Recent Control & Safety Systems Discussions
Similar Posts
Process Historian is Crazy Slow- What tools to help Diagnose?
Melted terminals on M-Series Backplane - ever seen it?
#MES and #PervasiveSensing are seen as a critical part of supply chain management
DO Charms IO Error
DI CHARM: Bad Hardware Error
Share

Help diagnosing: mA error seen at the DeltaV AI CHARM

Was hoping to get some insight into diagnosis and troubleshooting of a signal error I am seeing. I have a (calibrated) Rosemount 1151 wired to an AI 4-20mA HART CHARM.

With the instrument at 4.0mA output, I am seeing a 0.5% (4.08mA) at the AI. With the instrument at 20mA, I see 102.2% (20.35mA) at the AI in DeltaV.

This behavior has been duplicated by removing the instrument from the loop, and sourcing 4/20 from TREX.

Here is a screenshot of the CHARM in Diagnostics with the Instrument at 4.0 mA:

SimHogan

19 Replies

  • Andre Dicaire
    Please log a CTS call with the GSC. There are no calibration options in the DeltaV Analog IO channels. These are calibrated at the factory. There are no moving parts (potentiometers) and such that can drift, and hence no calibration is needed.

    FYI. The current signal in the AI CHARM is dropped across two resistors in series. One is a high precision resistor that generates the internal voltage for the system and the second brings the total resistance to 250 ohms so that a 1-5 V reading can be measured at the terminal without breaking the circuit to use an Amp meter. the accuracy of the voltage reading is slightly less and it is intended for quick read of the signal.

    Have you tried another CHARM? Just curious if the same issue is present with a different AI CHARM. That might be an interesting data point for the GSC investigation.

    Andre Dicaire

  • SimHogan

    In reply to Andre Dicaire:

    Hello Andre
    Thank you for the prompt response. I will do both: Replace the CHARM and re-test, and I have just logged a CTS with Guardian.
    I will report back after re-testing the replacement CHARM.

    Best Regards,

    Sim

  • fkitzmann
    Sim,
    sometimes it has nothing to do with the AI CHARMs. When calibrating an instrument on the bench, the 4-20 mA signal output does not get adjusted for conditions in the field. The same is true when using the TREX. The communicator will put out 4 mA based on internal measurements.
    If replacing the AI CHARM does not produce the result you seek, I suggest you perform an analog signal trim (analog calibration) on the transmitter to get 4.00 mA and 20.00 mA at the AI (not as measured with a meter anywhere else).
  • Andre Dicaire

    In reply to fkitzmann:

    I'm curious. I grew up believing that current signals are constant over the entire loop. The current at the transmitter is the same as the current through the load resistors of the AI input. If a transmitter is truly putting out 4 mA, then one can expect 4 mA to arrive at the terminals of the AI CHARM. Since this error is proportional to the signal, it can not be an induced voltage which would produced a fixed bias. What sort of field installation issues cause a current signal to increase by 2% between the transmitter and the AI channel?

    Anyway, by connecting the TREX directly to the AI channel, the field installation issues are bypassed. That means either the TREX is improperly calibrated and is outputting slightly more than 4.00 mA/ 20.00 mA, or the AI CHARM is improperly factory calibrated. The product data sheet states a 0.1% error at 0-60 C. that would be 3.996 to 4.004 mA.

    Confucious: A man with one watch knows what time it is. A man with two watches is never sure... (did they even have watches in Confucious' time)

    From what I'm seeing, this looks like a span error of about 2%. If you do decide to adjust the transmitter trim, make sure the HART PV value matches the AI Block value. If the issue is with the CHARMs factory calibration, this issue would be present on other CHARMS. It's important to get to the root cause and to follow up with the GSC if you think it is an issue with the CHARM calibration.

    good luck.

    Andre Dicaire

  • Jonas Berge

    In reply to Andre Dicaire:

    Sim,

    Any signal conditioner like safety barrier or signal splitter between the transmitter and the CHARM?

    When you inject current from TREX, is that done in the field-end or directly on the CHARM terminals?
  • SimHogan

    In reply to fkitzmann:

    Thank you for the suggestion, this will not address my problem, but it will mask the error. The instrument is correct, responding with 4-20mA correctly at the corresponding process values.
    Your suggestion would be correct if there were a trim function in the AI CHARM.
  • SimHogan

    In reply to Andre Dicaire:

    This is what I was taught as well Andre, in fact, everything we are taught about current and voltage is based on it.
    By the way, this is not the first time I have seen this behavior since adopting CHARM. I've never been able to fully understand it yet.
  • SimHogan

    In reply to Jonas Berge:

    There is no splitter, or safety barrier. Our testing was done in the field, at the transmitter. Great suggestion to eliminate the field wiring and inject a signal directly at the CHARM.
    We will do this test to isolate the field wiring, as well as test after replacing the CHARM.
  • Tinh Phan

    In reply to SimHogan:

