS for particularly long stretches of time.

[pseudohadamard]

You can, or used to be able to, buy 5680 rubidium references recycled from cellular base stations for around USD 100, and then get a third-party interface board to deal with the weird voltage and comms requirements. Not sure if they're still available at that price.

[jacquesm]

Go hunt up a Rubidinium standard on Ebay, $150 or thereabouts and you'll have a pretty good standard.