4.3.2.5  Drift Correction 

When you train the PEC, it is assumed that the only error being corrected is the periodic 

error in the worm.  Because of the periodic nature of this error, the star will appear to 

drift back and forth in RA during the period of the worm's rotation, but will always end 

up in the same spot at the end of every rotation.  Theoretically then, at the end of a 

training run, the total amount of correction to the west should exactly equal the total 

amount of correction to the east.  If, however, there was systematic drift of the object 

caused by something other than periodic error (i.e. misalignment of the mount, 

differential refraction, etc.), then the corrections made during the training run would 

reflect this error also. 

This is not a problem if (1) the object being imaged is the same as (or very near) the 

object used to train the PEC and (2) the mount has not been moved since the training.  In 

that case, the same systematic errors that were present during the training run will also be 

present during playback, and both the periodic and systematic errors will have been 

corrected.  However, in many cases, you will want to use the same PEC data for widely 

scattered objects over several observing sessions when the telescope has likely been 

moved between the sessions.  In this case, the amount of systematic error will likely be 

different than during the training run, so just playing back the corrections made during 

training will actually introduce errors. 

Drift Correction solves this problem by effectively removing the systematic drift 

component from the recorded PEC data.   If everything were perfect and there were no 

systematic drift, corrections in both directions would cancel each other out, and this 

feature would not be needed.  If, however, they do not cancel out after a worm period, the 

existence of systematic drift is assumed.  

Selecting this menu item removes this non periodic component.  Gemini sums up the 

corrections (negative and positive) made in both directions, and spreads them evenly 

throughout the worm period. It does not, however, change the tracking rate itself, because 

to do so would lead to imprecise tracking in other regions of the sky.  

Because you may choose to train the PEC on the same object as the one that is to be 

imaged, Drift Correction of the PEC data is not done automatically after each training 

run; this helps minimize the guiding corrections necessary. You should only select the 

  Drift Correction   function if you want to use the PEC in a region of the sky other than 

the one where you trained it, or if you plan to use the PEC data across multiple sessions 

where the mount will be moved between sessions. 

After you have selected this function, 

the drift correction will remain active until a new training session overwrites the data. 

Using this function multiple times is not necessary, but also is not harmful. The amount 

of drift compensation is displayed, and should become zero after subsequent calls. 

4.3.2.6  Smooth Data 

In addition to periodic error and systematic drift, the corrections recorded during PEC 

training runs can also contain random errors caused by scintillations (  seeing  ) and over 

corrections.   The Smooth Data function helps eliminate these random influences by 

Gemini Users Manual 

4  

0