in DNDN's trials one unique thing is that there are almost no censored patients - also I wonder if your generated trials were there both "regular/classical" and "random/odd" shaped curves - but you thought of all that already ?
I did not test different amounts of censoring (although I can think of no conceivable reason why that would make a difference). I did test different types of curve shapes (albeit all roughly in agreement with the Log-Rank rules - e.g. no significantly crossing TRUTH curves).