APW9/APW9+ Power Supply Repair & Diagnostics Guide: Complete S17/T17 PSU Manual

Introduction

The Bitmain APW9 and APW9+ are the power supply units that powered the entire S17/T17 generation of Bitcoin miners — arguably the most temperamental and repair-intensive generation Bitmain ever produced. Every Antminer S17, S17+, S17 Pro, S17e, T17, T17+, and T17e that shipped from the factory between 2019 and 2020 was paired with an APW9 or its improved successor, the APW9+. These PSUs sat at the intersection of Bitmain’s transition from the S9 era to the S19 era, and they inherited some of the growing pains of that transition. If you are running S17-series or T17-series hardware, understanding the APW9/APW9+ is not optional — it is the difference between a miner that runs and one that sits on a shelf waiting for a diagnosis.

The APW9 is a high-power, single-output switching power supply that converts 200–240V AC from the wall into regulated 14.5V DC at up to ~248A, delivering approximately 3,600W of continuous power. The APW9+ is the enhanced revision with improved efficiency and better thermal management, but the core architecture remains the same. Inside the metal enclosure you will find a PFC front end, an LLC resonant converter, synchronous rectification, a microcontroller managing communication with the miner’s control board, and a suite of protection circuits that monitor for every conceivable fault. It is a serious piece of power electronics — and it demands serious respect, both for what it delivers and for the lethal voltages present inside.

The S17/T17 generation earned a reputation in the mining community for reliability problems, and the PSU is frequently part of that story. Connector overheating, PFC failures, blown input fuses, and fan bearing wear are issues we see on the D-Central repair bench daily. This guide is your complete reference for diagnosing, repairing, and maintaining the APW9 and APW9+. We cover external diagnostics, internal architecture, every common failure mode, component-level repair procedures, the differences between the APW9 and APW9+, preventive maintenance, and the honest assessment of when to repair versus when to replace. This is the guide our technicians wish existed when S17-series units started arriving at our bench in volume — so we wrote it.

APW9/APW9+ Technical Specifications

Before you diagnose anything, you need to know exactly what the APW9 and APW9+ are designed to deliver. These are the baseline specifications you will compare your measurements against during diagnostics. Any deviation from these values tells you what has failed and where to begin your investigation.

APW9/APW9+ Miner Compatibility

Safety Warnings

PSU repair is fundamentally different from miner repair. Inside the miner, the dangerous voltages are limited to the PSU connectors. Inside the PSU itself, you are face-to-face with the full fury of rectified AC mains voltage. This section is not legal boilerplate — it is the knowledge that keeps you alive on the bench.

We follow this protocol on every single PSU repair at D-Central. No exceptions, no shortcuts, no matter how many thousands of units we have worked on. Complacency around high voltage is what gets experienced technicians hurt. Respect the energy stored in those capacitors every single time.

Required Tools

For a comprehensive guide on using a multimeter for ASIC diagnostics — including proper probe technique, measurement modes, and safety practices — see our Multimeter Guide for ASIC Repair.

External Diagnostics (No Disassembly Required)

Before you reach for a screwdriver, you can diagnose the majority of APW9/APW9+ failures without opening the enclosure. External diagnostics eliminate entire categories of faults and often identify the problem outright. Follow these steps in order — do not skip ahead.

Step 1: Visual Inspection (Power Off, Unplugged)

With the APW9/APW9+ completely disconnected from AC power and the miner, inspect the unit using your eyes and nose. A surprising amount of diagnostic information is available without a single measurement:

