Advanced

This function allows the miner process to automatically restart if the device’s current performance falls below the set threshold.

Description and Settings

The user sets the minimum performance threshold in the range from 0% to 99%, where 0 disables this function. The system compares the actual hashrate with the nominal (theoretical) value.

Theoretical Hashrate Calculation

The system compares actual performance with the theoretical value calculated as:Theoretical Hashrate = (Chip Frequency × Number of Cores) / 1000

Where:

• Chip Frequency — current operating frequency;
• Number of Cores — total number of device computational cores.

If the current hashrate falls below the specified percentage of nominal, the system initiates a miner restart.

Algorithm and Check Cycles

Hashrate checks follow these timing rules:

• Initial Delay: Checks start no earlier than 5 minutes after mining starts, or 10 minutes for quick start (to stabilize readings).
• Exception: The setting is inactive during preset changes or within the first 5 minutes after startup. If less than 3 minutes have passed since start, the check is also skipped. Does not operate during autotune.
• Cycle Frequency: After initial stages, the system checks hashrate against the threshold every 90 seconds.

Limits the number of consecutive system restart attempts when errors occur.

Restart Conditions

System reaction depends on this parameter:

• Limited Attempts: The miner will restart the specified number of times.
• Infinite Loop: If set to 0, restarts are unlimited (checks and restarts continue cyclically until performance is restored).

Temperature limit (recommended range 50–120°C, default 90°C). When reached, mining stops to prevent chip damage.

Overheat Protection Description

Regardless of the value set in this field, if the maximum temperature on the miner exceeds it, the system will immediately shut down. This is the primary overheat protection.

The following rules are used to trigger protection:

• Data Source: The system relies on the temperature of the hottest chip or the hottest sensor (if chip-level measurement is unavailable).

Stops the device if the number of active hash boards drops below the specified number.

Operation Logic and Checks

The system checks the status of each chain (board) regardless of whether it is enabled or disabled in the settings. The check verifies whether all chips on the board are detected.

Example:

• If the miner has 4 chains and this field is set to 4, then if even one chain fails (leaving only 3 operational), the miner will not start.
• If the value is set to 3 in the same scenario, the miner will start and continue operating with the three functional boards.

A scale to set the waiting time (0–300 seconds) before mining starts after device power-on. This allows power stabilization and system initialization.

Randomization Logic and Network Load

The user selects a time range, e.g., 160–200 seconds. The system automatically picks a random value within this range for each device.

This time is used as a pause before initializing all equipment and affects both start mining and resume mining processes.

Operation Principle:

• If a wait time is set, the miner simply waits for this period before initializing the power supply and other components.
• Due to randomization, one miner may have a delay of 160 seconds, another 200 seconds.
• This prevents a sudden peak load on the electrical network, ensuring smooth equipment startup and avoiding voltage spikes.

Accelerated miner initialization without performing some checks.

Description and Effectiveness

Initialization occurs much faster by bypassing all delays. This function allows the miner to start quickly and apply load to the equipment as soon as possible.

Model-specific Notes:

Although the feature is available on all devices, its effectiveness depends on the miner’s hardware characteristics:

• On some models, only one sensor operates during initialization, and the system requires time to determine the required cooling level.
• For example, on the S21 Pro model, the cooling phase is fixed at 120 seconds, so using Quick Start has no visible effect on this device.

How to determine if Quick Start is effective:

To check if Quick Start works on your device, monitor the logs during startup:

• If the line Temperatures 38 - 41 C shows the cooling process with temperature differences (e.g., 6, 5 degrees) and the miner continues initialization — Quick Start will be effective.
• If you see a cooling countdown (120 seconds, 110, 100, etc.), Quick Start is ineffective, as the delay is built into this model’s algorithm.

Monitors even voltage distribution and chip stability within domains.

Operation Logic and Hashrate Recovery

This function checks performance within each voltage domain. If a domain’s hashrate becomes unstable and drops below 50% of nominal (calculated based on current frequency), the system attempts to stabilize the chips for that chain.

Stabilization under changing conditions:

Useful when external operating conditions fluctuate. For example, if the room housing the miners first cools down and then warms up, chips may become unstable. In such cases, they often cannot recover normal operation on their own. The miner performs a forced chip restart to stabilize hashrate if technically possible under current conditions.

Conditions and Limitations

This option has several operational limitations:

• When the function DOES NOT operate:
  • If Autotune is active (Autotune).
  • If the miner has not fully initialized.
  • During preset changes.
  • If the miner is not actively mining.
  • If less than 2 minutes have passed since mining started.
• Check cycle:
  • After stabilization, domain balance checks run every 10 minutes. If a domain’s hashrate drops more than 50%, the function triggers for that chain.

