A servo drive that faulted on overvoltage every time a vertical axis descended. The axis was a Z-axis lifting 50 kg. The servo was 750W. The drive had no braking resistor. When the axis descended, the motor became a generator — the kinetic energy charged the DC bus. The bus voltage climbed from 310 V to 400 V, triggering the overvoltage fault. This is about braking resistor sizing.

The regenerative energy

When a vertical axis descends, gravity accelerates the load. The motor must hold back — it operates as a generator. The energy goes back into the DC bus. The DC bus voltage rises. If no resistor bleeds the energy, the bus overvoltages and faults.

The energy per descent:

E = m · g · h · η

Where m is the load mass (kg), g is 9.81, h is the descent height (m), and η is the efficiency (0.8 for a ball screw). For 50 kg descending 300 mm (0.3 m): E = 50 × 9.81 × 0.3 × 0.8 = 118 J per cycle. At 10 cycles per minute: power = 118 × 10 / 60 = 19.7 W. That’s small. But the peak power during deceleration is much higher.

The braking resistor sizing

The resistor must dissipate the peak braking power. The peak power during a controlled stop:

P_peak = m · g · v

Where v is the descent speed (m/s). For 50 kg at 0.2 m/s descent: P_peak = 50 × 9.81 × 0.2 = 98 W. The resistor must handle 98 W peak. But the resistor also must handle the average power (19.7 W continuous). I spec a resistor with 2x peak rating: 200 W. The resistor resistance is determined by the DC bus voltage:

R = V_bus² / P_peak

For a 230 V AC drive, the DC bus is 310 V. At 400 V (regenerative threshold): R = 400² / 98 = 1633 Ω. I use 1000 Ω (a common value) which gives 160 W peak. That’s enough for 98 W peak.

What I changed

1. Added a 100 Ω, 100 W braking resistor. Wait — I need to recalculate. The drive’s internal braking transistor switches at 380 V. The resistor must be sized so the power at 380 V doesn’t exceed the resistor rating. P = 380² / R. For R = 100 Ω: P = 1444 W. That’s too much — the resistor burns. For R = 1000 Ω: P = 144 W. That’s fine for a 200 W resistor. I installed a 1000 Ω, 200 W resistor. The overvoltage fault disappeared.

2. Added a dynamic braking resistor for multi-axis. For a machine with 3 vertical axes, I use a shared regen resistor (or a regenerative power supply). The energy from one axis braking feeds another axis accelerating. The shared resistor is smaller because the peaks don’t coincide.

3. Added mechanical holding. For vertical axes, I always add a mechanical brake (spring-engaged, pneumatically released). The brake holds the axis when power is off. The servo doesn’t have to hold the load at rest. The braking resistor only handles the dynamic braking during descent — not the static holding. This reduces the resistor size by 50%.

The resistor mounting

Braking resistors get hot (200W continuous = 200W heat). I mount them on a metal panel (not plastic). I keep them away from sensitive electronics. I add a thermal switch on the resistor (set at 150°C) that faults the drive if the resistor overheats. A burned-out resistor causes an uncontrolled drop of the vertical axis.

The resistor I size: 2x peak braking power, resistance chosen so P = V²/R stays within rating. The overvoltage fault wasn’t the drive — it was a vertical axis with no regen resistor. Add a 1000Ω/200W resistor and a mechanical holding brake. The brake handles static; the resistor handles dynamic.