IceRiver Temperature Abnormal Warning on Web Interface

Read the dashboard before touching anything. Capture Temp1/Temp2 per board, fan RPM per fan, hashrate per board, ambient at the inlet, and the exact warning timestamp from the event log. This baseline is what you compare every fix against - skipping the read is the #1 reason DIY repairs fail because you cannot tell if you helped or hurt without a starting point.

Verify ambient at the intake with an IR thermometer aimed at the front grille. IceRiver spec: <=35 C Normal Mode, <=30 C Performance Mode. If you are over, fix the room before anything else - open a window, point a box fan at the intake, lower the thermostat, relocate the miner. Most warnings on healthy hardware clear within 15 minutes once inlet is back in spec.

Clear obstructions. Walk around the miner. Check for >=15 cm clearance front and rear per IceRiver guidance. Move boxes, cables, anything in the airflow path. Confirm the exhaust isn't dumping into the intake of a neighbouring miner - stacking is the silent killer of Temperature Abnormal warnings.

Read the chassis LEDs. D2/D3 flashing simultaneously confirms thermal warning (this page). If you also see D1/D3 (fan fault) or solid D2/D3 (overheat trip), you are already past the warning stage and need a different page from Related Errors.

Read the thermal log. The web UI exposes a rolling event log accessible via the dashboard's log/event panel. Filter for Temperature Abnormal. Frequency tells urgency: once a day at peak ambient = environmental; multiple times an hour = active hardware issue. Co-occurring events (fan RPM dips, hashrate drops, network blips) point at the proximate cause.

Clean the intake filter and grille. Power off at the wall. Pull the filter (KS-class designs vary, typically a slide-out mesh on the front). Shop-vac it. Wipe the grille with microfiber and 99% IPA. Compressed air through the chassis intake from front to back. Reassemble. Restart and observe 30 minutes.

Cross-check sensor readings against reality. With the miner running and the warning active, IR-thermometer the heatsink on each board. Compare to the dashboard Temp2 value. Agreement within 3-5 C means sensors are honest. Disagreement points at sensor drift - escalate to the 300/301/302 page if the gap is >5 C.

Slot-swap the hottest hashboard. KS3/KS5 series only. Label slots 0/1/2 with tape. Move the suspect board to a known-good slot. Power up, observe 15 minutes. Hot follows the board = board issue (paste, pad, board fan duct). Hot stays in the slot = chassis/airflow issue (slot fan, ducting, that section of the airflow path).

Check fan RPM under command. From the dashboard, note commanded RPM vs actual RPM per fan. A 12038 fan commanded to 6000 should read 5800-6000. If actual lags command by >500 RPM, the fan is wearing out and underperforming silently - schedule replacement before it triggers a fan-fault code.

Check exhaust temperature with an IR thermometer. A healthy KS-class miner under load exhausts air +15-25 C above ambient. Significantly less = airflow restriction (filter, ducting, fan). Significantly more = chips dumping more heat than usual (paste failure, OC, ambient).

Refresh thermal paste on the suspect board. Power off at wall. Open chassis. Pull suspect hashboard. Remove heatsink (typically 4-6 screws, may have spring-load washers - note arrangement before removing). Clean old paste from chip and heatsink with 99% IPA. Apply Arctic MX-6 or Thermal Grizzly Kryonaut - uniform thin layer, don't glob. Re-mount heatsink with even torque. Reassemble, restart, observe 30 minutes hashing.

Replace thermal pads on PMIC and voltage-domain ICs. While the hashboard is open, inspect thermal pads on the surrounding voltage-domain components. Crumbled, dried, or compressed past usefulness = replace. Match thickness (usually 1.0 or 1.5 mm) and W/m K rating from the original. Don't skip this - dried pads on PMICs are a common cause of Temp2 climbing while Temp1 looks fine.

Replace a worn fan. If step 9 found a fan lagging command, swap it. Stock IceRiver fans on 12038-class chassis are DF1203812B2UN or equivalent at 6000 RPM. See the dedicated 12038 fan replacement page for the full DIY procedure. New fan = back to spec airflow.

Rollback firmware if a recent update preceded the warning. Pull the prior version's image from your records or contact IceRiver support - prior versions are not always public. Apply via SD-card recovery per IceRiver's documented procedure. Some firmware revisions have shipped with thermal-warning thresholds set too aggressively - rolling back is a valid diagnostic and sometimes the actual fix.

Address PSU heat soak. If diagnostics fingered the PSU, options are: external airflow directed at the PSU surface (a small 120 mm USB-powered fan placed near the PSU intake works), PSU replacement with the same model from a fresh production batch, or full PSU upgrade if available. KS5L/KS5M low-headroom PSUs are the documented case (VoskCoin teardown thread on voskcointalk.com).

Stop DIY and ship to D-Central when: (a) all Tier 3 fixes applied and the warning still fires, (b) per-board isolation pinpoints a board that warms up despite fresh paste + verified airflow + healthy fans, (c) controller-side thermal-driver issue suspected (warning fires across all boards simultaneously even after firmware rollback), (d) a Temperature Abnormal warning has already escalated once to a hard-trip code and now recurs after you cleared it. Book a D-Central ASIC repair slot.

D-Central bench process: test fixture probes the I2C sensor bus directly to confirm sensor honesty independent of the controller; per-board thermal imaging at controlled load to identify hot-spots invisible to dashboard Temp1/Temp2; paste/pad refresh with workshop-grade materials; fan replacement with matched parts; controller-side thermal-driver diagnostics; post-repair 24-hour burn-in at nameplate with continuous logging to verify the warning doesn't recur. D-Central is positioning as the Western English-language IceRiver repair authority.

Ship safely. Pack the suspect hashboards or full miner in anti-static bags, double-box with >=5 cm foam every side. Include a printed note with: model and serial, exact warning string seen (Temperature Abnormal), chassis LED pattern observed, dashboard readings at warning time (Temp1/Temp2 per board, fan RPM, ambient), firmware version, what you have already tried (cite the step numbers from this page), and contact. Saves diagnostic time and your repair cost.