DC Bus Overvoltage Trips on Regenerating Runs
A 2005 geared traction elevator (3,500 lb, 350 fpm, 480 V three-phase) uses a closed-loop VVVF drive with a DC bus, an IGBT chopper and a dynamic braking resistor. It now trips on 'DC bus overvoltage' during deceleration, but only when the car is descending with more than about 2,000 lb aboard, or ascending empty. The same floors in the other direction, and a half load in either direction, never trip it. Three months ago it happened once a week; this week it has happened two or three times a day, and twice the car stopped about 100 mm off the floor with a leveling fault.
With the car idle you read 495, 492 and 498 V phase-to-phase; during a run the lowest reading is 486 V. The braking resistor measures 9.6 ohms against a 10 ohm, 5 percent nameplate and looks heat-discolored but continuous. The main contactor and pre-charge circuit check out. The drive log also shows two 'chopper overcurrent' events in the same week.
1. Explain the energy flow that produces the trips, and why that particular load and direction combination triggers them.
Read the worked answer
In those two conditions the motor is not driving the load; the load is driving the motor. A loaded car going down and an empty car going up are the two cases where gravity does the work, the machine acts as a generator, and power flows back into the drive instead of out of it. The drive rectifies that energy onto the DC bus, the bus capacitors charge, and the chopper is supposed to switch the braking resistor across the bus to burn the energy as heat. When the chopper or its resistor circuit cannot absorb energy fast enough, bus voltage climbs until overvoltage protection trips.
That explains the dependence exactly. Half load, or the opposite direction, is motoring or nearly balanced, so little or no energy is regenerated and the bus never approaches the threshold. It also explains the progression: a degrading chopper or an intermittent resistor connection absorbs a smaller share of the energy each month, so the trips go from weekly to daily. And 495 to 498 V idle is at the top of the tolerance band, which reduces the headroom before the trip point is reached - the incoming voltage is not the cause, but it shrinks the margin.
2. The resistor measures within tolerance. What else do you test, and how?
Read the worked answer
A resistor can measure correctly cold and open when it is hot, so one reading proves very little. Test the whole circuit. Check the resistor's thermal switch or fuse, terminal torque, the cable from the chopper to the resistor, and the resistor's mounting and ventilation. Then test the chopper module itself through the drive diagnostics, or with the module isolated and a meter across the switching device, and compare against the 'chopper overcurrent' signature in the log.
With the drive off and the DC bus proved dead and discharged, inspect and measure the bus capacitor bank for bulging, leakage and lost capacitance - a bank that has aged has less ability to ride through switching and the bus ripple rises. Check the encoder coupling and feedback for a speed-loop overshoot on deceleration that commands more braking than the profile needs. Finish with a loaded test in the regenerating direction, recording DC bus voltage, stopping accuracy and the fault log. That test is the only proof the repair worked; a bench measurement is not.
3. Your supervisor suggests raising the overvoltage threshold and, if that fails, disconnecting the braking resistor because the car only uses it going down heavy. Respond.
Read the worked answer
Both suggestions remove a protection instead of repairing it. The energy from a loaded descending car does not disappear because the threshold moved; raising the trip point simply lets the bus run closer to the capacitor and IGBT ratings until something fails, and the failure mode is a destroyed drive or a car that cannot be stopped electrically. Disconnecting the resistor removes the only path that energy has. The first heavy down run sends the bus straight to the raised threshold, and in the meantime the chopper sees an open load and can fail shorted.
The correct path is to prove the chopper, the resistor circuit and the bus capacitors, replace what failed, and verify with a full-load regenerating test. If the repair cannot be made today, the car comes out of service, not the protection. Put the recommendation and your refusal in writing on the record, because 'we raised the limit and it stopped tripping' is exactly the history that hides a failure until it injures someone.