geneally, the post-hoc test should help you to spot where the significant difference is.
I do not know why in your case it does not work.
What tool did you use? Did you do the right calculations?
May be that a look at your data could help us to see what's going on.