Miner Over-Temperature — Troubleshooting Guide
Diagnose and resolve over-temperature warnings, thermal shutdowns, and hashrate throttling on Antminer, Whatsminer, and Avalon ASIC miners.
Symptoms
You are experiencing one or more of the following:
- Web UI displays "Temp too high" or "Over temperature protection" error
- Miner shuts down automatically after running for some time
- Hashrate is lower than expected due to thermal throttling
- ASIC chip temperatures exceed 90-100C in the status page
- Board inlet/outlet temperature readings are abnormally high
- Fans running at 100% speed constantly
- Error log shows "WARN: temp is too high" or "ERROR: chain temp over limit"
Safety Warning: Over-temperature conditions can cause permanent damage to ASIC chips and in extreme cases create fire hazards. Do not disable thermal protection features. Address the root cause rather than masking the symptom.
Normal Operating Temperature Ranges
Understanding what is normal for your miner model helps determine whether you actually have a problem:
Bitmain Antminer
| Model | Chip Temp (Normal) | Chip Temp (Warning) | Chip Temp (Shutdown) | PCB Temp (Normal) |
|---|---|---|---|---|
| S9 | 55-75C | 80C | 95C | 40-60C |
| S17 / S17 Pro | 55-75C | 85C | 100C | 40-60C |
| S19 / S19 Pro | 55-80C | 90C | 105C | 40-65C |
| S19 XP | 55-80C | 90C | 100C | 40-65C |
| S21 | 55-85C | 95C | 110C | 45-70C |
| S21 Pro | 60-90C | 100C | 115C | 45-70C |
MicroBT Whatsminer
| Model | Chip Temp (Normal) | Chip Temp (Warning) | Chip Temp (Shutdown) |
|---|---|---|---|
| M30S / M30S++ | 55-75C | 85C | 95C |
| M50 / M50S | 55-80C | 90C | 100C |
| M60 / M60S | 60-85C | 95C | 105C |
Canaan Avalon
| Model | Chip Temp (Normal) | Chip Temp (Warning) | Chip Temp (Shutdown) |
|---|---|---|---|
| A1246 | 55-75C | 85C | 95C |
| A1366 | 55-80C | 90C | 100C |
| A1466 | 60-85C | 95C | 105C |
Modern miners like the S21 series operate at higher temperatures by design due to improved chip packaging and thermal management. A chip temperature of 85C on an S21 is perfectly normal, whereas the same temperature on an S9 would be concerning.
Quick Checks
- What is the ambient temperature? Measure the air temperature where the miner intake pulls air. Above 35C (95F) is problematic for most miners.
- Are all fans spinning? Listen and look -- a single failed fan can cause overheating. Check the web UI for fan RPM readings.
- Is airflow obstructed? Check that nothing is blocking the miner's intake or exhaust. Miners need clear, unobstructed airflow.
- When was the miner last cleaned? Dust buildup on heatsinks and fans is the most common cause of gradual temperature increases.
- Has anything changed recently? New firmware with higher clocks, new miner placement, building HVAC changes, or seasonal temperature increases.
Diagnostic Flowchart
Over-temperature warning
│
├─ Check ambient temperature
│ ├─ Above 35C (95F) → Improve room cooling first
│ └─ Below 35C → Continue diagnosis
│
├─ Check all fans
│ ├─ Any fan stopped or slow? → See "Fan Errors" guide
│ └─ All fans running normally → Continue
│
├─ Check airflow
│ ├─ Intake or exhaust blocked? → Clear obstruction
│ ├─ Miners facing each other (hot air recirculation)? → Rearrange
│ └─ Airflow is clear → Continue
│
├─ Check dust buildup
│ ├─ Heatsinks clogged with dust? → Clean the miner
│ └─ Clean → Continue
│
├─ Is only ONE board overheating?
│ ├─ YES → Likely degraded thermal paste or failed temp sensor
│ │ ├─ Compare inlet/outlet temps across boards
│ │ └─ One board significantly higher → Thermal paste issue
│ └─ NO (all boards high) → Environmental or system issue
│
├─ Is the miner overclocked?
