Restoring Water-Damaged Yaskawa F7 VFDs: Field Recovery Guide
Drives recovered from scrap, flood, or long-term outdoor storage arrive with a familiar signature: standing water, dried mud, debris in card cages, and corroded busbars. The good news is that modern IGBT-based drives such as the Yaskawa F7 (CIMR-F7) and the Saftronics G5 (a rebadged Yaskawa CIMR-G5 platform) are mechanically and electrically recoverable in the majority of contamination cases, provided you follow a disciplined procedure.
The single most important rule is that pure water is rarely the killer — it is what the water brings with it. Alkalines, acids, petroleum residues, road salts, and dissolved minerals all conduct current long after the visible moisture has evaporated, and those contaminants are what cause PCB leakage, copper migration, and electrolytic capacitor failure. Reverse this sequence by removing the contamination, drying thoroughly, and bringing the bus up gently, and a scrap-yard drive can be put back into service for the cost of your time.
Platform Reference: Yaskawa F7 and Saftronics G5
Identify the platform before powering anything up. The two drives in this article share a manufacturer lineage but are not the same architecture.
| Feature | Yaskawa F7 (CIMR-F7) | Saftronics G5 (CIMR-G5) |
|---|---|---|
| Generation | Late 1990s vector drive | Mid-1990s high-performance drive |
| Power range (400 V class) | 0.4 to 300 kW (0.5 to 400 HP) | 0.4 to 300 kW (0.5 to 400 HP) |
| Control modes | V/f, V/f w/PG, open-loop vector, flux vector | V/f, open-loop vector, flux vector |
| Digital operator | JVOP-160 (LCD) or JVOP-161 | JVOP-140 / JVOP-145 |
| Max input voltage parameter | 510 V (E1-01 upper limit) | 510 V (E1-01 upper limit) |
| Field-replaceable subassemblies | 7 PCBs typical (control, gate drive, snubber, power supply, charging, IGBT interface, operator) | 5 to 6 PCBs typical |
| Brand origin | Yaskawa Electric, Japan | Saftronics became Yaskawa subsidiary in 2006; pre-acquisition G5 units are rebadged Yaskawa G5 |
Both platforms cap the input voltage parameter at 510 V, which is why they work directly with the 500 V bus that a back-fed 220/440 V transformer produces. The G5 is mechanically simpler and easier to dismantle; the F7 has more onboard intelligence and a more capable fault-diagnostic display, which actually helps post-recovery troubleshooting.
Safety Precautions and Initial Assessment
Document the following before applying power or removing the cover:
- Visible contamination type: clear water, river mud, salt water, petroleum, sewage. Salt and petroleum contamination are recoverable but require distilled-water rinsing; sewage should be considered biohazard and discarded.
- PCB visible condition: white residue (oxidation, recoverable), green/blue corrosion (recoverable if localized), black carbon tracking (often fatal on that trace).
- Mechanical damage: bent busbars, cracked insulators, broken heat sink mounting tabs. Bent busbars on IGBT modules are usually repairable with careful re-bending; cracked insulators must be replaced with identical-form-factor ceramic or thermoplastic stand-offs.
- Connector corrosion: AMP/Molex connectors with green verdigris can usually be cleaned with a fiberglass pen and contact lubricant; sealed connectors with internal green corrosion are usually scrap.
Disassembly Procedure
Take photographs at every step. Document screw locations, cable routing, and torque values. The F7 is built from approximately seven subassemblies; the G5 from five to six. Lay boards out in the order removed and label each with a piece of masking tape and the corresponding torque value.
- Power down, lockout/tagout, wait five minutes, then measure DC bus voltage at the charger board test points.
- Remove the digital operator (JVOP-160 on the F7) by depressing the spring tabs on each side and pulling straight back.
- Remove the front cover screws (typically four M4 captive screws on the F7) and lift the cover vertically to clear the operator ribbon.
