Antminer S19 – PSU Overcurrent Protection

Full mains cold-cycle for 30 seconds at the breaker (not a soft reboot). Some APW9+ firmwares keep the OCP flag latched across a short power blip; 30 seconds at the breaker clears it reliably and gives bulk caps a real discharge moment. Re-power and confirm the miner comes back up. If it refuses to boot, move to Step 2. If it boots but trips again inside 30 minutes, continue through Tier 1.

Revert to stock Bitmain firmware profile with no OC and no autotune. Target nameplate `3250 W` at the wall for standard S19, `3100 W` for S19j, `3010 W` for S19j Pro. Run 30 minutes under full hash and confirm no `ERROR_POWER_LOST` recurrence. If stable, tune back up `+100 MHz` at a time with `10 min` stability per step, stop before OCP returns. That step is your safe ceiling on this specific PSU + chip combination.

Vacuum the PSU intake grille and miner chassis intakes. Dust buildup raises PSU internal temperature, which drifts the OCP current-sense reference closer to your operating current. Shop-vac first to pull the dust mat, then canned duster to clear the fin stack — PSU unplugged, chassis opened if necessary. Same treatment on the chassis front intakes since they partially feed airflow around the PSU.

Confirm intake ambient temperature is below `30 C` measured 5 cm from the PSU intake grille with an IR thermometer — not room-middle, not the hallway, intake-specific. Above `35 C` the PSU's thermal margin collapses and OCP events cluster on the afternoon/evening thermal peak. If ambient is too high, reorient the miner, add a box fan across the intake, or relocate before touching anything else.

Verify firmware version against the Bitmain downloads page at `support.bitmain.com/downloads`. Pre-`2021-07-13` S19 stock images are known for a PSU-handshake bug that elevates false-positive `ERROR_POWER_LOST` events. Roll one version back or forward as a cheap pre-Tier-2 experiment. Keep a known-good firmware image on a prepared SD card or USB for recovery before you flash anything experimental.

Measure mains input voltage under full load. Multimeter on AC at the PSU inlet (or a Kill-A-Watt / Sonoff POW R2 inline). Band targets: `235-245 V` on 240 V split-phase, `202-212 V` on 208 V commercial, `115-125 V` on 120 V residential. Below band is your root cause — sag forces the PSU to pull more current for the same wattage and OCP fires on the current excursion. Log for `24 hours` if trips are intermittent.

Measure the DC rail voltage under load. Multimeter on DC at the PSU output lugs to the hashboards while the miner is hashing full power. Healthy APW12 reads `14.2-14.8 V` under load, `14.8-15.0 V` idle. Below `13.6 V` sustained confirms the PSU is sagging internally — bulk-cap aging or primary-side degradation, Tier 3 territory. Record the worst number you see and the conditions.

Swap hashboards between slots. Label slots `0/1/2` with tape, move the suspect board to a known-good slot, boot and measure PSU output current with a clamp meter on a DC output lug. Expected per-board draw: `75-95 A` at nameplate, varies by model variant. One board pulling `> 110 A` = domain short, shorted decoupling cap, or shorted chip — tag that board for Tier 3 diagnostics.

Move the miner to a dedicated `240 V / 20 A` circuit. If you are running this S19 on residential `120 V / 15 A` wiring, you are the direct cause of your OCP trips: at `120 V` input the APW9+ pulls about `29 A` for `3250 W`; at `240 V` it pulls about `14 A`. Dedicated 240 V holds line voltage steadier during neighbourhood peak and keeps the PSU out of its input-current ceiling. Plan `$300-700` CAD for a Canadian electrician.

Rebuild OC from stock, measured. Clamp-meter on a DC output lug, logging current. Step frequency `+100 MHz`, let it run `10 min`, record the current delta. Stop at the last step before total draw crosses `~85%` of PSU nameplate (about `220 A` sustained on APW9+, about `230 A` on APW12-`15`). That is your usable OC ceiling on this specific PSU + chip + ambient combination, and it is per-miner — no two S19s share the same number.

