We have a thermo finnigan survey LCQ instrument. When I used it, I got noisy signal problem. This only happens to the ms data, not for the PDA results.

Here are the conditions I used:
Analyte: Caffeine at a concentration of 10 mg/L (Injection volume: 10 µL)
Mobile phase: A: 0.1% formic acid in H2O (40%); B: 0.1% formic acid in ACN (60%). [Isocratic]
Column: C18
Flow rate: 0.5 mL/min
Sheath gas flow rate: 90 arb
Aux gas flow rate: 20 arb
Capillary voltage: 3.5 kV
Capillary temp: 350 ºC
Cone voltage: 39 V

I also attached an example of a chromatogram below: Image

What could be the most possible reason? Actually this problem also happens to other analytes even when I try to optimize the ionization parameter. Any suggestion will be greatly appreciated!