- Disconnect the operator ribbon cable, the cooling fan lead, and any option card connectors (PG-X2, SI-B, SI-P, SI-N).
- Remove the control board by releasing the plastic standoffs and lifting straight up — do not pry. The control board is the most expensive single PCB in the assembly.
- Remove the gate drive board, the snubber board, the power supply board, and the IGBT module interface board in that order.
- Remove the cooling fan assembly if water has reached it. Submerged fans should be dried and re-lubricated; motors with seized bearings are field-replaceable.
Cleaning Methods: From Dry Brush to Distilled-Water Wash
Match the cleaning method to the contamination type. Aggressive cleaning is wasted on a board that only needs a brush, and gentle cleaning is wasted on a board with five years of road salt.
Method 1 — Dry Brushing
For fully dried mud and dust with no salt or petroleum residue, a soft natural-bristle paint brush is sufficient. Brush in the direction of component leads to avoid prying loose through-hole parts. Follow with a vacuum cleaner fitted with an ESD-safe plastic nozzle — never a metal nozzle near live conductors or ungrounded boards.
Method 2 — Distilled-Water Rinse
For salt contamination or any residue that does not come off dry, rinse the board under warm distilled water with a soft brush. The rinse water resistivity should be above 1 MΩ·cm. Rinse for at least 5 minutes per board, then perform a second rinse with fresh distilled water. Tap water leaves mineral deposits that form leakage paths when the board is energized.
Method 3 — Dishwasher Cleaning
Field-proven on PCBs up to 600 mm square, the no-soap dishwasher cycle is effective for heavy contamination. Run three cycles with no detergent, but replace the tap-water fill on the final cycle with distilled water poured directly into the tub. This is the only way to guarantee a mineral-free final rinse. Air-dry the boards; do not use the heated dry cycle.
Method 4 — Pressure Washer (Heat Sink and Busbars Only)
A pressure washer held at 1 to 1.5 m (3 to 4 ft) is useful for the heat sink and busbar assemblies, but never point it at PCBs, connectors, or the cooling fan motor. Hold the board above a drain and direct the spray at an oblique angle so that the debris is sheared off rather than driven into the connector cavities.
Cleaning agents to avoid:
- Isopropyl alcohol (IPA) on certain plastic housings — it attacks ABS and polystyrene.
- Acetone or MEK — dissolves conformal coating and most plastic connectors.
- Commercial contact cleaner sprays that contain petroleum distillates — they leave a residue that attracts dust.
- Tap water as a final rinse — always finish with distilled water.
Drying: Avoiding Secondary Damage
Drying is where most recoveries fail. The goal is to remove moisture from under every IC, particularly the IC sockets, and from the inside of any inductive component. A board that looks dry on the surface can still have condensate under a PLCC socket.
| Method | Temperature | Duration | Best for |
|---|---|---|---|
| Ambient air dry | 20 to 30 °C | 48 to 72 h | Boards with low contamination |
| Hair dryer (low) | 40 to 60 °C | 30 to 60 min per board | Boards with trapped moisture under ICs |
| Convection oven | 60 to 70 °C (150 °F) | 6 h minimum | Boards with confirmed water ingress |
| Rice burial | 20 to 30 °C | 24 h | Boards in transport (pre-oven) |
| Desiccant cabinet | 20 to 30 °C, <20 % RH | 24 h | Final stage for any board |
Compressed air is dangerous as a primary drying tool: the rapid expansion cools the air below the local dew point and can drive condensation deeper into the assembly. The hose itself typically carries oil and moisture from the compressor. If compressed air must be used, fit a desiccant filter and a 0.3 µm particulate filter, and apply the air warm (above ambient) so that it picks up moisture rather than depositing it.
Capacitor Reforming: When and How
Aluminum electrolytic capacitors form a thin aluminum oxide dielectric layer on the anode foil. When a drive has been unpowered for more than six months, this layer slowly degrades and the capacitor leakage current rises. Applying full bus voltage across a degraded capacitor typically ruptures the vent and reduces the bus capacitance permanently.
