Field Sensor Loop Reads High After Panel Rebuild
A technician rebuilt a control panel that reads a 4–20 mA pressure transmitter via a 250 Ω shunt into a 12-bit ADC (0–5 V, 0–4095 counts). Before the rebuild the loop read 8.00 mA at zero pressure and 20.00 mA at 10 bar. After the rebuild the ADC reports 2450 counts with the transmitter vented to atmosphere, and the reading drifts upward when a nearby VFD runs. The transmitter is two-wire, loop-powered from a 24 V supply, and the loop now shares a cable with a 480 V motor feeder for 6 m. The technician proposes recalibrating the zero in software to compensate.
1. Calculate the loop current implied by the 2450-count reading and explain whether software zeroing is a defensible fix.
Read the worked answer
Counts to volts: 2450/4095 × 5 V ≈ 2.99 V. Across 250 Ω that is 2.99/250 ≈ 11.96 mA, close to 12 mA mid-scale, which at zero pressure is wrong by roughly 8 mA. A correct 4 mA zero would give 1.0 V ≈ 819 counts. Software zeroing would hide a fault that is almost certainly an offset or wiring error, not transmitter drift: possible causes are a missing or reversed shunt, an extra parallel resistance, or a ground loop. Zeroing also destroys the ability to detect a genuine open loop later. The correct action is to break the loop, inject a known 4 mA and 20 mA with a calibrator, and confirm the count values at the ADC before touching any scaling.
2. Explain the VFD-correlated drift and give two corrective measures ranked by effectiveness.
Read the worked answer
The 6 m shared run couples VFD switching noise capacitively and inductively into the high-impedance signal pair; common-mode noise converts to differential error because the two conductors see unequal coupling. Most effective fix: separate the signal cable from the motor feeder, use shielded twisted pair with the shield grounded at one end only, and route at right angles where crossing is unavoidable. Second: add a passive RC or ferrite filter at the ADC input and confirm the transmitter's own filtering. Software averaging is a last resort because it does not remove the coupling mechanism and slows response. Verify improvement by running the VFD at several output frequencies while logging counts.