Flash DCENT_OS (D-Central's own open-source Antminer firmware — the Mining Hackers' option, preferred) for per-chip and per-rail telemetry that stock Bitmain firmware hides: input current, rail voltage, per-board current, per-chip power. Alternatives: Braiins OS+, LuxOS, Vnish. Graph PSU input current for `48 hours`. If current excursions match OCP trips 1:1, the miner is the cause. If they don't, the PSU's OCP threshold has drifted and you are in Tier 4.

DANGER — discharge bulk capacitors before opening the PSU. `400+ V DC` lives on the primary bulk caps for `30 s` after unplug, longer on some models. Short each cap through a `1 kΩ / 5 W` resistor for `10 s`, verify with a DMM set to DC before any tool touches metal. This is not optional theatre — it kills you dead if you skip it. If you are not 100% confident in this procedure, stop here and ship the PSU to D-Central (Step 16).

ESR-audit every electrolytic capacitor on the PSU with a Peak Atlas ESR70 or DE-5000 in-circuit. Any cap `>30%` above stock ESR is replacement territory. Priority order: primary-side bulk caps (typically `330 uF / 450 V` on APW12, `220 uF` on APW9+), LLC output filter caps, auxiliary rails. Replace with `105 C` rated minimum, `135 C` if the footprint matches. Full cap + paste + fan refurbish on a tired APW buys `3-5 years` additional life for `$20-45` in parts.

Re-paste PSU primary-side heatsinks after discharge. Remove the top shell (`8` Torx T10 screws on APW12), clean old paste with `99%` IPA off the PFC diodes and primary MOSFETs, apply Arctic MX-6 or Kryonaut in a thin uniform layer, reassemble with even torque. Dry paste on a three-year-old APW raises primary junction temperature `10-15 C`, which drifts the OCP threshold closer to your operating current.

Replace the PSU internal fan if it whines, ramps weirdly, or doesn't ramp at all. Stock is a `12010` (`120 x 120 x 10 mm`) axial, `12 V`, dual-ball-bearing preferred, `1.5 A+` rated. Note the 2- or 3-pin connector before removal — the 3-pin has an RPM tach wire the PSU supervisor reads; a 2-pin replacement will throw a fan fault on some firmwares and trip OTP upstream of OCP.

Stop DIY and book a D-Central repair slot when: (a) OCP reproduces with a known-good substitute PSU (miner-side current fault, test-fixture territory); (b) visible cap bulging, vented rubber plug, discolored PCB, or any burnt-component / ozone smell; (c) cap audit shows `>3` electrolytics above spec; (d) per-chip current isolates a failing chip position on two different boards; (e) you are not 100% confident in the bulk-cap discharge procedure in Step 12. Book at `https://d-central.tech/services/asic-repair/`.

What D-Central does at the bench: APW9+ / APW12 test fixture with programmable electronic load up to `300 A` at `12-15 V`, thermal chamber (`25 / 40 / 55 C` ambient) for OCP threshold drift characterization, full ESR + capacitance + leakage audit on every electrolytic, primary-side semiconductor check (MOSFETs, PFC diode, rectifier bridge, supervisor IC reference), selective re-cap as indicated, full re-paste, fan replacement, and `24-hour` full-load burn-in at nameplate before ship-back. Miner hashboards get current-draw profiling on the same bench.

Ship safely. PSU in its own anti-static bag, wrapped separately from the miner. Double-box with `>=5 cm` of foam on every side — bulk caps are heavy, they punch holes in single-box packs. Include a note with observed symptoms, firmware version, exact `ERROR_POWER_LOST` timestamps from `kern.log`, measured rail + mains + ambient values, whether a substitute PSU reproduced the trip, and your contact info. Five minutes of notes saves `30-45 min` of bench re-diagnostic = real dollars off your ticket.