• Scorch marks or discoloration — Look at the enclosure surface near the AC input terminal and near the output connector bank. Darkening or blistering of the paint indicates a severe thermal event occurred inside.
• Burnt smell — Put your nose near the ventilation slots. A distinct acrid smell means a component has burned. MOSFETs produce a sharp electronic burn smell; capacitors smell slightly sweet and chemical; PCB material smells like burnt phenolic resin. Any of these smells means something has already failed inside.
• Connector inspection — This is critical for the APW9. Examine every 6-pin output connector for signs of heat damage: discolored pins (should be silver/gold — black, blue, or brown means heat damage), melted plastic housing, pitting or carbon deposits on pin surfaces. Burnt connectors are the single most common APW9 failure mode.
• Fan blade condition — Look through the fan grilles for cracked, broken, or missing fan blades. Spin each fan by hand through the grille — it should rotate freely with minimal friction. Grinding or catching indicates bearing wear.
• Rattling sounds — Gently tilt and shake the unit. Listen for loose components. A rattling sound means something has broken free from the PCB. Do not power on.
• Physical damage — Dents in the enclosure, bent AC input prongs, cracked casing. Shipping damage is common, especially with used units purchased online.
• Cable condition — Inspect the output cables for fraying, exposed copper, melted insulation, or kinks. The 6-pin cables carry high current — damaged insulation is a short circuit waiting to happen.

Step 2: LED Indicators & Fan Behavior

Connect the APW9/APW9+ to a known-good 200–240V AC outlet without connecting it to any miner. Observe what happens in the first 5 seconds:

• Fans spin up — The two cooling fans should begin spinning within 1–2 seconds of AC power being applied. This confirms the auxiliary power supply circuit is functional and the fans themselves are alive.
• No fans, no lights — Completely dead. Most likely cause: blown input fuse or catastrophic PFC failure. See Failure Mode 1.
• Fans spin briefly then stop — The PSU is attempting to start but a protection circuit is triggering immediately. This usually indicates a short on the output side or a PFC stage fault.
• LED indicator (if present) — Some APW9/APW9+ revisions have a status LED near the AC input. Green = normal standby. Red or no LED with fans running = fault condition. No LED at all may simply mean your revision does not have one — not all do.
• Abnormal fan speed — If fans run at full speed immediately (maximum RPM, very loud) without any load, the temperature sensor or fan controller circuit may have failed, defaulting to maximum cooling as a safety measure.

Step 3: Output Voltage Check

With the PSU connected to AC power and no miner attached, use your multimeter to check the output connectors.

Important: The APW9/APW9+ main output may or may not be active without a miner connected, depending on the firmware revision and whether the PSU requires an enable signal from the control board. On many APW9 units, the main output is active by default when AC power is applied (unlike the APW12 which requires an explicit EN signal). This behavior varies by revision.

To measure: set your multimeter to DC voltage mode. Place the red probe on any pin in the 6-pin output connector and the black probe on a ground pin. The 6-pin connector has three power pins and three ground pins — power pins are typically on one side, ground on the other. If you read the expected voltage range, the PSU’s output stage is functional at least at no-load. If you read zero, and the fans are running, the main conversion stage has failed or the PSU is waiting for an enable signal.

Step 4: Detailed Connector Inspection

This step deserves its own section because connector failure is the defining issue of the APW9/APW9+. The 6-pin PCIe-style connectors on the APW9 carry enormous current — each connector may handle 25–40A depending on configuration, and the S17/T17 series uses multiple connectors per hashboard.

Remove each 6-pin connector from the miner (if connected) and inspect individually:

• Pin surface condition — Clean pins are bright silver or gold-plated. Any discoloration (dark spots, bluing, blackening) indicates heat damage from resistance at the contact point.
• Pin tension — Each female socket should grip the male pin firmly. If a connector slides on and off with almost no force, the contact tension has been lost (usually from overheating). This is a connector that will fail.
• Plastic housing — Look for warped, melted, or discolored plastic. The nylon housing has a melting point around 260°C. If it is deformed, the internal pins were running far too hot.
• Carbon deposits — Black residue on or between pins is carbon from arcing. This indicates the connection was making and breaking contact under load — an extremely dangerous condition that generates enormous localized heat.

Disassembly Procedure

If external diagnostics point to an internal fault, or if you need to perform preventive maintenance (cleaning, thermal paste replacement, fan swap), you will need to open the APW9/APW9+. Follow this procedure exactly.