│ ├─ YES → Reduce to stock settings and retest
│ └─ NO → Possible thermal paste degradation or sensor fault
│
└─ Temperature reading seems impossibly high (>130C)?
└─ Likely a failed temperature sensor (false reading)Causes
1. Blocked Airflow or Dust Buildup
Probability: Very High
The most common cause of overheating. ASIC miners push enormous volumes of air through the chassis. Any restriction reduces cooling effectiveness dramatically.
Symptoms specific to this cause:
- Gradual temperature increase over weeks or months
- All boards show elevated temperatures, not just one
- Fans running at maximum speed but temperatures still high
- Visible dust on fan blades or heatsinks
Diagnosis:
Inspect Intake and Exhaust
Check both ends of the miner. The intake side should have clear, cool air access. The exhaust side should have unrestricted space for hot air to escape. Common obstructions:
- Miners placed too close together or against a wall
- Cables draped over intake or exhaust
- Nearby equipment blowing hot exhaust into the miner's intake
- Shelving or racks blocking airflow
Inspect Internal Heatsinks
Power off the miner and remove the top cover (or fan shroud). Look at the heatsink fins on the hash boards. Compressed dust between fins acts as insulation, preventing heat transfer to the air.
Antminer S19/S21 series use aluminum heatsinks bonded to the hash boards. Dust tends to accumulate between the fins and on the leading edge (intake side). The heatsink-to-chip interface uses thermal paste or thermal pads.
Whatsminer models use bolt-on heatsinks. Dust collects between fins and around the fan mounting area. The fan shroud design on M30/M50 series makes internal dust harder to see without disassembly.
Avalon miners use a modular design with individual heatsinks per module. Remove each module to inspect for dust buildup between the module and the chassis airflow channel.
Clean the Miner
Use compressed air (60-100 PSI with a nozzle, or a datavac blower) to blow dust out of the heatsink fins and fan blades. Blow from the exhaust side toward the intake to push dust out the way it came in.
Do not use a household vacuum cleaner near circuit boards -- it generates static electricity. Use compressed air or an ESD-safe electronics blower. Hold fan blades still while blowing air through them to prevent over-spinning and damaging the bearings.
Fix: Clean the miner thoroughly and ensure proper airflow in your installation. Establish a regular cleaning schedule based on your environment -- dusty environments may need monthly cleaning, clean server rooms may only need quarterly.
Prevention: Install intake air filters if your environment is dusty. Maintain positive pressure in your mining room. Keep miners spaced at least 6 inches apart. Ensure hot exhaust air is directed away from other miners' intakes.
2. Failed Fan
Probability: High
If one or more fans fail, the remaining fans cannot compensate for the lost airflow. The miner will overheat quickly, even in a cool environment.
Symptoms specific to this cause:
- One fan shows 0 RPM or very low RPM in the web UI
- Temperature rises rapidly (within minutes of powering on)
- Audible difference in fan noise (quieter or rattling)
- Miner may show "Fan speed error" alongside the temperature warning
Diagnosis:
Check the fan RPM readings in the web UI. Each fan should report a speed within the normal range for your model (typically 4,000-7,000 RPM). A fan reporting 0 RPM is either disconnected, jammed, or has a dead motor.
Fix: Replace the failed fan. See Fan Errors for detailed fan troubleshooting and replacement procedures.
Prevention: Clean fans regularly. Fans are wear items -- budget for replacement every 1-2 years in continuous operation.
3. Degraded Thermal Paste
Probability: Medium
Thermal paste (or thermal pads) transfers heat from ASIC chips to the heatsink. Over time, thermal paste dries out and loses effectiveness, especially at sustained high temperatures. This is a gradual failure that worsens over months.
