[time-nuts] Simple solution for disciplining OCXO with 1 PPS

[Tom Van Baak]

> Can a pure analog design access the sawtooth correction?  My GPS
> receivers send sawtooth as a digital message on a serial port.   I
> don't know if saw tooth correction is required to meet his spec.

Hi Chris,

Many GPS/1PPS receivers don't output sawtooth information and yet they work really well. Sawtooth correction isn't necessary except for high-end GPSDO. The same is true for zero-D timing mode.

Examples include:
https://www.adafruit.com/products/746
https://www.parallax.com/product/28509
as well as any number of equivalent GPS/1PPS boards at half the price on eBay. See also:
http://www.leapsecond.com/pages/MG1613S/

In fact many simple GPSDO rely on sawtooth dither to improve their performance; that is, it's a feature to be exploited, not a bug to be corrected for.

A dirt cheap GPS receiver these days may have 1PPS error of 20 ns RMS. Over 1000 s integration that's 2e-11 in frequency stability; far better than the OP's modest target of 1e-10 at an hour.

Here's another thought. OP wants STS of stability of 1e-10 at an hour. No mention of accuracy, or long-term performance. So if stability is all that's needed I would just use a old 10811-class OCXO, perhaps one that's been running faithfully for a few weeks or months. That will get you 1e-10 at an hour without any GPS any antenna any analog any digital any complexity any tuning.

/tvb