Noisy 4–20 mA Flow Signal on a VFD-Driven Pump
A municipal water plant uses a PLC to control a 100 hp VFD-driven pump. A magnetic flowmeter is located 300 ft from the PLC cabinet and sends a 4–20 mA signal. The signal cable is shielded twisted pair, with the shield grounded at both the flowmeter and the PLC cabinet. The analog input card is a single-ended 4–20 mA card powered from the same 24 Vdc supply that powers relay coils and the VFD interface. When the VFD runs, the flow reading fluctuates by ±5% and occasionally drops to 0 mA; when the VFD is off, the reading is stable. A handheld meter at the flowmeter terminals shows a steady 12.0 mA. The plant has already replaced the flowmeter and the analog card, with no improvement.
1. What is the most likely cause of the fluctuating and occasional 0 mA reading?
Read the worked answer
The handheld meter is a floating, battery-powered instrument, so it does not create a ground reference and reads the loop current correctly. The PLC input card is single-ended, meaning its measurement is referenced to the PLC cabinet ground. Grounding the shield at both ends creates a ground loop between the flowmeter location and the PLC cabinet. VFD operation raises the local ground potential through common-mode noise on the motor cable and the drive grounding system, so a voltage appears between the two ground points. That common-mode voltage can drive current through the shield and can exceed the input card common-mode range, causing offset errors, noise, and intermittent loss of signal. The occasional 0 mA reading may indicate input-card saturation or a temporary loss of loop compliance, not a failed flowmeter. Sharing the 24 Vdc supply with relay coils and the VFD interface adds switching noise and ground contamination. Replacing hardware without correcting the grounding and isolation scheme explains why the problem persists.
2. What corrective measures would you implement, and in what order?
Read the worked answer
First, break the ground loop by grounding the cable shield at one end only, normally at the PLC cabinet, and leave the flowmeter end isolated. Second, separate the analog signal from the VFD interface and relay-coil power; use a dedicated 24 Vdc supply or add an isolated DC/DC converter for the loop. Third, add a signal isolator or use an isolated analog input card if the common-mode voltage cannot be reduced below the card rating. Fourth, route the signal cable away from VFD motor cables and cross power cables at 90 degrees. Fifth, improve VFD grounding by terminating the motor cable shield with a 360-degree clamp at both the drive and motor, and add coil suppression at relay coils. Verify loop resistance and compliance voltage: 300 ft of typical 20 AWG cable is only a few ohms, but the input card burden plus any isolator burden must be supplied by the 24 V loop supply at 20 mA. Measure AC and DC common-mode voltage between the flowmeter ground and PLC ground before and after changes.
3. How would you prove the fix is effective under real operating conditions?
Read the worked answer
Use a loop calibrator to inject 4, 12, and 20 mA at the flowmeter end and verify the PLC reading at the input card, both with the VFD off and at several VFD speeds. Measure common-mode voltage between the flowmeter enclosure and the PLC cabinet ground with the VFD running; it should be reduced to a level well within the input card isolation rating. Trend the flow signal while the pump ramps and while other relays and VFD interface signals operate. Confirm that the reading no longer drops to 0 mA and that the fluctuation is within the loop accuracy tolerance. Also verify that the shield is grounded at only one point and that the signal return is not bonded to the VFD power ground at both ends. Document the final wiring, supply, and grounding arrangement so future maintenance does not restore the ground loop.