Symptoms specific to this cause:
- One specific board runs hotter than the other boards despite identical airflow
- Temperatures have increased gradually over many months
- Cleaning the miner did not resolve the temperature increase
- The miner is more than 12-18 months old and has never had thermal paste replaced
Diagnosis:
Compare Board Temperatures
Check the web UI for per-board temperatures. If one board is consistently 10-15C hotter than the others, and they are in the same airflow path, that board likely has degraded thermal paste.
Visual Inspection
Remove the heatsink from the suspect board (this requires removing the board from the miner and unbolting the heatsink). Inspect the thermal paste:
- Dried, cracked, or powdery paste = degraded, needs replacement
- Wet, uniform coverage = paste is still good, look elsewhere for the cause
Fix:
Remove Old Thermal Paste
Clean all old thermal paste from both the chip surfaces and the heatsink using isopropyl alcohol (99%) and lint-free wipes. Ensure no residue remains.
Apply Fresh Thermal Paste
Apply a thin, even layer of quality thermal paste (Arctic MX-6, Thermal Grizzly Kryonaut, or equivalent) to each chip surface. For hash boards with many chips, a stencil application method ensures consistent coverage.
Do not use too much thermal paste. A thin, even layer (0.1-0.2mm) provides the best heat transfer. Excess paste actually insulates rather than conducts, and can overflow onto surrounding components.
Reassemble and Test
Reinstall the heatsink with proper torque on the mounting bolts (follow a cross pattern to ensure even pressure). Reinstall the board and power on the miner. Monitor temperatures over the next hour to verify improvement.
Prevention: Use high-quality thermal paste. Plan for thermal paste replacement every 12-24 months in continuous operation. Some operators use thermal pads (like Fujipoly or Laird) for longer service life, though pads typically have slightly lower thermal conductivity than paste.
4. Ambient Temperature Too High
Probability: Medium
ASIC miners are designed to operate within a specific ambient temperature range. Most miners are rated for up to 35-40C ambient. Above this, even a perfectly maintained miner will overheat.
Symptoms specific to this cause:
- All miners in the same room are showing elevated temperatures
- Problem is seasonal (worse in summer)
- Problem is time-of-day dependent (worse in afternoon)
- Room feels noticeably hot
Diagnosis:
Measure the intake air temperature with a separate thermometer (do not rely on the miner's board temp sensor for ambient). If intake air is above 35C, the environment needs cooling before the miner can operate normally.
Fix: Improve room cooling. Options include:
- Increase HVAC capacity
- Add exhaust fans to remove hot air from the room
- Use evaporative cooling (in dry climates)
- Reduce miner density (fewer miners per room)
- Immersion cooling (for large-scale operations)
- Reduce miner power/frequency to lower heat output
Prevention: Plan your cooling capacity for the hottest expected ambient temperature plus a 5C margin. Monitor room temperature with independent sensors.
5. Overclocking
Probability: Medium
Custom firmware or factory overclock profiles increase chip voltage and frequency, which increases power consumption and heat output. If cooling was adequate at stock settings, overclocking can push it past the thermal limit.
Symptoms specific to this cause:
- Problem started after applying overclocked firmware or increasing power limit
- Miner draws more watts than the stock specification
- Hashrate was higher than stock before thermal throttling began
Diagnosis:
Check your firmware and power settings:
- Are you running stock firmware or custom firmware (Braiins, VNish, etc.)?
- What power profile or frequency is configured?
- Compare the miner's actual power draw (from the PSU or wall meter) to the stock specification
Fix: Reduce the frequency, voltage, or power limit to a level where temperatures stay within safe ranges. On custom firmware like Braiins OS+, use the auto-tuning feature which automatically adjusts for temperature. On stock firmware, restore factory default settings.
Prevention: When overclocking, increase settings gradually and monitor temperatures at each step. Do not exceed the cooling system's capacity. Ensure the PSU can handle the increased power draw.
6. Broken Temperature Sensor (False Reading)
Probability: Low
Sometimes the temperature sensor itself fails, reporting impossibly high or low temperatures. The miner's protection logic reacts to the sensor reading, not the actual temperature, so a faulty sensor can cause unnecessary shutdowns.