The threshold is approximately six months of unpowered storage. A drive that has been in a scrap yard for two to three weeks does not need reforming. A drive pulled from a flood and stored in a workshop for a year does.
Reforming procedure with a current-limited DC supply:
- Set the supply to 0 V, current limit to 50 mA.
- Connect to the DC bus (with the drive's own rectifier disconnected if practical; otherwise across the bus capacitors directly).
- Raise the voltage slowly: 0 to 50 V over 5 minutes, then 50 to 200 V over 10 minutes, then 200 V to rated (560 V DC for a 400 V class drive) over 30 minutes.
- Hold at rated voltage for 1 hour while monitoring the supply current. The current should fall below 5 mA as the capacitors reform.
- Disconnect the supply, wait 5 minutes, and verify that the bus holds voltage (capacitors healthy if voltage decays slowly over several minutes).
Reassembly and Visual Inspection
- Reinstall the boards in reverse order of removal. Confirm that each connector is fully seated and that the locking tabs engage.
- Inspect the busbar bolts for proper torque — typically 4 to 6 N·m for the IGBT module busbars on a 5 HP drive. Loose busbars cause arcing and CPF24 faults on first power-up.
- Verify the cooling fan rotation is correct (intake-to-exhaust arrow on the housing).
- Reconnect the operator ribbon, option cards, and control wiring.
- Perform a final visual check for any stray solder, wire strands, or metallic debris before closing the cover.
Initial Power-Up and Parameter Configuration
Bring the drive up in stages. Do not connect a motor to the output terminals on the first power-up.
- Apply input power with the run command disabled (b1-02 = 0 for the F7, A1-02 = 0 on initial verification).
- Observe the operator display. A healthy F7 reads the firmware version, then displays the run-time screen (frequency reference 0.00 Hz, output frequency 0.00 Hz).
- Check the DC bus voltage in monitor U1-07. For a 460 V class input, the bus should read 620 to 680 V DC after the charge cycle completes.
- Verify input voltage parameter E1-01 is set to match the actual supply. The F7 accepts 155 to 510 V in 1 V steps.
For a 550 V motor fed from a 500 V bus, configure the F7 as follows:
| Parameter | Function | Recommended setting |
|---|---|---|
| A1-02 | Control method | 0 (V/f) for first test, 2 (open-loop vector) for production |
| b1-01 | Frequency reference source | 1 (analog input A1) or 0 (operator keypad) |
| b1-02 | Run command source | 1 (terminal) for production, 0 (keypad) for bench |
| C1-01 / C1-02 | Accel / decel time | 10 s / 10 s for first test; reduce to 3 to 5 s for production |
| E1-01 | Input voltage | 500 V (actual measured back-fed transformer output) |
| E1-04 | Max output frequency | 60 Hz |
| E1-05 | Max output voltage | 500 V (rated 550 V motor; for 440 V motor, set 440) |
| E1-06 | Base frequency | 60 Hz |
| E1-09 | Min output frequency | 0.5 Hz (V/f) or 0.0 Hz (vector) |
| E2-01 | Motor rated FLA | Per motor nameplate |
| L1-01 | Motor overload protection | 1 (general-purpose motor) |
| L3-04 | Stall prevention on decel | 1 (enabled, factory default) |
Run autotune (T1-01) before connecting the load. Rotational autotune (T1-01 = 0) requires the motor to be uncoupled. Stationary autotune (T1-01 = 1) is acceptable for first test if the mechanical load cannot be decoupled.
Back-Fed Transformer Configuration for 550 V Service
A back-fed single-phase transformer — 220 V or 240 V primary on the 440 V secondary — produces 500 to 520 V on the 220 V primary side. This is a legitimate technique for field operation of 550 V Canadian-spec machine tools when the available utility service is 240 V single phase. The 510 V upper limit of both the F7 and the G5 provides just enough headroom to avoid the oV (DC bus overvoltage) fault on light load.
