Noisy Flow Signal and a Flattened Air Line
You are commissioning a new magnetic flowmeter on a 4-in. chemical transfer line. Its 4-20 mA output runs to a PLC analog input through 18 AWG twisted-pair instrument cable with an overall foil shield. For 60 ft of a 180 ft total run, that cable is pulled through the same 2-in. conduit as a 480 V feeder to a variable-frequency drive. The shield is landed on the ground bus at the flowmeter junction box and again on the ground bus inside the PLC cabinet. With the drive off, the PLC indication is steady; with the drive running at 40 Hz and above, the indicated flow swings plus or minus 8% around an otherwise steady reading on the flowmeter's local display. Separately, the 1/2-in. copper instrument air tube feeding the control valve positioner has a 90-degree bend formed over roughly a 3/4-in. radius, and the tube is visibly flattened at the bend. The valve strokes noticeably slower than its twin on the next line.
1. Diagnose the signal swing. Identify each installation error and explain how the symptom pattern supports your diagnosis.
Read the worked answer
Two separate electrical errors, plus the routing.
Shield grounded at both ends. The two ground points are not at the same potential, so current flows along the shield. That current couples capacitively into the twisted pair and appears as signal error, and its magnitude follows whatever is driving the ground difference - here, the drive.
Signal cable sharing a raceway with a VFD feeder. Motor leads carry fast PWM edges with high dv/dt. A 60 ft parallel run couples that energy into adjacent conductors efficiently. Instrument (Class 2/3) conductors are not permitted to share a raceway with power conductors unless separated by a barrier or otherwise permitted; good practice keeps them in separate raceways regardless.
Why the symptoms fit: steady with the drive off, plus or minus 8% with the drive running above 40 Hz - that is electrical interference, not a process problem. A real flow disturbance would not care whether the drive was energized.
Also verify the flow tube's own grounding. An ungrounded or poorly grounded mag meter tube is an independent and very common source of 60 Hz noise, and it would mask whether the shield fix worked.
2. State the corrections you will make, in order, and how you will prove the fix.
Read the worked answer
Corrections:
- Separate the signal cable into its own conduit for the shared 60 ft, ideally for the whole run. Where it must run parallel to drive cable, keep at least 12 in. of spacing; where it must cross, cross at right angles.
- Single-point ground the shield. Lift it at the flowmeter junction box and insulate the end. Land it only at the PLC/panel end, with a short pigtail or, better, a 360-degree clamp to the cabinet ground bar.
- Verify the cabinet ground bar has continuity to the plant ground grid.
- On the drive side: use shielded VFD motor cable with the shield bonded 360 degrees at both the drive and the motor, and confirm the motor frame and drive enclosure are bonded.
- Bond the flow tube to the pipe with grounding rings or straps per the manufacturer.
Proof: re-test with the drive swept through 40-60 Hz under load and confirm the swing is inside the loop's acceptable noise band, not just quieter. Record the before/after values. A fix that only works at one drive speed has not been proven.
3. Evaluate the air tube bend. Is 3/4-in. radius acceptable for 1/2-in. copper, and what do you do about it?
Read the worked answer
Not acceptable.
The usual minimum for soft copper is about 3 times the OD - roughly 1.5 in. for 1/2-in. tube - and many manufacturers call for more. A 3/4-in. radius is well under that, and the visible flattening confirms the tube has yielded rather than bent.
Consequences:
- The flattened section reduces the flow area, so the positioner sees restricted air and the valve strokes slowly - which matches the observation against the twin valve on the next line.
- The crease is a stress riser and a fatigue crack waiting to happen on a vibrating line.
What to do:
- Isolate and vent the air supply, confirm zero pressure.
- Cut out the damaged section and re-bend new tube with a lever or spring bender to at least 1.5-2 in. radius.
- Do not try to straighten or re-bend the flattened piece, and do not add extra fittings just to dodge the bend if the run can be re-routed.
- Leak-test at operating pressure, then re-check valve stroke time against the twin to confirm the restriction is gone.