Whatsminer Immersion – Dielectric Fluid Temperature High

Pull the trendline. Read chip_temp_max from BTMiner and bulk fluid temp from the BMS for the last 30-90 days. A 5 °C+ climb over weeks at constant hash load means creep — this fault tree applies. A sudden jump points at an acute fault (pump failure, fan stall, leak) — different page, check pump-failure and water-cooling-leak siblings instead. Confirms what kind of fight you're in before you start cleaning anything.

Walk the dry cooler with a flashlight. Inspect the fin pack from 10 cm. Look at every fan for visible rotation. Check intake clearance for boxes, snow drifts, pollen, cottonwood, biological growth. Most diagnostic value on this entire page comes from this 10-minute walk. Free, fast, often the fix.

Clear gross obstructions. Cardboard leaning on the intake. Snow against the side panel. Pollen accumulation on the grilles. Cottonwood drifts in late spring. Insect debris. Anything that visibly blocks airflow. Document with photos before you clear so you have a baseline for the next inspection.

Read the BTMiner log. SSH or MinerTool to each miner, pull /var/log/btminer.log, and grep for temp. Confirm the alarm is the chassis temperature trip and not a hashboard or PSU fault flagged with the same wording. Cross-reference the timestamp against the fluid-temp trendline from Step 1 to confirm correlation.

Baseline the four numbers. With the tank at steady-state hash load, measure: tank bulk fluid temperature (mid-tank thermocouple, away from return jets), HX inlet fluid temp, HX outlet fluid temp, ambient air at the HX intake face. Compute HX fluid ΔT and approach temperature (HX outlet fluid minus HX inlet air). Healthy: bulk fluid ≤ 45 °C, fluid ΔT 8-12 °C, approach 5-15 °C. Write the numbers on the side of the HX with a paint marker — that's your permanent baseline.

Power off the dry cooler. Fin-comb or fin-comb-and-brush the air side. Soft-bristle brush from the intake side; vacuum from the discharge side; fin comb (matched TPI) to straighten flattened fins. For grease / oil films, alkaline coil cleaner (Nu-Calgon Evap Foam or equivalent HVAC product) plus low-pressure water rinse. Patience over pressure — high-pressure rinses bend fins and make the next cleaning harder.

Audit every fan. Visual rotation, RPM with an IR tach (target: matches commanded duty within 5 %), bearing health by sound. Fans grinding, slow, or stalled get replaced. Most axial-fan motors on dry coolers are commodity ECM units, replaceable in 30-60 minutes with hand tools. A stalled fan reporting 100 % commanded duty contributes zero CFM — three of eight stalled fans = 62 % of design airflow on the cooler.

Install permanent thermocouples at HX fluid-in, fluid-out, and ambient intake if they're not already there. K-type probes plus a CAD 50 datalogger gives you every diagnostic on this page going forward. Anyone running an immersion loop in production should have these from day one. The day you stop having to pull a multimeter to read HX delta-T is the day you start catching creep two months earlier.

Verify intake clearance and check for exhaust recirculation. Hand-anemometer at the intake face, then compare intake air temperature to outdoor ambient. If intake is more than 5 °C above outdoor, you have recirculation — fan exhaust looping back into the intake face. This is the single most common Mining Hacker DIY error on home immersion builds and it has more leverage on bulk fluid temp than any other intervention.

Check the pump. Listen for cavitation (high-pitched buzz or rattling at the pump body). Touch-test the pump body — slightly warm is fine, hot to the touch is not. Inspect the suction strainer for debris and clear it. Check the mechanical seal for visible weep on the shaft. A pump quietly losing flow is the second-most-common cause behind air-side fouling, and BTMiner won't tell you about it directly.

Re-baseline after every change. Single-variable changes only — clean fins, wait 4 hours, record the four numbers from Step 5; replace a fan, wait 4 hours, record again. Without disciplined re-baselining you cannot tell which intervention bought the cooling, and you'll waste a Saturday cleaning fins on a HX that wasn't the problem.

Clamp-on ultrasonic flow meter on the pump discharge. Confirm flow is at nameplate at the current head pressure. Pump impeller wear and seal weep are silent until they aren't — flow data is the truth the BMS does not show. Borrow or rent the meter if you don't own one; CAD 400-900 to buy and pays back the first time it tells you the pump is at 60 % flow.

Pull and inspect the pump. Impeller wear pattern (cavitation grooves on the eye, leading-edge erosion), mechanical seal face condition (scoring, cracks, weep tracks), volute scoring, bearing play. Rebuild kit (impeller, seal, gasket set) is 120-550 CAD; full pump replacement is 400-2200 CAD depending on flow rating and chemical compatibility.

CIP the plate-HX water side on fluid-to-water builds. Circulating descaler chemistry — citric acid 5 % solution or a commercial CIP product — through the water side at 40-50 °C for 2-4 hours, then neutralize and flush. Document HX delta-T before and after to confirm scale removal worked. If delta-T does not recover, the plates need replacement, not cleaning.

Add evaporative pre-cooling on the dry cooler. Wetted-pad system or atomized-spray array upstream of the fan intake drops effective intake temperature 5-10 °C on dry days. Effectiveness collapses in humid weather, but on dry summer afternoons it pulls bulk fluid temp down 3-7 °C for the cost of water and a CAD 800-2500 retrofit. Cheapest first move on Canadian summer-creep installs.

Re-engineer recirculation. Through-wall mount the dry cooler so intake faces outdoor air. Add an exhaust duct that vents outside. Add intake baffles or a Z-cabinet plenum to break the path between fan exhaust and fan intake. Required clearances vary by HX vendor — 1-2 m of free airflow above and around a vertical-discharge dry cooler is the typical minimum. Plumbing-shop work, not HX work, but high leverage on bulk fluid.

Upsize the dry cooler. Steps 1-16 done, ambient genuinely above design, HX is structurally undersized for current load and climate. Next-larger-frame dry cooler with +25-50 % heat reject capacity is the structural fix. Budget CAD 3500-9500 for the cooler plus CAD 1000-2500 for install labour. Sized against the HX vendor's published heat-rejection curve at peak ambient — not a directional rule of thumb.

Add a second HX in parallel. When a single dry cooler can't be upsized further on a constrained site, paralleling a second cooler distributes load and adds redundancy. Plumb in matched check valves, balance the flow split with throttling valves, document the split percentages so the next operator can re-balance after maintenance.

BTMiner power_percent derate as a stopgap. While the structural fix is on order — HX upsize shipping, fan motor on backorder, summer heatwave running long — drop BTMiner's power_percent to 90 % or 85 % to reduce heat output and keep the rig hashing without crossing the trip. Document the hashrate hit, document the kWh saved, set a calendar reminder to undo the derate once the structural fix lands. Stopgap, not a fix.

Stop DIY when: cleaning the fins and verifying flow has not closed the gap in 48-72 hours, plate-HX water side needs CIP and you don't have the equipment or chemistry experience, multiple top-bank hashboards show thermal-stress signatures (discoloured BMC chips, lifted BGA, scorched solder mask), or you've cleaned the same HX three times in 12 months and creep is recurring. Book D-Central immersion service for loop teardown, fluid lab analysis, pump rebuild, plate-HX CIP or replacement, and dry-cooler retrofit recommendation with a 48-hour post-recommissioning baseline burn-in so you ship home with documented healthy numbers.