Monitors the integrity of chip chains to prevent improper operation.

Working Principle

The system periodically polls all chips on the hashboards, checking their response. If a chip stops responding (e.g., due to failure), the system reacts immediately to protect the device.

System Response

Actions when a faulty chip is detected depend on hardware type:

• Miners with PIC controllers: Only the chain (board) with the unresponsive chip is shut down.
• Miners without PIC controllers: If any chip fails, the entire miner is powered off.

Determines miner behavior on startup after a reboot or power interruption.

Operation Logic

This setting allows the system to either restore or ignore the paused state depending on the user’s choice:

• Option Enabled: If the miner was paused and a power outage or reboot occurs at that moment, upon powering on, the device will automatically resume mining, ignoring the previous pause.
• Option Disabled: The system remembers its last state before shutdown. If the user paused the miner and the power went out, after initialization upon power-on the device will automatically resume the paused state, preserving the previous status.

Automatic switch to a lower performance profile in case of overheating or chain errors.

Trigger Logic and Reasons for Preset Reduction

If the miner overheats on the currently selected profile (board or chip) or a chain break is detected, the system automatically lowers the preset by one level.

Main reasons for this function activation:

• Overheating: Reducing the preset helps decrease overclocking and, consequently, the operating temperatures of the device.
• Chain break: Power supplies have a consumption limit. If exceeded, the unit may go into protection and cut off power. At this moment, chips stop responding to system requests. The system assumes the PSU limit was exceeded and lowers the preset to reduce the load.

Operation Features

• Auto-tune: If the profile has not yet been tuned, the system will automatically start the tuning process for it.

Automatically triggers a retune process when critical chain errors occur.

Operation Logic

If the miner detects a “chain break” error during operation, the system initiates a retune of the current profile.

Main triggers:

• Integrity Violation: Error detected during regular chip chain checks.
• Power Supply Limit: A “chain break” occurred due to PSU power limit being reached, causing chain instability.

With this setting enabled, the system does not merely restart mining; it resets the current overclocking profile and starts the auto-tuning process (if a profile was previously set).

Monitors compliance of supplied voltage with configured parameters.

Scope and Logic

This function operates only on miners equipped with PIC controllers. It activates whenever voltage changes — whether through manual user adjustments, autotune operation, or other system corrections.

Working principle:

• The system measures actual voltage supplied by the PSU to each chain via the PIC controller.
• If the supplied voltage does not match the intended value, the affected chain or the entire miner will be stopped.
• Note: it monitors the voltage from the PSU to the chain, not the internal voltage of the chips themselves.

Hardware Considerations

• Miners with PIC controllers: Allow disabling the problematic chain individually if voltage mismatch is detected.
• Miners without PIC controllers: This function has no effect, as the hardware cannot perform the necessary measurements.

Automatically adjusts voltage based on current chip temperature. The miner analyzes temperature and modifies voltage (up or down) to maintain stable hashrate.

Operation Logic and Adaptation

The system continuously monitors temperature changes, fan speeds, and other parameters.

Example: if the device operated at 60°C and the temperature drops to 50°C, a cooler chip may need more voltage for stable mining. Adaptation occurs continuously.

Behavior by Cooling Mode:

• Automatic Mode (temperature support): If fan duty increases by 4% compared to the previous measurement, the system tries to lower voltage. This helps the miner cool more easily and operate stably as ambient temperature rises.
• Manual Mode or Immersion Cooling: If temperature increases by 3%, voltage is lowered; if it decreases, voltage is raised to maintain stability.

Conditions and Limitations

This function operates only under specific conditions.

The function DOES NOT operate (standby mode):

• Until the miner is fully initialized.
• Until the final working voltage is set (as specified on the “Mining” page in the Global voltage / Volt control field).
• If High Voltage Offset is active (e.g., +1V or +0.8V), it activates only after switching to the main working voltage.
• During preset switching.
• If working frequencies are not set.
• During Autotune operation (Autotune).

In all other standard modes, the system continuously monitors temperature and fan speed changes to adjust parameters.

Fan speed is limited to 50% during device startup or shutdown.

Mode Description

As the name implies, this mode is designed to reduce noise during miner transitions. When enabled, fan speed is capped at 50% during startup and shutdown.

Operational Notes

Due to reduced fan speeds, maintenance timing changes:

• Cooling Time: The period required to cool chips before pause and immediately before mining start is doubled.