Opening the APW9/APW9+ Safely

1. Disconnect AC power cord from the wall outlet. Not from the PSU — from the wall. Then disconnect from the PSU as well.
2. Disconnect all DC output cables from the miner.
3. Wait a minimum of 5 minutes. The internal bleed-down resistors will discharge the PFC capacitors. Set a timer — do not estimate.
4. Place the PSU on an anti-static mat or grounded work surface. Attach your anti-static wrist strap to the PSU chassis.
5. Remove the enclosure screws. The APW9/APW9+ uses Phillips head screws along the top cover edges and sometimes along the sides. There are typically 6–8 screws. Keep track of their locations — some may be different lengths.
6. Lift the top cover carefully. Some revisions have internal cables (fan wires, sensor wires) attached to the top cover. Do not yank it off — lift slowly and check for any connected wires. Disconnect fan headers if the fans are mounted to the top cover.
7. Before touching anything: measure PFC capacitor voltage. Set your multimeter to DC voltage, 600V range. Locate the large electrolytic capacitors (they are the tallest cylindrical components on the board, usually two in series or parallel near the AC input section). Measure across their terminals. Voltage must be below 10V.
8. If voltage is above 10V: use your discharge resistor (10KΩ 10W wirewound) across the capacitor terminals. Hold until voltage drops below 10V. At 400V, the initial discharge current is ~40mA — safe for the resistor but enough to be lethal to you, so use insulated leads.
9. Confirm discharge on all large capacitors. The APW9 may have multiple banks of capacitors. Check each one.
10. You are now safe to inspect and work on the internal components.

Internal Component Overview

Understanding the APW9/APW9+ internal architecture is the foundation of effective diagnostics. When something fails, knowing the signal path from AC input to DC output lets you systematically narrow down the fault to a specific stage. The APW9 converts power through four main stages.

Stage 1: AC Input & EMI Filter

Power enters the APW9 through a single AC input terminal (unlike the APW12 which has dual inputs). The first circuit encountered is the EMI filter, which prevents high-frequency switching noise from feeding back into your building’s electrical system.

• Input fuses (F1, F2) — Glass or ceramic cartridge fuses rated for full input current (20A typically). First line of defense against downstream shorts. When a fuse blows, it is a symptom — something downstream caused it.
• Metal Oxide Varistor (MOV) — Absorbs voltage spikes from the AC line. Sacrificial component — it absorbs hits so the rest of the circuit does not have to.
• Common-mode chokes and Y-capacitors — Filter high-frequency noise.
• NTC thermistor (inrush current limiter) — Limits the massive current spike when bulk capacitors first charge. Without this, the inrush current would blow the fuses or trip the breaker every time you plug in.

Stage 2: Bridge Rectifier & PFC (Power Factor Correction)

After the EMI filter, AC power hits the bridge rectifier, converting AC to pulsating DC. This raw DC enters the active PFC (Power Factor Correction) stage — one of the most critical and failure-prone sections of the APW9.

The PFC stage uses a boost converter topology to:

1. Correct power factor — Shapes input current to follow input voltage, achieving PF > 0.99. This is legally required for commercial equipment and reduces wasted reactive power.
2. Boost voltage — Converts rectified AC (~310V DC peak from 220V AC) up to a regulated 390–420V DC bus stored on the bulk electrolytic capacitors.

Key PFC components:

• PFC MOSFETs — High-voltage switching transistors (600V+ rated). These switch at tens of kHz and handle the full bus voltage and input current. They run hot and are a common failure point in the APW9.
• PFC diodes — Fast-recovery or SiC diodes handling the boost inductor flyback current.
• PFC boost inductor — Large toroidal inductor that stores energy during each switching cycle.
• Bulk electrolytic capacitors — Large capacitors (420V rated, typically 330–470µF each) that store the 390–420V DC bus. These are the components that retain lethal charge after power-off.
• PFC controller IC — Manages switching frequency and duty cycle to maintain target bus voltage.

Stage 3: LLC Resonant Converter (DC-DC Conversion)

The PFC stage produces a stable ~400V DC bus. The LLC resonant converter steps this down to the required 14.5V output. The APW9 uses a half-bridge LLC topology, achieving high efficiency through zero-voltage switching (ZVS).

