That is such a large jump that this warrants further investigation.
Did you change anything about your setup or testing? i.e. did you use the same power meter and type of test?
If you did make changes, post them here and we can help you figure out a good path forward.