Run stops with Mass spec fault during acquisition: 16384

I am running a 7890A with a 5975C, using a Gerstel MPS PAL LHX2H-xt using Gertel Maestro, and MSD Chemstation G1701EA E.02.01.1177

Several times now, I have started a sequence successfully, collecting the first data file, but when it gets to the second injection it stops the run with: 

Fatal sequence error detected.
1101,"({SP00} :RUN:RDY ON;): Mass spec fault during acquisition: 16384"

I thought that 16384 was a communication error. It seems that the Mass Spec does not communicate with the GC according to the PING on the front of the MS, but I thought since they are both connected to Chemstation that they communicate through the msinsctl.exe

After the first one, I power cycled the MS and the computer. I was able to run an air and water check afterwards and the filament turned on fine. It started another run without turning on the filament. I requested to run the method from the 5975 soft key and it worked fine (although it overwrote the last data point from the previous batch!?!). 

Nothing major happened since the last successful batch. We likely had to vent it due to our too-frequent, scheduled power outages, but the vacuum is in the low 6s with really low atmospheric gases.

  • Mass Spec faults can be decoded in tune. Status, MS Error Codes, type the error, <enter>, and this is what shows for 16384

    The GC may have shut down the MS if GC detector, temperature, pressure, or flow setpoints have not be reached. The MS will remain in standby until the problem is fixed or the APG cable is disconnected.

    The simplest test is to remove the syringe and try a sequence of a few vials without it.  If it works, it's related to the injection. If it doesn't, it's probably related to a method setpoint.  Are you running Splitless or a Pulsed Splitless injection and the inlet is struggling to get to the initial pressure?

    Any beeps on the GC?

  • This was helpful. I had not used Status: MS Error Codes before. So that's a new tool for me.

    Even more helpful, is realizing that the problem did not originate with the MS!

    The method is a 10:1 split injection. There is elevated temperature and flow in the post-run, but this has worked in the past. It _could_ be part of the problem because combing through the logs, I have seen this:

    ACQ5975 Run started 22:12:50 07/11/2024
    7890 Front inlet pressure not ready Run Time 0.00 22:20:50 07/11/2024
    METHOD Acquired CO24193A_05_L3.d 22:23:53 07/11/2024
    7890 The run was stopped by the GC's Stop button or 22:28:22 07/11/2024
    Stop input.

    Although it says the Run Time is 0.00, this has occurred as it starts the post-run, as evidenced by being exactly 8 minutes after the injection, and the data not being written for another 3 minutes. That the GC's "stop button" isn't pressed for another 5 minutes makes me wonder if it thinks it is taking too long to cool down?

    I have tried many things without resolution. Best case, it will run 2-4 samples and then the GC will cause it to stop. I say GC because of the message above, but it is ultimately because the MS decides to go to standby.

    Thank you for your assistance,
    Matt

  • In Windows Explorer underneath your RUN_METHOD_Example.M subdirectory is a file called acqmeth.txt. Please upload it.

    10:1 split is very low if the column flow is typical GC/MS 1.2 ml/min.  The total flow should never be lower than 20 ml/min and the inlet pressure should never be <5psi for reasonable pressure/flow control.  These both can happen with too low total flow/too low split ratio, and too short/too wide a column at too low an oven temperature.

    On the GC edit parameters page, there are the GC Calculators.  I use the Vapor Volume and Pressure/Flow Calculator quite often.

  • The method is for headspace analysis. It was originally intended to have cryotrapping, but it has been modified to work without it, but I believe that the 10:1 was also used in the reference method.

    The split flow does drop to zero momentarily when it transitions to the post-run, so that could cause issues. We are leaning towards the GC network card though, since the errors do not always occur at that time.

    I came back in to the office to look up "The run was stopped by the GC's Stop button or Stop input." since this occurs at times when I would not expect the GC to be communicating anything. 

    AcqMeth.txt:

    https://drive.google.com/file/d/1qfDcDVCxLqR_dExKGCwStSzw-64xViE2/view?usp=sharing

  • (reply is hung up in moderation, presumably due to the attachment)

  • Recommendations.

    Increase the oven temperature equilibration time to 1.0 to 1.5 minutes or longer. It takes time for the 250° oven final temperature sheetmetal to cool all the way down and become stable for the next run.  If the temperature is making it to 30° C for 0.2, it might go ready but then might go not ready again as the temperature hasn't sufficiently settled down.

    30° C oven start temperature is very difficult to achieve in most laboratories. There's enough extra heat from the inlet and transferline that below 35° to even 37° C is tough at some places. If you really need to start that low you may have to use oven cryo. 

    The liner recommended for headspace only runs is the 1.5 mm i.d. one - 18740-80200 | Agilent (1 pk), 5183-4709 | Agilent (5 pk), 5183-4710 | Agilent (25 pk).  The sample does not need expansion volume so the low 140µL volume keeps the peaks a bit narrower than a larger one.

    Please set the MS transferline temperature to 230° C...to match the MS Ion Source.  That might aid source thermal stability.

    1.9328 ml/min column flow is higher than optimal for the ion source and higher linear velocity than optimal with the 20m×250µm×0.18µm column.  The orthogonal ion source works best between 0.8 to 1.4 ml/min of helium flow.  Lowering the column flow may improve your sample's peak shapes.

    Are you getting at least 10 scans across your earliest eluting peaks?  If not, you may need to change the A/D from 2^3 down to 2^2 ...that will reduce the averaging and may result in slightly increased visible noise, but better peak quantitation.

    That method is from Wed Jan 19, 2011 ?

Was this helpful?