Key components:

• Main switching MOSFETs — Form the half-bridge legs. Switch at hundreds of kHz in the resonant frequency range.
• Resonant inductor and capacitor — Form the resonant tank circuit with the transformer’s magnetizing inductance.
• Main power transformer(s) — Step down high-voltage primary to low-voltage secondary. High-frequency operation (100+ kHz) allows compact size despite enormous power throughput.
• PWM controller IC — Controls switching frequency to regulate output. Works with feedback from the secondary side via optocoupler isolation.

The LLC stage is the most complex section of the APW9 and the most expensive to repair. MOSFET failure here frequently causes cascading damage to the gate driver, controller IC, and sometimes the transformer. When this stage fails, the repair-vs-replace calculation often tips toward replacement.

Stage 4: Synchronous Rectification & Output

On the secondary side of the transformer, the APW9 uses synchronous rectification — MOSFETs instead of diodes for the rectifying function. This reduces conduction losses and contributes to the APW9’s efficiency.

The output side contains:

• Synchronous rectifier MOSFETs — Low-voltage, high-current MOSFETs performing rectification. Degradation of these components causes voltage droop under load.
• Output filter inductors — Smooth pulsating DC into clean DC.
• Output filter capacitors — Low-ESR electrolytic and ceramic capacitors filtering remaining ripple. These are the components most affected by age and heat — and one of the most common APW9 repair targets.
• Output bus bars / heavy copper traces — Carry up to 248A to the output connectors. These get hot under full load.
• Current sense resistors — Shunt resistors for overcurrent protection measurement.
• 6-pin output connectors — Soldered to the PCB. These are where the PSU meets the miner, and where connector overheating failures originate.

Common Failures and Repairs

After repairing hundreds of S17/T17-series miners and their APW9/APW9+ power supplies at our Laval facility, these are the failure modes we see most frequently. Understanding the symptom-to-cause mapping lets you diagnose faster and order the right parts before you even open the unit.

Failure 1: Connector Overheating & Burn (Most Common)

Symptom: Miner drops hashboards intermittently. PSU may shut down under load. Upon inspection, one or more 6-pin connectors show heat damage — melted plastic, discolored or blackened pins, carbon deposits. In severe cases, the PCB traces beneath the connector are visibly burned.

Root cause: The 6-pin PCIe-style connectors were not designed for the continuous high-current loads that Bitcoin mining demands. Each connector carries 25–40A continuously, 24/7, for months or years. Any increase in contact resistance — from loose fit, oxidation, dust, or micro-arcing — creates a hot spot that progressively degrades the connection.

Repair procedure:

1. Cable-side only damage: If only the female cable connector is damaged, replace the entire cable assembly. Do not attempt to re-pin a burnt connector. Cost: $10–20.
2. PCB-side damage: If the male pins soldered to the PSU board are burnt, you must open the PSU (following the safety protocol), desolder the damaged connector, clean the pads, and solder a new connector. Use only high-quality connectors rated for the current. If PCB traces are burned through, trace repair with copper wire jumpers is possible but fragile.
3. Prevention: After any connector repair, apply dielectric grease (DeoxIT D5 or similar) to the connector pins. This prevents oxidation and maintains low contact resistance. Inspect connectors quarterly.

Failure 2: Completely Dead PSU (No Output, No Fans)

Symptom: AC power is connected and verified at 200–240V. No fans spin, no LED indicators, no output voltage. Completely unresponsive.

Common causes:

• Blown input fuses (F1/F2) — The most common cause of a completely dead PSU and the easiest to fix. Fuses blow from voltage spikes, MOV failure, or a downstream short. Test with multimeter in continuity mode.
• Failed bridge rectifier — Shorted diodes in the rectifier, often caused by the same event that blew the fuses. Test each diode: should show ~0.5–0.7V forward, open-circuit reverse.
• PFC MOSFET failure — Shorted MOSFETs in the PFC boost converter. Test drain-to-source in diode mode: should show body diode in one direction, open in the other. A short in both directions means the MOSFET has failed.
• NTC thermistor open-circuit — The inrush current limiter has burned open, breaking the power path. Test continuity — should read low resistance (2–10Ω at room temperature).

