lms_gain default setting not zero

[zhuber21]

Describe the bug

Right now, if you don't supply the lms_gain parameter to tracking_setup(), it defaults to None and then uses the code

if lms_gain is None:
    lms_gain = int(9 - np.ceil(np.log2(100/reset_rate_khz)))

to set it to four. Since we are supposed to be setting it to zero right now, having the default be something other than zero caused problems during our testing. As long as lms_gain is supposed to be set to zero, Shawn suggests that the default should be zero to avoid mistakes.

[zhuber21]

This is related to the issue Daniel brought up in https://github.com/simonsobs/smurf-issues/issues/36

[swh76]

This is related to JIRA https://jira.slac.stanford.edu/browse/ESCRYODET-825. Need to fix that first.