Negative dips in baseline all of a sudden

[EmpowersBane]

Im running a UPLC gradient run with buffered mobile in Line B, and 100% Acetonitrile in Line A. The last few runs have given me a dip in the baseline, which almost looks like the solvent/unretained front and its happening in the middle of the 40 minute injection. Im wondering is it something to do with the detector perhaps? I haven't given it an acid wash in a while. I use a PDA detector and again, this has only popped up the last few runs and I haven't put on a new column or loop etc.
any ideas?

[Rndirk]

Could you post such a chromatogram please?

[BostonFSE]

monitor the baseline with the flow off, do you see the spikes or flat line?
are the spikes very symmetrical? if so, that indicates electronic noise. power cycle all modules.

hard to tell with no chromatogram but if its noisy baseline, I use the trick below:

while monitoring the baseline, if you remove the flow cell with flow on and the baseline goes completely flat you have a dirty flow cell/lens. if it stays "noisy" you may need a new lamp.

[DR]

Could also be a function of the age of the MP.