// draft // 11-Mar-2006

Keywords: Symmetricom Datum Sigma-Tau Hydrogen Maser Cavity-Auto-Tuned Short-term Long-term Stability Allan deviation

Introduction

The Register Speed Multiplier (RSM) of the Sigma Tau MHM-2010 hydrogen maser is configured by two small toggle switches labeled "4x" and "2x" located behind the front panel. Interpreting these as two binary digits allows one to choose one of four RSM rates: 1x, 2x, 4x, or 8x. The expectation is that 1x gives better short-term stability and 8x better long-term stability (?)

It has been quite difficult for me to accurately measure the difference in maser performance for each switch setting. It requires, over the span of many days, a very stable lab environment and a very stable reference; down to the levels around 1×10^-15.

My lab is still not temperature stable to 1 ºC levels (goal is 0.1 ºC) and the reference in this case is a KVARZ CH1-75 active hydrogen maser; which during this time had its own set of performance challenges. Nevertheless, below are two runs made in October 2005 which give a hint of the performance difference in the RSM rate. I believe the auto-tuning of the CH1-75 reference was disabled during these runs (mode F_runM rather than AFCC-I).

As seen in the photo below the RSM toggle switches are located under the top right corner of the front panel. The entire front panel must be removed in order to access the switches (they are not intended to be changed often; but the operation can be done live while the maser is operating).

The labels aren't totally clear but the email below clarifies the situation.

> Are the 4x and 2x RSM switches behind the front
> panel "on" when pushed to the left or right? I.e.,
> is 1x (slow) when both are left and 8x (fast) when
> both are right?

Your assumption is correct, right is multiply (Fast), left is X1 (slow)

> Approximately what is the time constant at either extreme?

I don't remember an exact time constant without going through the system and
doing a calculation, however, in the slowest speed the action of the tuner
is usually seen as a slight bump in the stability curve around 10000s, and
at the fastest the bump moves to the left and peaks around 1000s.

OK, now for the plots.

Stability with RSM = 1x

Both RSM switches set LEFT (1x)

Stability with RSM = 8x

Both RSM switches set RIGHT (8x)

Conclusion

Although it would have been nice to have had longer runs (tau out to a day, 10⁵ s, or even 1 week or 10⁶ s), or multiple runs (to give some confidence of run-to-run consistency), or runs at each of all four switch positions, this pair of runs gives a good idea of the difference between the 1x and 8x positions.

These plots show that the ADEV most affected by the change is tau 200 s to 4000 s. In particular, at around tau 1 hour, the ADEV gets about 2½ times better (about 2.5×10^-15 vs. 6×10^-15) in the RSM 1x position.

Other thoughts:

• It is possible that this 2½ factor is exaggerated by the temperature instability in the lab. In other words, if the lab had perfect temperature control perhaps the difference would not be so great. I don't know. That's yet another experiment to try.
• One idea to better measure the difference between RSM settings is to use 1x on odd days and 8x on even days while collecting continuous phase data. Gradual changes in the environment or the reference standard would then more equally apply to "both" runs. After raw data is collected for several weeks the 1x and 8x days could be split out in software to create two separate ADEV plots.

Links:

• Symmetricom's MHM 2010 Active Hydrogen Maser
• USNO Hydrogen Masers