Selection rules for the back-fed transformer:
- Three-phase VFD requires three-phase supply. A single-phase transformer feeding a three-phase VFD only works if the VFD is rated for single-phase input. The Yaskawa F7 400 V class with 5 HP / 3.7 kW rating accepts single-phase input at the 200 V class terminals but requires three-phase at the 400 V class. Verify the drive's single-phase input rating in the nameplate data and the hardware manual before attempting this configuration.
- Calculate the transformer kVA from the motor FLA. For a 5 HP, 550 V, 5.5 A motor:
kVA = sqrt(3) × V_LL × I_line / 1000 = 1.732 × 550 × 5.5 / 1000 = 5.24 kVAthree-phase apparent. A 3 kVA single-phase transformer will supply 13.6 A continuously, which is sufficient for a 3 HP 550 V motor at 4 A. A 5 HP motor at 5.5 A on a 3 kVA single-phase source will overheat the transformer within 30 minutes. - Size the transformer to 1.25 × motor kVA minimum to allow for VFD switching harmonics and inrush.
Verification and Functional Testing
- With the motor uncoupled, run the drive from 0 to 60 Hz and back in 5 Hz steps. Monitor output voltage and current on the operator display (U1-06 output voltage, U1-03 output current).
- Verify the output voltage at 60 Hz reads within 5 % of the E1-05 setting.
- Verify the output current at no load is less than 30 % of motor FLA. Higher no-load current suggests bearing failure or a partial short in the motor windings.
- Couple the load and repeat the frequency sweep. Watch for L3-04 stall prevention activity — if it engages, extend the decel time in C1-02.
- Run the loaded motor at rated speed for 30 minutes. Monitor the heatsink temperature (U4-08) — should stabilize below 70 °C on a 5 HP drive with normal duty.
- Verify input current does not exceed 80 % of the transformer secondary rating at full load.
Troubleshooting Matrix: Common Post-Recovery Faults
| Fault code | Meaning | Most likely post-recovery cause | Action |
|---|---|---|---|
| UV1 | DC bus undervoltage | Bus capacitor not charged; charge circuit damaged | Verify input voltage; check charge resistor and contactor |
| UV2 | Control power undervoltage | Control transformer tap loose | Reseat connector; verify +15 V rail on power supply board |
| UV3 | Inrush circuit fault | Contactor coil damaged by water | Replace inrush contactor; verify auxiliary contact closes |
| oC | Overcurrent | IGBT module damaged by bent busbar; insulation breakdown | Megger the motor leads phase-to-phase and phase-to-ground; check IGBT modules with curve tracer |
| oV | DC bus overvoltage | Input voltage setting E1-01 too high for actual supply | Reduce E1-01 to match measured input; check for regen on decel |
| GF | Ground fault | Residual moisture or contamination on output side | Re-dry; clean busbars with IPA-free contact cleaner; re-megger |
| SC | Output short circuit | IGBT module shorted from mechanical damage | Test each IGBT module with multimeter diode function; replace module |
| oH | Heatsink overheat | Cooling fan seized; heatsink fins clogged with debris | Replace fan; clear fins with compressed air; verify thermistor |
| oH1 | Heatsink thermistor fault | Thermistor lead corroded | Inspect thermistor leads; check resistance at 25 °C (typically 10 kΩ) |
| CPF00 / CPF01 | Control CPU fault | EEPROM corruption from moisture | Initialize to factory (A1-03 = 2220), re-enter parameters |
| CPF21 / CPF22 | Operator communication fault | Operator ribbon cable oxidized | Clean ribbon contacts with fiberglass pen; reseat connector |
| Err | EEPROM read/write error | EEPROM socket contamination | Remove EEPROM, clean socket, reseat; if persistent, replace EEPROM |
| oPE01 | Drive capacity setting error | Control board swapped from different capacity drive | Verify control board part number matches drive rating |
Field-Proven Caveats and Lessons Learned
A drive that has been sitting outdoors and then through a wash cycle can appear perfectly clean and still harbor faults. The most common post-recovery failure modes are:
- Corrosion on connector pins that was not visible until power-up. Always inspect the male pins under 10× magnification and clean with a fiberglass pen before reconnection.