This maintains acoustic comfort while ensuring equipment safety through a longer, smoother heat dissipation process.

Allows the device to continue operating even during temporary internet disconnection.

Operating Logic

When this option is enabled, in case of internet disconnection, the miner continues to operate normally and consumes the same amount of power as before the disconnection. All hash rate generated during this period is not credited — the device simply continues the computation process without sending data to the pool.

Main Use Cases

This function is necessary in the following cases:

• Equipment protection at low temperatures: For users operating in extremely low ambient temperatures, stopping the miner may cause excessive cooling and damage the equipment. Offline Mining helps maintain the working temperature of devices.
• Stabilizing power grid load: In case of a mass internet outage at a site, simultaneously stopping multiple miners creates a sudden load drop, and restarting them causes a sudden spike. This option helps keep the load stable and avoids fluctuations.

Intelligent frequency balancing within rows to stabilize temperature without losing hash rate.

Working Principle

The system tries to detect chips with poor heat dissipation (heating more than others) and balances temperatures by redistributing frequencies.

Main Advantages

• Profit retention: The total row hash rate does not drop due to overclocking “cold” chips.
• Targeted protection: Prevents wear of specific chips by avoiding their overheating relative to neighbors.
• Stability: Allows the miner to operate at high presets even with minor cooling defects (uneven thermal paste layer).

This function allows ignoring faulty temperature sensors if it does not threaten the safety of the device.

Sensor Logic and Priorities

The system divides sensors into two groups depending on their location and importance:

• Front sensors (at the input): They measure the temperature of incoming cool air. If a front sensor fails, with this option enabled the system will ignore the failure, and the miner will continue operating.
• Rear sensors (at the output): These sensors detect the maximum chip temperatures and signal possible overheating. Rear sensor integrity is critical for equipment safety.

Function Limitation

Even if the Skip broken temperature sensors option is enabled, it only applies to front sensors. If at least one rear (hotter) sensor fails, the miner will immediately shut down to prevent chip damage.

This option is relevant for models where the system can read each chip's temperature individually. These include:

• L9, S21, T21 and other devices based on similar chips.

Operating Logic

During polling, chip sensors may return clearly erroneous values (e.g., abnormally high numbers or sudden jumps of 50 degrees or more), indicating incorrect sensor readings.

• With the option enabled: The system ignores such incorrect readings from individual chips when making protection decisions.
• Temperature control: Temperature monitoring and overheating protection continue to operate based on general board sensors located outside the chips.

This function allows forcibly disabling individual hash boards on miners that do not have PIC controllers for power management.

Why This Option Is in the “Danger Zone”

On miners without PIC controllers, the system cannot physically cut power to each board (chain) individually via the power controller. The only way to programmatically “disable” such a board is to hold it in a constant reset state.

Risks of Use:

• When this function is enabled, a Reset signal is constantly applied to the board or pin while other chains continue operating normally.
• According to Bitmain technical recommendations, this method is extremely dangerous: prolonged reset voltage on a board under critical conditions may burn out components.

Application

If this option is enabled, the user can disable boards on devices without PIC controllers, but this is done entirely at the owner’s risk due to high risk of physical damage.

This function temporarily increases voltage (offset) during chip initialization at device startup.

Purpose of the Function

During miner initialization, certain chip models require higher voltage for correct core startup. Without using the offset, cores may fail to start in normal operating mode.

Configuration and Safety Features

This setting is provided as a separate regulator because constant high voltage operation may be unsafe:

• Overheat protection: High voltage increases chip temperatures. Using the offset only at startup prevents overall miner overheating during operation.
• Power supply limits: At high ambient temperatures, high voltage may exceed the PSU consumption limit, causing it to enter protection and shut down.
• When overclocking the miner above stock parameters, it is recommended to disable this function (set this setting to 0).

This parameter sets the criterion by which the autotuner classifies chips as “bad” and excludes them from full optimization.

Autotune Logic

In the first tuning stage, the system analyzes the performance of each individual chip. If a chip produces a hash rate below the set threshold at the given frequencies and voltages, it is considered “bad.”

Consequences for “bad” chips:

• The system does not spend time tuning them.
• Frequencies for such chips are automatically lowered so they can continue operating in a safe mode without affecting overall chain stability.

Threshold Configuration

The selected value directly affects the depth of device optimization:

• High threshold (e.g., 80%): The system will be more demanding on chips. The higher the threshold, the more chips the autotuner will attempt to bring to nominal hash rate during testing.
• Low threshold: The lower the percentage, the fewer chips the system will try to bring to nominal power, leaving more chips at reduced frequencies.