Please can you help me. We have been having repeatability problems on the HP7890B for a long time and can't get to the bottom of it. Analysing for VOC. I have run lots of RSD measurements (10 injections). A common thing we are seeing is excellent RSD(%) and then every so often we get an increase in area of a peak which can throw the RSD(%) our quite a bit as shown in the table below. Sometimes the result is higher and we get RSD(%) of about 7%. This one outlying result can throw a whole analysis out and there doesn't seem to be any pattern to when this one outlying result will occur. Another problem being when we look at bigger amounts in standards the result if on a bad injection can be in the magnitude of 100s of ppm difference to what it should be. the table below shows that both the Benzene and Toluene peaks increase with the same magnitude.
Autosampler Turret - G4513A
Syringe - Agilent Gold Tapered (5181
Septa 5183-4757 Bleed/Temp optimised, non stick
liner - 5183-4647 (deactivated glass wool, single tapered 4mm)
Solvent pre washes, sample washes, sample postwash all x4
Temp - 250oC
Total flow 36.3ml/min
Split flow ratio 2:1
Split flow rate 22.2 ml/min
Temp - 250oC
Air Flow - 350ml/min
H2 flow - 35ml/min
Makeup N2 flow - 12ml/min
Column flow - 11.1ml/min
Pressure - 6.5PSI
Velocity - 100.6ml/min
Initial oven temp - 38oC
Constant type - Pressure
Runtime - 17.3 mins
Post run temp - 50oC
38oC for 9 minutes
ramp 40oC/min to 90oC hold time 0 minutes
ramp 15oC/min to 180 and hold for 1 minute
Things we have tried
Maintenance - regularly changing Syringe, liner,septa, o-ring
Checked the Autosampler turrets and run on front and back FID - same results
Run the same standards on another GC in the lab a HP6890 with same method and HP6890 RSD all ok, HP7890 still problem. Used the column from te HP6890 in the HP7890 and same problem occuring.
Had Agilent visit and check the machine and they see nothing wrong with the instrument.
Looked H2 generator, and fitted a regulator so we would get a definite steady flow of Hydrogen.
Tried Splitless and split analysis
Looked at Viscosity delays and injection techniques all look fine as no change in results.
Changed the method parameters to constant flow rather than constan pressure, still the same.
Increased method time to see if peaks elute later on. there was no peaks eluting later on.
If anyone has had similar problems or have an suggestions of how to solve this problem then I would be forever in somebodies debt as we have has this probm for around 3 years and never found the solution and because of this we have a 20 year old HP6890 running everything and a newish HP7890 sat there like an expensive brick of plastic doing nothing much.