- Capacitor vent rupture on first power-up. Reform the bus if the drive has been unpowered for more than six months.
- Wire creep: water can wick up inside stranded conductors and remain there for weeks. Cut back 25 mm of any wire that was submerged and re-strip.
- ESD damage during handling. Ground yourself before touching the control board; a 100 V static discharge can latch up the CPU or corrupt the EEPROM without leaving any visible damage.
- PCB delamination: if the board has flexed noticeably when wet, the inner layers may be separated. Tap the board gently and listen for a dull thud (delaminated) versus a sharp ring (intact).
Verify that all field-installed signal and motor cables meet the manufacturer's shield and grounding guidance. The shield should be terminated to ground at one end only to avoid circulating currents that inject common-mode noise into the drive's analog inputs and the protective earth path. The Xylem "Top 10 VFD Topics" technical bulletin documents this rule and the related ground-loop behavior in detail.
For long-term reliability, add a 5 % line reactor on the input when a recovered drive is fed from a back-fed transformer, since the transformer's higher source impedance distorts the input current waveform and stresses the input rectifier. The reactor is also effective at reducing the common-mode voltage that contributes to premature motor bearing failure.
When commissioning a recovered drive on a 550 V motor for the first time, set the maximum output voltage (E1-05) to the motor nameplate rating — never to the bus voltage. The F7 will otherwise over-flux the motor at base frequency and trip oC on the first load step.
Can a Yaskawa F7 VFD be recovered after being submerged in water and mud?
Yes, in the majority of cases. Remove all PCBs, clean with distilled water (or a no-soap dishwasher cycle with distilled final rinse for heavy contamination), dry for 24 to 72 hours at 60 to 70 °C, inspect for corrosion, then reassemble and reform the DC bus capacitors if the drive has been unpowered for more than six months. Most field-recovered F7 units will pass the first power-up sequence after this procedure.
What parameter sets the maximum input voltage on the Yaskawa F7?
E1-01 sets the input voltage. The F7 accepts 155 V to 510 V in 1 V increments. For a 550 V Canadian-spec motor fed from a back-fed single-phase transformer, set E1-01 to the measured transformer output (typically 500 V) and E1-05 (max output voltage) to the motor nameplate rating.
How long can a VFD sit unpowered before the capacitors need reforming?
Approximately six months. Below that threshold, the aluminum oxide layer on the anode foil is generally intact and a direct power-up is acceptable. Above six months, reform the bus using a current-limited DC supply at 50 mA, ramping from 0 V to rated bus voltage (560 V DC for a 400 V class drive) over 45 to 60 minutes.
Is a back-fed 220 to 440 V transformer safe for feeding a 460 V VFD?
For shop and test-bench service, yes, provided the VFD's input voltage parameter (E1-01) is set to the measured secondary voltage and the maximum input is below the 510 V drive limit. For production or commercial service, the configuration must meet local electrical code — for example, NEC Article 450 (USA) or CEC Rule 26-700 (Canada). Add a 5 % line reactor to limit input current harmonics and protect the bridge rectifier.
Which is more recoverable, the Yaskawa F7 or the Saftronics G5?
The Saftronics G5 (CIMR-G5) is mechanically simpler with fewer subassemblies, making it marginally easier to clean. The F7 has more onboard intelligence, including a more capable fault-diagnostic display (CPF00 through CPF24), which actually helps the post-recovery troubleshooting. Both share the same 510 V input ceiling and the same general recovery procedure.