Agilent 6890 valve switching problem

[nick576]

Hi,
In our lab we have started to experience a problem with a valve switching during a run, causing flow rate changes. Here is some more on the system and problem.

We use an Agilent 6890 to measure CH4 and CO2 in air samples using an FID, the GC is also fitted with a methanizer. In the GC there is a 6 port air actuated valve, which is used for backflushing a pre-column to vent using the AUX carrier gas inputs. The backflush starts during the run (valve switched ON) and is turned off at end of the run (Valve switched OFF) and the GC is ready for the next injection.

This method has worked extremely well for us. Recently we have seen two consecutive negative dips/peaks in the FID signal at the end of the run, taking about 2-3 min to settle again. The dips can be re-produced just switching the valve from the ON to OFF position. It does not occur switching in the other direction. The column flow rates (measured at FID outlet) drops to about 12 ml min-1 when the negative peak is seen, then gradually increases again. To me it suggests the valve is leaking, when it switches, but why only when it is switched OFF? Or am i completely on the wrong track?

And any ideas how this could be fixed would be greatly appreciated.
Thanks, Nick

GC parameters:

Air :400 ml min-1
Hydrogen: 45 ml min-1
Make up gas (N2): 20 ml min-1
Carrier (N2): 20 ml min
Column: packed Hayesep Q
Pre-column: short Hayesep Q
Oven: 50 C
Inlet: 100 C
Methanizer: 375 C
Detector: 300 C

[larkl]

I'd guess (and it is just a guess) that the valve rotor is scored from a particle or something. I'd replace just the rotor and see how it looks.