    You definitely have unique situation. I am sure you might have check these already but just want to make sure. Have you check your Ground loop for the HART device? Make sure you only have one single point of grounding. Are these low voltage wiring for the transmitter are ran separately from the high voltages? Are you using shielded twisted pair cable?
    Let us know what you find out the root cause. I am sure we all are curious of the root cause.
  • Jeffrey Mach

    In reply to Andre Dicaire:

    I'm going to second Andre here. While it is more likely something you see with aging control systems, since it is often too difficult to calibrate the analog input, you simply adjust the analog output of the field device so that the instrument and control system agree what "4" and "20" mA are. Even if those are not exactly those values to NIST standard, the important thing is that the device measurement is accurately understood by the control system. Thankfully, the HART protocol devices can easily allow you to adjust your output at LRV and URV so that the transmitter matches the receiver.

    Now, if I was to see a brand new Analog Input ask for a significant (more than a few %) alteration from "true" values, I would worry about a factory defect, so there is still a point to putting an ammeter in-line when calibrating to the analog input and recording the deviation from true. If the control system was getting up in years, but is otherwise fully functional, I would not worry so much if it drifted a few percent over its lifetime, so long as it still had a reasonable signal span and the rate of drift was minimal. Your tolerances for absolute error and drift will vary by application.
  • SimHogan

    In reply to Jeffrey Mach:

    I'm still pending further testing, so I have nothing new to add yet.

    In reply to Jeffrey
    Our system is not aging, as it is currently supported technology, the actual installation is less than 5 years old. In fact, I am taliking about a new Rosemount transmitter and new AI CHARM.
    I never saw this problem in 20 years on the legacy DCS (ABB MOD300 with GE Genius IO).

    It has been suggested "adjusting the instrument" to compensate for the error seen in the DCS. I cannot at all wrap my thinking around this (any more than I can understand why 4mA at the instrument is not seen as 4mA at the CHARM, other than possible injection of external noise or induced current. The error clearly is not with the current sourced from the instrument, in fact, it's LRV and URV are correct, validated and verified. Introducing bias at the instrument to make it 'look right' in DCS seems wrong and dangerous to me. Maybe someone can explain what I am missing.

    I am hoping to have more information from additional field testing soon, and will update.
    We are going to isolate the field wiring to eliminate that error, and replace the CHARM to see if the error persists.
    I appreciate the insight and discussion

    Best Regards to All!
  • LaurentB

    In reply to SimHogan:

    Before crying in all direction, let start with basic tests requests. Change Charm module and/or inject mA directly at charm terminal wiring.
    Then and only then after you can tell that you never saw that in 20years
    And if it’s still wrong, next year you can say that you already saw that
  • SimHogan

    In reply to LaurentB:

    Greetings LaurentB
    I am conducting the tests as prescribed, documenting as found, working with Guardian and within this group to effect a diagnosis.
    I hope to update the thread soon, Sadly, In the era of COVID19 simple tasks take longer than usual.

    Just to clarify -
    My reply was to Jeffery's statement "it is more likely something you see with aging control systems" -
    Outside if induced current from high voltage exposure (which I have seen), I had not seen this offset before.

    I am looking forward to diagnosing and understanding the condition.

    Thanks again
    Sim
  • Andre Dicaire

    In reply to SimHogan:

    SimHogan, Based on the data you provided, the error is proportional to the signal and not a fixed bias. A an induced voltage would present as a bias.

    When was signal generate (TREX?) calibrated to certified reference? I would at least use an Amp meter in series to see if both indicate the same current signal connected directly to the CHARM. Highly unlikely both meters would drift over time to the same error.

    CHARMS are all calibrated at the factory so a common calibration error is possible. Multiple CHARMS might read the same error. If CHARMS read different values with some within 0.1% spec and others out (i.e. 2%) then we know there is a factory calibration issue because all the CHARMS are not reading the same (+/- 0.1%).

    If you Amp Meter and Trex don't agree, you don't know which is off, but you do know you need to calibrate both of them.

    Do not use the 1-5V reading across the CHARM signal to common terminals as this will not be as accurate as the Current signal.

    Andre Dicaire

  • SimHogan

    In reply to Andre Dicaire:

    Hello Andre
    I agree - The error is proportional to signal, and not a fixed bias. We are scheduled to do more advanced loop testing on Tuesday (9/14).
    Looking forward to reporting findings soon. I will get more information what was used to calibrate (and validate) the TREX and the Transmitter.

    Thanks to all again for the input.
    Regards
    Sim

This is the official online community site of the Emerson Global Users Exchange, a forum for the free exchange of non-proprietary information among the global user community of all Emerson Automation Solution's products and services. Our goal is to improve the efficiency and use of automation systems and solutions employed at members’ facilities by sharing our knowledge, experiences, and application information.

User Groups | World Areas | Community Guidelines | Legal Information | Privacy Policy | Contact Community Manager

Website translation provided by Website Translator

© 2015-2022 Emerson Global Users Exchange. All rights reserved.