Performance Tests with SAO VLG-10-1 Maser
(Serial P2) at Morehead State University

The Morehead State University Space Science Center operates a state-of-the-art 21 meter radio telescope and has recently become a station in the NASA Deep Space Network. That activity requires a very stable source of frequency and time, and to facilitate this Morehead was given use of a (quite old) hydrogen maser frequency standard that was built in 1974 by the Smithsonian Astrophysical Observatory. It is a model VLG-10-1 that was refurbished in 2016 at MIT Haystack Observatory.

Morehead is about 3 hours drive from where I live in Dayton, Ohio, and the director of engineering for the telescope is a ham radio friend of mine. A while ago we started talking about bringing some of my time and frequency measurement gear to MSU to compare it against the maser.

I have some pretty good frequency standards, but the maser should be more stable than any of them, except at very short (<10 second) and very long (>100000 second) averaging times. Comparing them to the maser will tell more about my standards than it does about the maser, but we can at least determine the worst the maser is doing.

On November 19, 2018, I loaded up my car with gear and drove to Morehead. I took along a TAPR TICC timestamping counter and a Miles TimePod, a/ka/ Microsemi 3120A stability and phase noise measurement probe to collect data, along with a couple of laptop computers for logging.

I brought four frequency standards, each intended to measure one area of the maser's performance:

There was also a PPS (pulse per second) signal from a GPS disciplined oscillator available in the measurement room. At the beginning of the campaign, we set up two sets of long-term measurement.

In one, the 10 MHz signal from the 5071A cesium standard was used as the reference to the TimePod analyzer, which was configured to measure the 10 MHz signals from the maser and the 5065A rubidium standard.

In the other, the maser provided a 10 MHz reference signal for the TICC timestamping counter. PPS signals from the GPSDO and 5071A drove the channel A and B inputs of the TICC. This configuration allowed the TimePod to also measure the maser vs. 5065A, and the TICC to also measure the 5071A vs. GPDSO by using the reference inputs as a transfer standard.

This setup provided several cross-checks to help validate the measurements:

This measurement ran until December 10, 2018 -- just short of 21 days, collecting 1.81 million samples.

Here is an analysis of the long term results.

And here is an analysis of the short term results.