Failure 3: PFC Circuit Failure

Symptom: PSU is dead or makes a soft clicking/ticking sound when AC is applied. Fans may or may not spin briefly. No output voltage. Upon opening (after safety protocol), the PFC bulk capacitors show no voltage or significantly low voltage (< 350V).

Common causes:

• PFC MOSFETs shorted — The most common PFC failure. These transistors handle the full bus voltage and high current. Thermal stress from 24/7 operation in hot environments is the primary killer.
• PFC controller IC failure — The IC that manages the PFC switching has failed, meaning the MOSFETs no longer receive proper gate drive signals. No switching means no voltage boost.
• PFC diode failure — A shorted or open boost diode prevents proper energy transfer in the PFC circuit.
• Bulk capacitor failure — Capacitors with severely degraded capacitance cannot maintain the bus voltage. Look for bulging tops, leaking electrolyte, or elevated ESR readings.

Difficulty: Advanced   |   Risk: High (primary-side components see 400V+)

PFC circuit repair requires identifying the exact failed component, sourcing an exact replacement, and performing SMD soldering on a board that carries lethal voltage during operation. This is professional-level work. If you are not experienced with high-voltage power electronics, send the unit to D-Central for professional repair.

Failure 4: LLC MOSFET Failure

Symptom: Fans may spin, but no DC output voltage. Or the PSU produces output but immediately shuts down. Burning smell may be present. Upon inspection, one or more MOSFETs in the LLC stage show visible burn damage on the PCB.

Repair considerations:

• LLC MOSFET replacement requires hot air rework and exact-match replacement parts.
• Cascading damage is common — when an LLC MOSFET fails, it frequently takes out the gate driver IC, controller IC, and associated passive components. You cannot replace just the MOSFET without checking every surrounding component.
• If the main transformer is damaged (rare but possible), the PSU is economically unrepairable — the transformer is a custom wound component that is not readily available as a spare.
• This is a replacement-level failure for most home miners. The parts cost, tooling requirements, and risk of re-failure make professional repair or unit replacement the economically rational choice.

Failure 5: Fan Failure

Symptom: One or both fans do not spin, or spin erratically with grinding/clicking sounds. PSU may work initially but shuts down after 10–30 minutes under load (OTP protection triggering as internal temperature rises).

Repair:

Difficulty: Beginner   |   Time: 15–20 minutes   |   Risk: Low

1. Open the PSU following the safety protocol.
2. Identify the failed fan. Spin each by hand — a good fan spins freely and silently. A failed fan grinds, catches, or does not spin.
3. Disconnect the fan from its header on the PCB. Note wire colors and connector orientation.
4. Remove the fan (typically held by screws or clips).
5. Install the replacement with matching specifications:
   • Size: 60mm × 60mm × 15mm (or 25mm depth, varies by revision)
   • Voltage: 12V DC
   • Connector: Match pin count (2-pin, 3-pin, or 4-pin PWM)
   • Airflow direction: Check the arrow on the fan frame — must match original orientation
6. Reconnect, reassemble, test. Fans should spin within 1–2 seconds of AC power application.

Failure 6: Output Capacitor Degradation

Symptom: PSU works fine for the first 10–30 minutes, then the miner begins reporting voltage errors, hashboards drop, or hashrate drops. Restarting temporarily resolves the issue (because the PSU cools down). Over weeks, the time-to-failure gets shorter as the capacitors degrade further.

Root cause: The output electrolytic capacitors on the secondary side have lost effective capacitance due to age and heat. They still function when cold, but as the PSU warms up during operation, the degraded capacitors lose capacitance and the voltage droops under load. An oscilloscope will show increasing ripple voltage as the PSU heats up.

