Investigation sequence
Preserve failure history
Record the original module number, replacement source, survival time and whether the new module failed instantly, during load or after days of operation.
Split field-side from drive-side
Before the board is blamed, record motor insulation, cable condition, output accessories and whether the fault remains with the output path isolated.
Compare all gate channels
Compare static resistance, bias rails, optocoupler output and gate waveform shape across homologous upper/lower U/V/W channels under a qualified method.
Inspect hidden small components
Prioritize zeners, gate resistors, optocouplers and small capacitors near the affected channel; leakage or aging may not look burnt.
Use low-energy proof first
Where a bench test is justified, verify driver permission and output balance before subjecting the module to full stored DC-link energy.
Close with a repeat-failure cause
The repair record should state the cause boundary: field-side, replacement module quality, driver channel, feedback/protection path or unresolved.
Stop conditions
- Another module fails during test
- Driver channel is asymmetric
- Negative gate bias is missing
- External output path is not cleared
- Protection must be bypassed to run
Linked records
The searcher is usually trying to decide whether the replacement module was bad, the motor/cable is still faulty, or a hidden gate-driver component is causing destructive switching. The first useful split is module authenticity and field-side load evidence versus six-channel driver symmetry and small-component leakage.
Repeat module failure usually means the original cause was not limited to the module; gate-drive, isolated supply, current detection, snubber/clamp, motor cable or load evidence must be reviewed.
The useful diagnostic question is whether the fast protection circuit detected a true output event or whether the gate driver itself failed to turn the module on hard enough. The page maps PWM isolation, push-pull gate current, negative bias and VCE detection into one repair route.
Maps the search-intent path for repeated Fuji G9/G11 module failures: PWM command, optocoupler/driver output, small capacitor leakage, zener/gate resistor condition, module input capacitance and motor/load evidence.
Turns the 616G3 discrete driver/protection explanation into a public diagnostic route: PWM isolation, push-pull amplification, negative gate bias, VCE rise detection and GF/OC feedback to the CPU.
A public safety-oriented map of why experienced repair benches reduce stored energy and use controlled load evidence before proving an output stage after destructive module repair.