On our volatiles instruments we clean once every 12-18 months(or whenever both filaments burn out). We run over 600 drinking water THMs per quarter but the system stays clean. Our waste water instrument runs everything from monitoring wells to sludge samples, same source cleaning schedule, but the purge and trap requires frequent cleaning from all the crud.
Semivolatiles, like the Quechers Food methods, does require more often source cleaning since a lot of high boiling junk makes it through the column to the analyzer.
The worse I ever ran was Derivitized Melamine samples on pharma excipients, those killed the source on the 7000QQQ after about 40 samples. They were derivitized in the vials so the reagents were probably the source of the problem.
Another odd thing I have noticed on the 7000 is that when the source gets dirty, there will be a current across the filaments even when they are off. Go into manual tune and there is 25-30uA reading and when you turn on the filaments they read 35uA but barely glow, it is like a offset is placed on the circuit that monitors the filament current. Cleaned the source per Agilent's recommendation and the base current dropped to 1uA, which is still odd to me.