it is often a non-optimal power/speed combination. The only thing you can do is test what suits your specific machine, but the test objects should not be too small. Circles or squares of min. 20x20mm give a reasonable picture of the effect from different speeds/power settings.
In a quick search here on the forum, I came across this comment on another topic from @jkwilborn which, although it doesn’t specifically address the issue, I believe provides good guidance.
