Polynoms are not matching Fourier Fits for AD CMi (DSCT) - although its a good data set ???

Bernhard,

I can comment only on the VStar data. It seems that you may have selected the entire light curve on which to perform the polynomial fits. This is not necessary, since you are interested only in determining the time of maximum (TOM). I suggest that you experiment with selecting (displaying) a more limited parts of the light curve, for example about half of the ascending and descending limbs and the peak, then in the View menu select Filter from Plot. Fit the polynomials to this filtered dataset.

I then fit polynomials starting with 4th order, and increasing (5th, 6th, etc) until it is clear the AIC reaches a minimum value (e.g, if AIC values were -281, -283, -284, -282, then -284 is the minimum value). VStar gives an RMS error. This is the root mean square of the residuals, which is not an error estimation of the actual TOM. In order to obtain the latter, you need to use Peranso or some other software.

I use Peranso. If you note the range of HJD values you used for your polynomial fit in VStar, note the polynomial order, then use those values in Peranso, you will find a TOM (it should be identical to that obtained in VStar) and an error estimation of the time.

Just a note on AIC. I use this and not BIC in VStar because limited experiments I've done indicate that, after fitting polynomials to several different parts of a light curve around the peak or trough, noting the AIC and BIC, then noting the TOM, the range of times is smaller using AIC than it is for BIC. In other words, in that experiment, AIC was less sensitive to the actual part of the light curve selected for polynomial fitting than was BIC.

Roy