A Contradiction: Setting the Gain Values

 The flicker noise gain, K_f, the white noise gain K_w and the clock bandwidth B have been set at values to allow the model to satisfy the masks. This manifestly contradicts an assertion made in the standards that the masks are set from the model. It is stated in [4] that, ``The model used to derive these numbers is described in (informative) Annex A'', Annex A being the part of the document detailing the clock model. The position at the time of writing is that the model has been changed to allow it to satisfy the masks although it is still stated that the masks have been derived from the model. From correspondence with the author of [14] reproduced in section A, it seems that the masks were defined through considerations other than the model and thus the model must satisfy them rather than vice versa.

For an Synchronisation Supply Unit, the latest proposed values are $K_f\leq11000$ and $K_w\leq3$ to allow. For a SDH Equipment Clocks the latest version of part 5 at time of writing [5], they are K_f=3.0 and $K_w\leq0.875$. However, it has been shown (section 6.4.1) and others [14] that these values will not satisfy the wander masks. This is due to the problems with the SEC mask discussed in section 6.4.1 although this is not at all clear from [14] and was only confirmed by correspondence with the author reproduced in appendix A. For this reason new values for K_f and K_w were suggested in [14] to allow the SEC noise model to produce noise below the masks even though the masks do not reflect the asymptotic behaviour of the clock.

The gain values suggested [14] are as follows: