Premium Content


This content is available if you have a software support plan or current system warranty. To request access to this content, please submit a support request here.

Load Cell Noise Troubleshooting - EMA Systems with Shock6 Software (Aerotech Nservo Controller)

(Credit: S. Neitz, MTS SIE)  

Rev 1, 2021-08-18

Contents:

Purpose

Introduction

Troubleshooting

Noise Mitigation

    Verify shielding is continuous from the load cell all the way through the electronics board

    Verify the ground wire is soldered onto the back of the electronics board

    Verify the additional ground wires are connected the electronics board draining to the enclosure     Items that may change load cell noise, but not fix the root cause    

Appendix 

    References

    Creating a PVP test in Shock6

    Running the PVP test in Shock6    


 

Purpose: This document outlines common causes, troubleshooting and mitigation for excessive load cell signal noise that appears to be consistent with actuator displacement on EMA systems that run on Shock6 software (using the Aerotech Nservo controller). Other types of load cell noise may exist but are out of the scope of this document.

Introduction: EMA systems built at MTS and EMA systems previously built by Roehrig may be susceptible to load cell noise that appears as interference at certain locations in actuator displacement (not as random white noise) if the proper load cell cable shielding and system grounding measures are not properly implemented. As a result, excessive load cell noise can be present seemingly at random times, making the system unreliable for customers to analyze shock absorber properties.

Troubleshooting:

To determine if the contents of this document apply to the situation with your system, run a minimum of 2 runs of the same low velocity test without any specimen installed on the system. See the appendix for a procedure to create and running this a test in Shock6 if you are not familiar with the software. After obtaining data, view the force vs displacement plots of multiple test runs of the same test (overlay the plots in Shock6). Note if the load cell signal has any of the following characteristics (see Figure 1 and Figure 2 for bad and good examples of noise):

  • The force signal from two or more separate runs of the same test speed overlay on top of one another and appear almost identical, including any disturbances that do not appear like typical random noise
  • The force signal has fictitious spikes and dips that are large in magnitude for what is expected from the load cell full scale
  • The force spikes are not randomly distributed, but occur at specific displacement locations and are repeatable if the test is run at the same speed several times

Figure 1: An example of bad load cell noise that is typical of an improperly shielded load cell signal cable and/or electronics board. Note: data is obtained with no specimen installed and all forces recorded are fictitious

Chart

Description automatically generated

Figure 2: An example of typical load cell noise

Noise Mitigation:

Noise is occurs due to EMI/RF interference finding a weak point in the load cell signal cable shielding which can then disturb the sensitive load cell signal as it travels from the load cell to the signal conditioner (also called the “electronics board”) and finally to the controller that is reading the final amplified analog value. To mitigate the load cell noise, the shield for the load cell signal cable must be visually inspected at every junction and over every stretch of cabling.

Verify shielding is continuous from the load cell all the way through the electronics board

  1. Use Figure 3 as a guide of which connections to inspect. If the cabling differs from the guide, use the concept that the shield must be continuous at every junction along the load cell signal path through the electronics board (from connector A to connector G).

Figure 3: The cabling diagram with the load cell signal path highlighted, noting cable ends with letters A through G

  1. Load cell cable on the EMA frame (MTS P/N 100-305-899)
    1. Connector end A in Figure 4 (load cell end)
      1. Remove the connector back shell and visually verify that the shield is connected to the back shell.
        1. A picture containing text, person, hand, tool

Description automatically generated

Figure 4: the load cell connector showing the back shell disassembled and proper shielding termination

  1. Plug the connector end A into the load cell. Using a multimeter, perform a continuity check between the load cell body and the assembled connector’s back shell. The resistance should be near 0.
  1. Connector end B (Frame connection end): Remove the back shell connector and visually verify that the shield is terminated to the back shell of the connector.
    1. A picture containing text, person

Description automatically generated

Figure 5: The D-sub connector showing the back shell disassembled and proper shielding termination

  1. Full cable: Disconnect both cable ends. Check continuity between back shells at either end of the cable (between ends A and B). The Resistance should be near 0.
  1. Load cell cable from frame to electrical console (MTS P/N 05841250X)
    1. Connector end C (Frame end): Remove the back shell connector and visually verify that the shield is terminated to the back shell of the connector (reference Figure 5)
    2. Connector end D (Electrical console end): Remove the back shell connector and visually verify that the shield is terminated to the back shell of the connector (reference Figure 5)
    3. Full Cable: Disconnect the cable at both ends. Check continuity between back shells at either end of the cable (between ends C and D). The Resistance should be near 0.
  2. Load cell cable between console wall and electronics board (MTS P/N 05841240X)
    1. Cable end E (panel mount connector on side of electric console): Remove the back shell connector and visually verify that the shield is terminated to the back shell of the connector (reference Figure 5)
    2. Cable end F (connects to electronics board): Remove the back shell connector and visually verify that the shield is terminated to the back shell of the connector (reference Figure 5)
    3. Full Cable: Disconnect both cable ends. Check continuity between back shells at either end of the cable (between ends E and F). The Resistance should be near 0.
  3. Load cell cable between electronics board and Nservo (058412301)
    1. Cable end G (connected to the electronics board): Remove the backshell connector and visually verify that the shield is terminated to the backshell of the connector (reference Figure 5)

Verify the ground wire is soldered onto the back of the electronics board

  1. Remove the electronics board to view the back of the board and verify the additional grounding wire has been soldered in place (item #3 in Figure 3 and circled in Figure 6). NOTE: there are other wires soldered in place on the back of the electronics board that might look similar to the required wire but may not be the required wire. It is critical to verify that a specific wire circled in Figure 6 has been soldered in place on the board between the two specified locations. If it does not exist, solder a wire in place exactly as shown.

Figure 6: a picture of the back of the electronics board with the grounding jumper wire properly installed

Verify the additional ground wires are connected the electronics board draining to the enclosure

  1. After checking the connections specified above, rerun the load cell test (without a specimen installed) and inspect the magnitude and randomness of the noise. If the noise has been reduced and is random (no spikes dependent on displacement), the system can be left as is. If the noise still exists, there are a few more extra grounding steps that can be implemented (wires #1 and #2 in Figure 3).
    1. See MTS DOD 700-009-921 for instructions on implementing the additional grounding wires

 

Items that may change load cell noise, but not fix the root cause:

  • Cable separation between load cell cable and power cables (main power cable or motor power cables)
    1. While this is best practice to separate analog signal cables and power cables, proper shielding should effectively mitigate the disturbance from the power cables in the load cell cable such that no special precautions are needed in placement of the load cell cable. If you notice that the load cell cable placement close to or far away from power cables has a large impact on noise, the load cell cable shielding must be inspected as it is not doing its job.
  • Toroids on the load cell signal cable
    1. Toroids can be used to combat very high frequency noise spikes (MHz range noise) because they act to filter signals flowing through cables. A toroid is something implemented on the load cell signal by default on Shock6 Nservo EMA systems. Additional toroids will not be effective in mitigating largescale low frequency disturbances such as the example in Figure 1. Filtering the load cell signal heavily using toroids would be unacceptable as it would cause delays in the analog signal, impeding the system’s ability to accurately measure force.
  • Software filtering
    1. The noise present on the load cell signal has some very low frequency content and can’t be filtered out without removing valuable information about data that the end user needs to see (damping force that is not phase shifted or altered in amplitude). The root cause of the noise issue should be fixed (load cell shielding) instead of filtering the signal.

Appendix:

References:

  1. MTS DOD 700009921

Creating a PVP test in Shock6:

The following test requires the system to be fully operational with the exception of the load cell noise.

  1. Remove any specimen installed in the system and do not install any specimen back in the system.
  2. Either lightly snug the clevis’s onto the actuator shaft and load cell so they don’t bounce around on the threads (or alternatively unscrew them and set them aside).
  3. Open Shock6 software
    1. Create a new test
    2. Select PVP test type
    3. No warmup
    4. No gas/seal drag test profile
    5. Three test speeds (sine wave)
      1. 100mm offset, 75mm amplitude, 0.5hz (0.24 m/s)
      2. 100mm offset, 75mm amplitude, 0.75hz (0.35 m/s)
      3. 100mm offset, 75mm amplitude, 1 hz (0.47 m/s)
    6. For systems that are set in English units these are approximately
      1. 3.937” offset, 2.953” amplitude, 0.5hz (9.276 in/s)
      2. 3.937” offset, 2.953” amplitude, 0.75hz (13.915 in/s)
      3. 3.937” offset, 2.953” amplitude, 1 hz (18.553 in/s)
    7. Leave the PVP test options as the default values
    8. Save the test with a name that you will remember such as “ThreeSpeedLoadCellTest”

Running a PVP test in Shock6

  1. With no specimen installed on the machine, run the “ThreeSpeedLoadCellTest” PVP test created in the previous step.
    1. Click the “Test” button
    2. Select the “ThreeSpeedLoadCellTest” from the list of tests and click the “Start Test” button
    3. Click “Ok” if/when the file properties window pops up (it may not show up on some systems as it can be turned off in settings).
    4. Enter 200mm (a value larger than the offset + amplitude of the test) into the “Damper Travel” field. This will allow the “Ok” button to become clickable. Click “Ok” to continue and run the test.
    5. Once the test has completed, a popup will occur to ask you to name and save the file. Enter a name for the data file that you will remember and save the file.
  2. The most recently run test should automatically display on the chart. Once the test has finished running, view the force vs displacement plot in Shock6 and note if there are what appear to be trends in the force signal.
    1. Click the “force vs displacement” icon on the upper bar of Shock6
      1. Graphical user interface, application, Word

Description automatically generated
    2. Use the file menu to open other data files. Check the box in the “S” or “show” column to display the data in the plot if the data is not displaying
      1. Graphical user interface, text, application, email

Description automatically generated

Did you find it helpful? Yes No

Send feedback
Sorry we couldn't be helpful. Help us improve this article with your feedback.