Thanks Crossrh, I ran both sweeps 2x per measurement. I did notice this before, where the results from a single and multiple sweep where very different, and the latter being much worse.
How would I set "synchronously averaged", sounds intresting.
I'll try single sweeps again tomorrow.