7693 autosampler and 7890B GC/MS with MassHunter 10.2.489 acquisition software injection error

Hello, we are using a 7693 autosampler and 7890B GC/MS with MassHunter 10.2.489 acquisition software. When running a batch some vials do not collect as expected. We only get a baseline through the injection. If the same vial is re-injected the expected chromatography is captured. This is very random... sometimes only one or two out of a batch of 30.  Thank you for any input.

  • Are the vial caps pierced and no data collected or are they not pierced?

    An old problem that could cause this was sample file folder and file names that were too long or had strange characters in them.   It's best to keep the total file name less than 40 characters.  The default D:\GCMS\1\Data\yourfolder\datafilename.D\ xxxx and more files underneath, adds up quickly to more than 40 characters if you like to nest folders and sequence - my example is 39 characters.   For sure do not use <>?/{}[]!@#$%^&*()\ ~`:;"',...   and I don't even like using spaces.   The underline in between words works.  

    A sequence log is created, can you share it?

  • Hello Paul,

    Thank you for the reply. It took a bit to get back to you as we wanted a fresh review of vial caps after injection. In our last batch we had one vial out of 20 that collected base line with no peaks. This vial was pierced for that run. We later had that vial re-injected and the chromatography was as expected. This was a Std 6 in our run.

    The file names are all consistent throughout a batch. The same path is used and the file name only changes by an auto increment. No strange characters are in the file name and the length is less than 40 for all. Again this current miss inject was a routine standard so the name has been used many times.

    Please see attached for that runs sequence log.


  • Agilent 5185-5838 9mm PTFE
    All are pierced in the center.

  • Which run of Standard 6?  That was injected for data009.D, data013.D, and data038.D ?   I looked at a few runs, and, for example, all of the lines associated with data file data038.D copied/pasted next to all of the lines associated with data037.D -- are identical except for the file name differences. The sequence table believes that the runs all completed. There are no sequence errors.  <Your sequence log is now the longest excel spreadsheet I've ever made/opened at 503707 lines. I can color code things there and compare injections easily

    Next I would want to look at the zipped data files. Is there background?  In MassHunter Qualitative Analysis, do all of the baselines overlay with the other runs?  Do you use inserts in your sample vials? Could the syringe have not touched sample?  Do you do enough sample washes and pumps?  What about at least 2 seconds or more of viscosity delay?  Do you use a PTFE tipped plunger? What size syringe for how much injected?

    it's a 7010x system?   Is the instrument called "GC-MS-MS #4"   ?   please don't use special characters even in system names. It probably can't hurt there, but I worry.

  • Hello Paul,

    The first injection of standard 6 was the baseline only run. #13 was an immediate re-inject and #38 was the end of batch run.

    Thanks for the huge Excel file to look closely at all possible errors.

    The baselines of the mis-inject runs overlay well with those of heptane cleaning injections.

    Yes, we do use inserts. Perhaps even with the septa hole being in the center the insert is off center until the needle drives in and nudges it a bit. We have not seen this on any other system using the same combination of vials, caps, and inserts. 

    The syringe is 10µL and the inject volume is 1.0µL. L1 air gap = 0 L2 vol. = 0.5, and L2 air gap = 0.2

    Solvent A washes = 2 at 4µL, Solvent B = 2 at 4µL. Sample wash = 1 at 2µL. Sample pumps = 3

    The autosample is a 7693 and the tower is a G4513A.

    Thank you for looking into a fix to this odd error. Usually one random inject errro per batch of ~20 injections.


Was this helpful?