Direction in a PWM drive is not a wiring property. The rectifier front end collapses the incoming supply into a DC bus, and every characteristic of the motor voltage after that point — frequency, amplitude, and phase sequence — is synthesized by the IGBT bridge firing order. Reversing a motor on a Mitsubishi FR-E520 means telling the gate logic to fire U-W-V instead of U-V-W. That is a command, not a cable.
Where Direction Actually Lives
Swap two conductors on the drive's input terminals and nothing happens to shaft direction. The diode bridge rectifies whatever phase relationship arrives; the bus capacitor holds DC. Phase sequence information is destroyed at the rectifier and re-created at the inverter bridge.
The number that matters is the commutation order, and the only thing that changes it is the run command state at the control terminals. On the FR-E520 those are STF (start forward) and STR (start reverse), referenced to the sink-logic common SD. Closing STF–SD runs forward; closing STR–SD runs reverse. The power terminals R/L1, S/L2, T/L3, U, V, W are untouched.
Three Ways to Reverse, and What Each Costs
| Approach | What it does | Cost / risk |
|---|---|---|
Control terminal STR (two-wire) |
Drive ramps to zero, inverts gate sequence, ramps up in reverse | One dry contact and one parameter change. No power-side hardware. |
| Reversing contactors on the drive output (U/V/W) | Mechanically swaps two motor phases | Breaking magnetizing current mid-flight; back-EMF strikes the contacts and the IGBT bridge. Overcurrent trips in the E.OC group, contact welding, possible module damage. |
| Contactors on the drive input (R/S/T) | Nothing — rectifier discards phase sequence | No reversal at all, plus repeated inrush charging of the bus capacitors and precharge-resistor stress. |
| Keypad (PU) or serial command | Direction commanded from the operation panel or communications | Works, but operation mode must exclude external terminal control; not usable for a field-mounted selector switch. |
Use the control terminal. Output-side switching is the one option that can destroy hardware: opening a contactor while the bridge is modulating into an inductive load produces an arc and a voltage transient the output devices are not sized to absorb. If a maintenance disconnect on the output is mandated by procedure, it must be interlocked so the drive is stopped and the output de-energized before the contacts move.
Terminal Wiring and Logic Common
Each input sources a small internal current; the field device only has to close a contact to the common. Two points cause most of the field failures:
- The common is
SDfor sink logic. Wiring a selector switch to a separate 24 V supply's negative rail without tying it correctly to the drive common leaves the input floating and the run command never registers. - Closing
STFandSTRat the same time is a stop command on FR-series drives, not a fault. A cheap two-position switch with make-before-break contacts will produce an unexplained momentary stop at every transition. Use a break-before-make selector, or a center-off three-position switch.
Keep the control wiring in shielded cable routed away from motor leads, with the shield landed at the drive end only. Induced noise on a run input from a parallel-run motor cable produces intermittent direction changes that look like a logic fault and are not.
Parameter Setup for External Two-Wire Control
- Isolate the supply, confirm the bus charge lamp is extinguished, and land the direction switch: common leg to
SD, forward contact toSTF, reverse contact toSTR. - Power up and set the operation mode selection parameter (
Pr.79) so the drive accepts the start command from the external terminals rather than the operation panel. Verify the mode indicator on the display reflects external operation. - Check the reverse rotation prevention parameter (
Pr.78). If it is set to inhibit reverse,STRis accepted electrically but produces no rotation — the most common reason a correctly wired reverse input appears dead. - Set acceleration and deceleration times (
Pr.7,Pr.8) to values the load inertia can actually follow, then bump the motor in each direction at low frequency before committing to full speed.
Ramp Physics During a Direction Change
A reversal is two events, not one: a controlled deceleration to zero, then an acceleration in the opposite sequence. During the deceleration half the machine inertia drives the motor above synchronous speed and it acts as a generator. That energy lands on the DC bus. This is heat and stored charge, not logic — if the deceleration time is shorter than the bus and any braking resistor can absorb, the bus voltage climbs past the trip threshold and the drive faults in the E.OV group. Lengthen the deceleration time, or add dynamic braking sized to the regenerated energy.
The acceleration half loads the motor from zero against whatever residual rotation remains. Stall prevention will hold output current at the configured level and stretch the ramp; if the mechanical load cannot tolerate the resulting torque reversal, the fix is on the ramp, not the drive.
| Quantity | Limit or target | Where to read it |
|---|---|---|
| Input phase sequence | No effect on shaft direction | Rectifier / DC bus topology |
| Run input contact duty | Dry contact, drive-sourced logic current | Control terminal specification table, instruction manual |
STF + STR both closed |
Stop command | Input terminal function description |
| DC bus voltage on reversal | Below overvoltage trip level | Drive monitor screen and fault history |
| Output current on reversal | Below stall prevention level | Output current monitor |
| Deceleration time |
Pr.8, long enough to avoid E.OV |
Parameter list |
Verification and Recurring Pitfalls
Verify with the motor uncoupled or the load in a safe state. Jog forward at low frequency and confirm shaft direction against the machine's defined positive direction. Command reverse and watch the operation panel: the drive should show the deceleration, pass through zero, and display the reverse indication as output frequency rises again. Cycle direction ten times while watching the output current and DC bus monitors, then clear and re-read the fault history — an empty history after ten cycles is the acceptance criterion, not a single successful reversal.
Recurring failures on this class of drive: a make-before-break selector producing phantom stops; reverse inhibited in parameters while the installer chases a wiring fault; an unshielded run-input cable bundled with motor leads; deceleration time left at a factory default that a high-inertia fan or centrifuge cannot follow; and a drive pulled from stores with remapped multi-function inputs. Check those five before opening the enclosure a second time.
Escalate when the fault repeats with the motor uncoupled, when an overcurrent code appears on a bare-shaft reversal, or when the drive accepts STR in external mode with reverse prevention confirmed off and still produces no output. Those point at the output bridge or the control board rather than the configuration, and belong with Mitsubishi Electric authorized service with the fault history and parameter set recorded.
FAQ
Why does swapping two phases on the inverter input not reverse the motor?
The input feeds a diode rectifier that converts the supply to DC on the bus capacitors, which destroys any phase-sequence information. The output phase sequence is generated independently by the IGBT gating, so only a run command at STR changes shaft direction.
Why does the FR-E520 ignore the STR input even though the wiring is correct?
Two causes dominate: the operation mode parameter (Pr.79) is not set for external terminal control, or the reverse rotation prevention parameter (Pr.78) is inhibiting reverse. Confirm the mode indicator on the display and read both parameters before troubleshooting the field wiring.
Why does the drive trip during a forward-to-reverse change but run fine in one direction?
The deceleration half of the reversal pushes the machine inertia back through the motor as generated energy onto the DC bus, and the bus voltage crosses the overvoltage trip level. Lengthen the deceleration time in Pr.8 or add a braking resistor sized to the regenerated energy.