Repair:

Difficulty: Intermediate   |   Time: 30–60 minutes   |   Risk: Medium

1. Open the PSU and identify the output capacitors (secondary side, near the output connectors). Look for bulging tops, leaking electrolyte, or signs of overheating on the PCB beneath them.
2. Note the exact specifications printed on each capacitor: capacitance (µF), voltage (V), temperature rating (°C), ESR if listed, and physical dimensions.
3. Desolder and replace with capacitors of equal or better specifications. Use high-quality 105°C rated capacitors from reputable manufacturers (Nichicon, Rubycon, Panasonic, Nippon Chemi-Con).
4. Observe polarity — electrolytic capacitors are polarized. The negative terminal is marked with a stripe. Reversing polarity can cause the capacitor to fail violently.
5. For PFC bulk capacitors (primary side, 400V+ rated): use only capacitors specifically rated for PFC/switching applications with high ripple current ratings.

Failure 7: Intermittent Shutdown Under Load

Symptom: PSU starts normally and the miner begins hashing, but the PSU shuts down randomly after minutes or hours. May restart on its own or require AC power cycling.

Systematic diagnosis:

• Check fans first — Feel for vibration, listen for grinding. A failing fan causes OTP shutdown under load. This is the most common cause and cheapest fix.
• Check for dust — Dust insulation on heatsinks causes thermal throttling. Compressed air cleaning from the exhaust side (blowing dust out the way it came in) often resolves this.
• Check input voltage during operation — Monitor your wall voltage with a multimeter while the miner is running. If it sags below 190V under load (common in hot weather when HVAC is running), UVP triggers. The fix is electrical infrastructure, not PSU repair.
• Check output capacitors — Marginal capacitors that fail under thermal load. See Failure 6.
• Check for cold solder joints — Thermal cycling creates cracks in solder joints on high-current paths. Visible under magnification. Re-flow the suspect joint.

Voltage Testing Reference

For technicians performing board-level diagnostics with the APW9/APW9+ open and powered (or measuring during controlled power-on), these are the expected voltages at key circuit points. Any measurement on a powered PSU must be performed with extreme caution — the PFC bus carries lethal voltage.

START: Is AC power present at the wall outlet? (Measure: 200-240V AC)
  |
  +-- NO --> Fix power source. Not a PSU issue.
  |
  +-- YES --> Plug in APW9/APW9+ (no miner connected). Do fans spin?
       |
       +-- NO --> PSU is completely dead.
       |    |
       |    +-- Check fuses F1/F2 (continuity test).
       |    |    |
       |    |    +-- Fuse blown --> Test bridge rectifier & PFC MOSFETs for shorts.
       |    |    |                  Fix root cause, then replace fuse.
       |    |    |
       |    |    +-- Fuses OK --> Auxiliary supply failure or NTC open.
       |    |                     Board-level repair or replace PSU.
       |    |
       |    +-- Check NTC thermistor (continuity: should be 2-10 ohms).
       |
       +-- YES --> Measure DC output at 6-pin connectors.
            |
            +-- No output (0V) --> PFC bus present? (Measure bulk caps)
            |    |
            |    +-- PFC bus OK (390-420V) --> LLC stage failure.
            |    |                             Advanced repair or replace.
            |    |
            |    +-- PFC bus absent/low --> PFC circuit failure.
            |                               MOSFETs, controller IC, diodes.
            |
            +-- Output present (14-15V) --> Connect to miner. Does it hash?
                 |
                 +-- YES, stable --> PSU is functional.
                 |
                 +-- YES, but shuts down after time --> Check fans/dust
                 |   (OTP), input voltage sag (UVP), or marginal caps.
                 |
                 +-- YES, but hashrate low / errors --> Measure output
                 |   voltage under load. If low: output caps or
                 |   rectifier MOSFETs. If ripple high: output caps.
                 |
                 +-- Miner reports hashboard missing --> Check specific
                     connector for that hashboard. Likely burnt connector.

Reassembly and Load Testing

After any internal repair, follow this verification sequence before connecting the APW9/APW9+ to a miner. Skipping verification steps risks damaging your miner on the first power-up — a PSU repair that did not fully address the root cause can produce overvoltage, shorts, or unstable output that destroys hashboards worth far more than the PSU.

