370XA alarms after every Auto Calibration and Validation but not Manual Calibration

Hi everyone, 

I have been experiencing 4 active alarm everyday when allowing Danalyzer 370XA GC to perform Automatic Calibration after 24 Hours. But when I untick auto Calibration and perform manual calibration the alarm does not appear for the case of Calibration and on Validation side alarm still appears even though when I perform Manual Validation.

The alarms are Stream 5 Validation Failure

  1. Stream 5 Validation Failure. 
  2. Stream 4 RF Deviation 
  3. Stream 4 Ideal RF Deviation
  4. Stream 4 RF Out Of Range

Please let us share the experience on resolving these alarms permanently on auto calibration and validation mode.

  • Hi Billy , please send me your contact details and I will ask tech support to contact you. My email is john.hay@emerson.com
    The RF deviation alarm will continue to be flagged every calibration until a RF is generated that is within the allowed deviation limits. Has the calibration gas been changed recently and the component data table matching the cylinder certificate ? Regards John
  • In reply to John Hay:

    Hi John,

    Correct the Calibration Gas has been changed and too bad there is no datasheet that shows exactly the content of this new installed Calibration Gas. I have instructed our team to look for data sheet so as we can compare.

    How about Validation? Is it also the function of RF?

    I have sent you my contact via your mail
  • In reply to Billy Hok:

    Billy ,
    In agreement with John Hay's reply .
    You confirmed you have change the calibration he cylinder . Before you attempt any further calibrations or validations you need to update the component table in the 370XA . If you do not have the certificate to hand , you may find the gas composition labelled on the cylinder . It should give you the mix . Compare this to the the existing values inthe GC companrnt table and update accordingly

    The GC may be working perfectly well . But calibration / validation failure occurs if the the calculated response factor generated during the calibration/ validation is outside the limits that you have set into the GC component table.

    The possible reason for the limit failure : If the new response factor is far removed from the original then it points to several potential errors. 1. The new calibration standard values have not been updated in the GC component table. 2. The calibration gas is not as the certificate suggests . That can be be checked by comparing the old calibration standard chromatogram picture with the new . If the mixes are very similar, old to new, then the graphic chromatogram should look similar. 3. The GC retention times have moved and some peaks may be only partially measured . This would be the first point to check . If the analyser is working normally all components should be holding the original retention time on the peaks .( and peak area should be reasonably matching )

    I hope this helps .
    John I just saw the message and jumped in with the comments.

    Best regards, Jim ( jim@mclelland.com)
  • Hi Jim,
    Remember me from the BASF/ Sinopec Nanjing job?  You helped me out greatly with the analyser hose.
    Hope all is well with you.  If you have a moment touch base.  We are working on another Emerson job here in Australia.
    Best Regards,
    Dave Evans |   | Lead Control Systems and Instrumentation  | Perth, Western Australia | Level 7, 197 St Georges Terrace, Perth, WA, 6000 | *: PO Box 7423, CLOISTERS SQUARE PO, WA, 6850 | (: +61 8 9347 4321 | ): +61 475 250 796 | *: Dave.Evans@albemarle.com | www.albemarle.com 
  • In reply to Jim McLelland:

    Cheers Jim, hope all is well.
  • In reply to Billy Hok:

    The Mon 2020 screenshots show that your current system is auto-calibrating and validating every 24 hrs, these are factory defaults, typically you should use validation to indicate when you actually needed to calibrate (i.e they would not be at the same intervals) , the calibration can hold for a number of days/weeks/months , validation simply checks that the calibration gas is reading correctly (to the certificate) within a tolerance set by the user (i.e 2 %) , please note the validation has no effect on the RF factors but it would have a validation alarm if it did not read the bottle correctly which seems the case if the calibration is failing, as pointed out by John and Dave there seems to be an issue with the calibration . I would suggest you save the diagnostics file, Mon2020 >> tools>>save diagnostic data, and send this to your local support team for review after you have checked your calibration bottle certificates.
  • In reply to Jim McLelland:

    Hi Jim,

    Onsite by bad luck, cylinder sticker was damaged and nothing is readable so far.

    We are now looking for information of the installed cylinder so as we can compare the values.

    I will give my feedback once I find the datasheet and correct installed information If they might be wrongly feed
  • In reply to jamie marsden:

    Hi Jamie,

    For now I have untick the validation and calibration tick. That leave the 370XA with only auto analysis from sample gas which so far has no problem. When I want to calibrate I do it Manually and the task finish successful without an alarm, now when I do validation this task Fail which now explain that the data does not match as you have said. As soon as I find the certificate I will change the data accordingly and will share the results.

    If the problem persist, I will take the diagnostic data as you have suggested.
  • In reply to Billy Hok:

    Billy ,
    All of the replies are sound and with the corrected cylinder composition values the 370XA should revert to running without alarms .

    That statement assumes that the basic chromatography is sound . With that in mind I can add some basic chromatography guidelines that are genric and not specific to the running of the Emerson 370XA model. If my notes considered too basic , you and your collegues can skip/ignore notes below.

    As John and Jamie highlighted the GC technology is very robust and therefore the analyser can run for weeks / months with little or no variation. REMEMBER IT IS AN ANALYSER SYSTEM , DO NOT OVERLOOK THE SAMPLE HANDLING SYSTEM. This stability of the system is reliant three critical functions
    1. SAMPLE SYSTEM must deliver representative sample in single phase ,( i will assume gas phase) . Even the smallest change in state will effect the sample being injected into the analyser . Dual phase at any level invalidates the measurement
    2. The GC SAMPLE INJECT, sample inject volume must remain constant .If there is any pressure difference between the sample gas injection and calibration gas injection you will see variations in result . The GC is typically built with atmospheric pressure balance to negate the effect .
    3. The GC CARRIER FLOW SYSTEM and GC OVEN TEMPERATURE must be stable , variations in either of these two will cause retention time shift of the measured components . Retention time change will create errors either reading low or high .

    4. CALIBRATION FREQUENCY : This can only be determined by initially running validation checks and confirming how stable the analyser runs without any correction factors being introduced. ( i.e. running calibration with wide response factor % windows ) . Any correction due to calibration updates hides any analyser drift effects either from sample inject , carrier flow or temperture changes or incorrect valve switching functions.
    Practical experience : Starting point once you have established a running GC . Run the GC on validation for three cycles . If stable return to process gas . 24 hours later run the validation again . If stable , hold of the validation of and repeat 4 days later , if still stable keep on process for say one week . Run validation two weeks later if still holding . YOU GET THE PICTURE YOU BUILD UP A HISTORY OF THE BASIC ANLYSER SYSTEM STABILITY . You may find it runns many months without needing calibration.
    That is why some opertors can run 6 months with no calibration and some operators have to calibrate on a more regular basis. THE NEED FOR CALIBRATION IS DRIVEN BY YOUR VALIDATION TOLERANCE ACCEPTANCE ( +/- CHANGES) AND IS ALL DEPENDENT ON THE APPLICATION REQUIREMENT.

    If you are tied to Gas metering for revenue accounting ,then 24 hour calibration may be requested . If your application is tied process control or process monitoring you can opt for wider validation windows and calibrate only when tolerances you set are exceeded.

    A bit lengthy , but hope it is useful to consider basics. ( Contact me by email if you need any more info )

    Good luck with the system