Calculating Variance Explained by Effect: High Intercept Sum of Squares Problem

Hi :)

Context: I'm using G*Power to conduct a post hoc power analysis. The study design is a 2x2 Two-Way ANOVA. ANOVA was done in SPSS.

Within G*Power, I'm using Protocol of power analyses -> F tests -> ANOVA: Fixed effects, special, main effects and interactions. Then to get the effect size F, I'm using the "Determine" button and filling in "Variance explained by special effect" and "Error variance." I'm pretty confident that "Error variance" is the square root of the MSE in the ANOVA table.

When calculating the variance explained by the special effect: I've done a lot of google-ing and reading, and the best estimate seems to be the SS of the factor in question divided by the total SS. (Please correct me if I'm wrong!) However, in the analysis of my raw data (which is normally distributed), about 90% of the SS is found in the intercept (e.g. total SS = 100, intercept SS = 90). This is cutting down my effect size a ton. I'd understand that if it was a really high error SS, but the error SS is low.

I ran a transformation of the data where I simply subtracted the mean data point such that the intercept now = zero. This removes the intercept SS from the total, such that the total SS is much lower and the effect size now seems more reasonable to me in comparison to the raw data.

How should I treat the SS of the intercept? Is it reasonable to do that transformation to get the variance explained by the effect? If not, what does it "mean" that the SS of the intercept is almost all of the total SS?

I did a search of the forum with keywords "intercept sum of squares" or "intercept power" and didn't find anything that addressed this issue.