Reassembly Checklist

1. Visual inspection — No loose components, no solder bridges, no tools left inside, all internal connectors seated, all fan headers connected.
2. Short-circuit test — With the enclosure open and no AC power, measure resistance across the output connectors (any power pin to any ground pin). You should read some resistance (from the output inductors and capacitors) — not zero ohms. Zero ohms means there is a short that will cause immediate failure on power-up.
3. Close the enclosure — Replace all screws. Ensure the top cover seats flat with no pinched wires.
4. Standby test — Connect AC power with no miner attached. Verify fans spin. If the unit has a status LED, verify it shows the expected state.
5. No-load output test — Measure DC output at the 6-pin connectors. Should read 14.2–15.0V DC.
6. Load test — Connect to a miner and monitor for at least 2 hours:
   • Watch for stable output voltage (within ±3% of target)
   • Monitor miner web interface for any PSU-related error messages
   • Check all hashboards are online and hashing at expected rates
   • Feel the PSU enclosure for excessive heat (> 60°C external = problem)
   • Listen for abnormal sounds (buzzing, clicking, high-pitched whining)
7. Aging test (professional standard) — Run at 80%+ rated load for 2+ hours continuously. This stresses repaired components and catches marginal repairs that would fail in the first days of operation.

APW9 vs. APW9+ Differences

The APW9 and APW9+ are often discussed interchangeably, but they are distinct hardware revisions with meaningful differences. Understanding these differences matters for sourcing replacements, selecting the right unit for your miner, and knowing what to expect during repair.

Bottom line: The APW9+ is the better unit. If you are buying a replacement, buy the APW9+. The efficiency improvement reduces heat generation, and the improved components extend service life. If you are repairing an existing APW9, the repair procedures are identical between the two — the PCB layout and component locations are the same, with only some component values and part numbers differing in specific positions.

APW9 in Bitmain’s PSU Lineage

Interchangeability Notes

• APW3++ on S17/T17: ABSOLUTELY NOT. The APW3++ delivers 1,600W. The S17 requires 2,500W+. The APW3++ will either fail to start the miner or overheat and fail catastrophically.
• APW7/APW8 on S17/T17: NO. Insufficient power capacity.
• APW9 on S19-series: MARGINAL ON T19, NO ON S19. The APW9 can technically produce enough wattage for a T19 (~2,800W), but leaves virtually no headroom. An S19 at 3,250W exceeds the APW9’s safe continuous rating. Operating a PSU at 95%+ capacity 24/7 is a recipe for premature failure.
• APW12 on S17/T17: POSSIBLE but check voltage. The APW12 can power S17/T17 series as it covers the required voltage range. However, the I2C communication between the APW12 and S17/T17 control boards may not function correctly in all configurations — the S17/T17 control boards were designed for the APW9’s communication protocol. If using an APW12, verify output voltage under load matches what the miner expects.
• APW9+ in place of APW9: YES. Drop-in replacement. Same form factor, same connectors, same pinout. Always prefer the APW9+ when given the choice.

Preventive Maintenance

The APW9/APW9+ is a workhorse, but it is not invincible. Heat, dust, and neglected connectors are its enemies. A simple maintenance schedule dramatically extends service life and prevents the catastrophic failures described in this guide. Prevention is always cheaper than repair.

Monthly: Dust Cleaning

Dust is the primary killer of power supplies in mining environments. It insulates heatsinks, restricts fan airflow, and holds moisture against PCB surfaces.

• Use compressed air (canned or electric blower) to blow through the PSU ventilation slots from the exhaust side.
• Pay particular attention to fan blades and grilles — these accumulate dust fastest.
• Perform cleaning with the PSU powered off and disconnected.
• In dusty environments (garages, basements, outdoor enclosures), increase to bi-weekly cleaning.

Quarterly: Connector Inspection & Treatment

This is the single highest-ROI maintenance task for the APW9/APW9+. Connector failure is the #1 failure mode, and it is almost entirely preventable.

