Rectifier Output Has Fallen and Potentials Have Moved
A 40 V / 20 A air-cooled rectifier feeds a 300 m horizontal anode bed in a drainage swale. At commissioning, panel meters read 32 V and 9.0 A with the taps as found. Nine months later, same taps and same AC supply voltage, the panel reads 40 V and 5.2 A. The drain-point instant-off potential has moved from -0.95 V to -0.88 V CSE, and the furthest test station TS-9 has moved from -0.86 V to -0.79 V. The swale has been dry for six weeks and the water table has dropped about 2 m. The panel ammeter was last calibrated two years ago and reads from a 75 mV / 10 A shunt.
1. Using the numbers given, decide whether this is a structure-side problem (coating/current demand) or a supply-side problem (rectifier/anode bed). What must be verified before you trust the conclusion?
Read the worked answer
Do the arithmetic first.
- Commissioning: R = 32 V / 9.0 A = 3.6 ohm
- Now: R = 40 V / 5.2 A = 7.7 ohm
Circuit resistance has roughly doubled while the driving voltage rose and the delivered current fell.
Direction of the argument: if the coating had failed, current demand would go up at a given voltage, not down. Falling current with rising voltage points to increased resistance in the loop - the anode bed, the header, the cables, or the rectifier itself.
Do not build on the panel meters yet. They were last calibrated two years ago. Verify:
- voltage with a calibrated DMM at the output terminals;
- current from the shunt with a calibrated DMM (75 mV / 10 A means 7.5 mV per amp - a 39 mV drop is 5.2 A);
- AC input voltage and tap position genuinely unchanged;
- shunt, fuses, and terminal connections clean and tight.
Only after the electrical readings are confirmed does the resistance conclusion hold. A single bad shunt connection can mimic this entire symptom.
2. How would you locate the increase in resistance within the circuit, and what field evidence supports each candidate?
Read the worked answer
Walk the circuit element by element, from the AC side outward.
- AC input / transformer: confirm input voltage and tap position unchanged; a dropped leg or a failing tap connection shows as abnormal AC volts or heat.
- Stack and shunt: check for a partially failed diode stack (output collapses under load) and for a hot or corroded shunt connection.
- Cables and header: look for damaged insulation, a corroded splice, or a loose header connection. A thermal camera or touch check after isolation is quick and revealing.
- Anode bed: the leading candidate here. Dry swale plus a 2 m water-table drop means the coke breeze has dried out and lost contact with the surrounding soil; bed-to-earth resistance rises, and gas blockage at the anodes can do the same.
Confirming the bed: measure the bed's resistance independently (bed-to-remote-earth, or a temporary rod array), and measure individual anode riser currents with a clamp where the header exposes them. Uneven division among anodes points to one failed or gas-blocked anode rather than a uniformly dried bed.
Why it matters: anodes that are simply dry can often be restored; anodes that are consumed or gas-blocked cannot. The remediation is different, so the diagnosis has to be specific.
3. The operator wants to raise the taps to 40 V. What is wrong with that as a first action, and how should the compliance picture be reported?
Read the worked answer
Why tap-up is the wrong first move:
- The unit is already at 40 V on a 40 V rating - there may be no headroom, and running it flat out risks overheating the transformer and stack.
- The bed is the limiting element. If the resistance is in the bed, more voltage produces little more current and simply dries the ground further around the anodes.
- It masks the fault. An unresolved resistance rise will keep climbing, and the next reading will be worse with no new information.
Fix or confirm the source first, then adjust output only after the bed can accept current.
Report the compliance picture honestly:
- The drain point at -0.88 V instant-off is more negative than -0.850 V, so it looks compliant - but drain-point compliance never demonstrates the whole line.
- TS-9 at -0.79 V instant-off fails the -0.850 V criterion, and the trend shows it degrading.
- Report both, with the trend, and evaluate the 100 mV polarization criterion using native potentials taken with the rectifier off.
- Flag the failing station, the uncalibrated meters, and the root cause as an open item with an owner and a date.
Do not let a compliant drain-point reading be presented as overall compliance.