weight must be constant within ID in panel data

[Alex Carr]

I first use logit to predict the likelihood of observations being treated. Then I calculated Inverse Probability Weighting (ipw) with 1/ipw for treated and 1/(ipw) for controled.

Then I try to run fixed-effect with IPW:

xtreg y t x1 x2 x3 [ps=ipw], fe

i get the error message "

weight must be constant within id

however, if I run OLS instead, with

 reg y t x1 x2 x3 [ps=ipw]

there is no error message

does anyone have any idea why this might be?