• Power down and disconnect. Remove each 6-pin connector from the miner.
• Inspect every pin for discoloration, pitting, or carbon deposits.
• Check that each connector grips firmly. If a connector slides on with little resistance, replace the cable.
• Clean pin surfaces with IPA and a cotton swab.
• Apply a thin film of dielectric grease (DeoxIT D5 or similar) to all connector pins before reconnecting.
• When reconnecting, push each connector firmly until it clicks or seats fully. Wiggle gently to confirm it is locked.

Annually: Thermal Paste Replacement

The thermal paste between power components and heatsinks degrades over time, especially in the high-temperature environment inside a mining PSU. Annually (or every 8,000 hours of operation):

• Open the PSU following the safety protocol.
• Remove heatsinks from MOSFET packages (if accessible — some are permanently bonded).
• Clean old thermal paste with IPA and lint-free cloths.
• Apply fresh thermal paste (thin, even layer) or replace thermal pads with matching thickness.
• Reassemble ensuring proper mounting pressure.

Ongoing: Input Power Quality

• Surge protection — Use quality surge protection. The internal MOV is sacrificial and degrades with each absorbed spike.
• Voltage stability — Monitor your wall voltage during peak load. Operation below 200V stresses the PFC stage. Consider an AVR or line conditioner if you experience frequent sags.
• Dedicated circuits — Run each S17/T17 miner on its own dedicated 20A/240V circuit. Sharing circuits with high-inrush loads causes transients that stress the PSU.
• Proper grounding — Ensure a solid ground connection. Poor grounding increases EMI and eliminates a critical safety path.

Ongoing: Temperature Management

For every 10°C increase in operating temperature above 85°C, electrolytic capacitor life roughly halves. Keep your mining environment as cool as practical:

• Ensure adequate airflow spacing around the PSU. Do not stack units.
• Ambient temperature below 35°C is ideal.
• Position the PSU so it draws cool air, not the hot exhaust from the miner.
• In Canadian climates, leverage the cold. Mining in a basement, garage, or purpose-built room with outdoor air intake provides free cooling for most of the year. This is one of the genuine advantages of mining in Canada — a point D-Central has been making since 2016.

Frequently Asked Questions

Why D-Central for APW9/APW9+ Repair

This guide gives you everything you need to diagnose the most common APW9/APW9+ failures and handle the repairs that are within reach of a competent home miner. But PSU repair — especially on the primary (high-voltage) side — is fundamentally different from miner repair. It requires specialized equipment, component-level soldering skills, and the confidence to work safely around 400V+ DC. When the repair exceeds your comfort level, skill set, or tooling, that is when you contact us.

D-Central Technologies has been repairing ASIC mining hardware since 2016. Our facility in Laval, Quebec has processed 2,500+ miner repairs, and the S17/T17 generation — with its notoriously challenging reliability profile — is hardware we know intimately. We have repaired every failure mode described in this guide hundreds of times over. We stock replacement APW9 and APW9+ units, replacement cables and connectors, internal components for board-level repair, and the test equipment to verify every repair to factory specification before it ships back to you.

Contact D-Central for APW9/APW9+ repair when:

• You suspect PFC or LLC stage failure (high-voltage, board-level work)
• You have cascading component failures (multiple parts damaged simultaneously)
• You do not have the soldering equipment for SMD rework
• You are not comfortable working near 400V+ DC
• You need a guaranteed, tested repair with accountability
• Your downtime cost exceeds the repair cost — we typically complete PSU repairs in 1–3 business days

Every hash counts. And every hash depends on clean, stable power from a healthy PSU. Whether you maintain your APW9 yourself or send it to us, the goal is the same — keep your S17/T17 running, keep your hash rate up, and keep contributing to the decentralization of Bitcoin’s hash power. That is what we are here for.

Related Guides:

• APW12 Power Supply Repair & Diagnostics Guide — For S19-series miners
• Multimeter Guide for ASIC Repair — Essential measurement techniques
• Antminer S17 Maintenance & Repair Guide
• D-Central ASIC Repair Services

Contact D-Central: 1-855-753-9997  |  Contact Form  |  ASIC Repair Page