Symptoms specific to this cause:
- One board reports a temperature far higher than the others (e.g., 130C+ while others show 70C)
- The board does not feel unusually hot to a brief touch (power off first!)
- The temperature reading is stuck at a fixed value or jumps erratically
- Only a single sensor shows the anomalous reading, not the whole board
Diagnosis:
Compare Sensor Readings
Most hash boards have multiple temperature sensors (inlet and outlet, or per-domain sensors). Compare all readings for the suspect board. A single anomalous sensor while others read normally suggests a sensor fault.
Physical Temperature Check
Power off the miner, wait 10 seconds, and carefully feel the heatsink over the suspect area. If the reported temperature was 120C but the heatsink feels merely warm (not burningly hot), the sensor is likely faulty.
If the miner reported extremely high temperatures, the heatsink may genuinely be hot enough to cause burns. Touch briefly with the back of your hand first to assess before making full contact.
Verify with External Measurement
Use an infrared thermometer or thermal camera pointed at the heatsink area near the suspect sensor to get an independent temperature measurement.
Fix: Temperature sensors (typically NTC thermistors or digital sensors like LM75 / TMP451) are soldered to the hash board. Replacing them requires desoldering the failed sensor and soldering a replacement of the same type and value. This is a relatively simple repair for someone with basic soldering skills. See Temperature Sensors for component details.
Prevention: Temperature sensor failures are uncommon and generally caused by electrical events or physical damage. Normal operation does not degrade sensors.
When to Seek Professional Help
Contact a qualified repair technician if:
- You smell burning or see discoloration on the hash board
- The miner overheats even with confirmed good airflow, clean heatsinks, and ambient below 30C
- Thermal paste replacement did not resolve the issue
- You suspect a voltage regulator is overheating (not the ASIC chips)
- Multiple boards show wildly different temperatures despite identical conditions
Frequently Asked Questions
Q: Can I run my miner with one fan removed temporarily?
No. Removing a fan destroys the airflow balance. The remaining fans cannot compensate, and the miner will overheat within minutes. Even a brief test without all fans risks permanent chip damage.
Q: Is it safe to run my miner in a garage during summer?
It depends on your climate. If garage temperatures exceed 35-40C, you need additional cooling. Many operators use exhaust fans to pull hot air out and intake vents for fresh air. Monitor temperatures closely during hot spells.
Q: How often should I replace thermal paste?
For continuous operation, every 12-24 months is a reasonable interval. If your miner consistently runs at higher temperatures, inspect the paste at 12 months. Lower-temperature environments can extend this to 24 months or more.
Q: My miner shows different temperatures for "PCB temp" and "chip temp." Which matters?
The chip temperature is the critical number -- this is what thermal protection uses for shutdown decisions. PCB temp is the board surface temperature and is always lower. Focus on keeping chip temperatures within the normal range for your model.
Q: Will immersion cooling eliminate over-temperature problems?
Immersion cooling (submerging boards in dielectric fluid) dramatically improves heat dissipation and essentially eliminates airflow-related temperature problems. However, it introduces its own complexity: fluid management, pump maintenance, and specialized enclosures. It is primarily used in large-scale operations.
Related Guides
- Fan Errors -- Fan troubleshooting and replacement
- Missing ASIC Chips -- Chips can fail from thermal damage
- How Hash Boards Work -- Understanding thermal management
- Power Supply Issues -- PSU overheating
- Miner Shows No Hashrate -- Thermal shutdown can cause zero hashrate
Missing ASIC Chips — Troubleshooting Guide
Diagnose and fix missing ASIC chips on your miner. Covers partial hashrate, incorrect chip counts, and specific chip position failures across Antminer, Whatsminer, and Avalon models.
Low Hashrate — Troubleshooting Guide
Diagnose and resolve low hashrate on ASIC miners including Antminer, Whatsminer, and Avalon. Covers missing chips, thermal throttling, pool issues, firmware settings, and hardware degradation.