IceRiver Error Codes: Complete Troubleshooting Reference (All KS Models)

Hard power-cycle at the breaker for 30 seconds. Not a soft reboot from the web UI — a full mains kill. Clears wedged daemon state and resets the I2C bus. On single-event codes from the 4xx and 7xx buckets, this alone clears 40% of tickets we see at the bench.

Pull the exact error code, the LED pattern, and the cadence. Web UI: `Status` → `Miner Log`. Photograph the live LED block. Note whether the code fires once or repeats. Write it all down before doing anything else — diagnostic discipline saves you from chasing the wrong subsystem.

Find the miner's IP. If the web UI is unreachable: open IceRiver's "Detect IP" tool from iceriver.io, or check your router's DHCP lease table. Verify the miner has a valid lease — no IP means the network bucket is your fault, not the hardware.

Shop-vac the intake filter and grille. Dust on filters = higher inlet temp = `350`-`352` overheat trips and `233`-`239` PSU thermal trips. Wipe the grille, vacuum the intake, verify nothing within 15 cm of the front of the chassis is blocking airflow. Five minutes of work resolves a meaningful share of thermal-bucket cadences.

Verify ambient with an IR thermometer at the intake grille. Target ≤ 35 °C for KS0/KS1/KS2 line, ≤ 30 °C for KS3/KS5 line under sustained load. Hot Canadian basements in summer hit 38 °C more often than people think — check before you blame the hardware.

Replace a failing fan. KS1/KS2/KS3 line uses a `12038` (120 mm × 38 mm) fan, typical part `DF1203812B2UN` or equivalent at 6000 RPM. KS0/KS0 Pro uses a smaller `4010` internal plus a `12025` or `12038` external. Power off, unscrew, unplug connector, swap, screw back, plug in, reboot. Codes 110/111/120/121/140 clear within 30 seconds of boot.

Reseat hashboard ribbon cables. Power off at the breaker. Open the chassis. Disconnect the ribbon and power connectors at each hashboard, visually inspect contacts for blackening or oxidation, reconnect firmly. Listen for the click. Sensor codes (`300`-`302`) on a single board often clear with this alone.

Try a firmware rollback through the web UI. If you started seeing the cadence right after an OTA upgrade, the new firmware is the suspect. Download the previous IceRiver-published version from iceriver.io/firmware-download/, flash through `System` → `Upgrade` over a wired connection. Most 7xx/8xx post-upgrade codes clear with rollback to last-known-good.

Verify AC line voltage at the wall under full mining load. Multimeter on AC, probe at the outlet while the miner is hashing at full power. Standard IceRiver PSUs accept 180 V-285 V AC. Sag below 190 V sustained = circuit undersized. Move to a dedicated 240 V circuit — single biggest improvement for North American Canadian installs running KS5/KS5L/KS5M.

Refresh thermal paste on the chip-side of the hashboards. KS5/KS5L/KS5M chips age fast on dried stock paste. Arctic MX-6 or Thermal Grizzly Kryonaut, uniform thin layer, don't glop it on. Reassemble carefully — torque heatsink screws evenly to avoid PCB bow. 350-352 cadences under healthy ambient often clear post-refresh.

SD-card reflash recovery. When `800`/`801`/`802` cycles persist after factory reset, the eMMC partition table is corrupt and OTA can't recover. Download the recovery image from iceriver.io/firmware-download/, write to microSD with balenaEtcher or `dd`, follow the model-specific procedure (boot button + insert at power-on for most KS units). Soak-test for an hour after reflash.

Replace the PSU on a `KS5L`/`KS5M`. The `BP-H-3640` family runs close to its rated headroom on the flagship KS5 line. Symptoms: repeating 233-239 codes under healthy ambient, intermittent reboots under hash load, audible PSU fan ramp at low ambient. Order a known-good replacement, swap it in (connector job, not soldering — within reach of an intermediate hands-on owner).

Replace a `1004LV100` ASIC chip. When a single board on a KS5L/KS5M reads zero hashrate or partial chip count (`9`/`26`/`52`), one or more `1004LV100` chips have failed silicon. Hot-air rework: preheat to 150 °C, hot-air at 310-330 °C for ~30 s, reflux, lift, replace, reflow. Practice on a scrap board first — KS5L hashboards aren't cheap.

Replace a temperature sensor IC. `300`/`301`/`302` codes that persist after ribbon reseat = dead I2C sensor (small SOT-23 / SOT-353 package near the hashboard's I2C bus). Identify under microscope, hot-air-replace, retest. If you don't own a microscope and a hot-air station, this is a Tier 4 job — don't risk the board.

Inspect electrolytic capacitors on the control board. Bulging tops or leaking electrolyte = end-of-life. `710`/`711`/`712` cadences on a 3+ year old KS unit are often caps drying out under sustained operating temperature. Replace with same-spec or higher-temp-rated equivalents.

Stop DIY and ship to D-Central when: (a) single chip-position failure repeats on two different boards in the same rig (PCB-level fault), (b) control board suspected and you don't own a hot-air station, (c) PSU shows scorched traces or burnt-component odour, or (d) repeated 7xx cadences across multiple firmware rollbacks. Past those points DIY costs more in time + risk than the repair itself.

D-Central bench process: test fixture with programmable load, per-chip isolation, sensor and PMIC continuity testing, hot-air rework with profile control, post-repair 24-hour burn-in at nameplate hashrate before sign-off. We're a Western retail repair authority — the only Western alternative for KS hardware is shipping to China (Zeus), and the trust profile of that path is not great.

Ship safely. Pack hashboards in anti-static bags, double-box with at least 5 cm of foam on every side. Include the original PSU and AC cable so we can reproduce the AC environment. Include a printed copy of the codes / LED states / cadence — saves us diagnostic time, which saves you money.