Perpetual Tune is an automated, continuous optimization feature built into UMC OS — ePIC Blockchain’s custom Antminer firmware — that monitors and adjusts voltage and clock frequency across a mining fleet in real time, without constant operator intervention. Rather than requiring a one-time calibration pass, it runs as an ongoing process, adapting to shifting conditions as hardware ages, ambient temperature changes, and power availability fluctuates. For enterprise operators managing large fleets of Antminer S19x through S21x units, it is the difference between a fleet that requires constant manual attention and one that manages its own efficiency.
Why Manual Tuning Does Not Scale
Every ASIC miner in a Bitcoin mining fleet behaves slightly differently. Chip variation means chips on the same hashboard operate at different efficiency points. Environmental variation means miners in different rows of a data hall experience different temperatures. Hardware aging means a miner deployed 14 months ago performs differently than one deployed last month, and differently again than it did at deployment.
Operators running stock Bitmain firmware have no mechanism to account for any of this. Stock firmware applies fixed preset modes — low, normal, or high power — and leaves every unit performing at the same conservative settings regardless of its actual hardware condition.
Custom firmware addresses this through tuning. UMC OS includes three calibration algorithms — Voltage Optimizer, Board Tune, and Chip Tune — that operators can run to set a miner’s initial optimized configuration. The difference between these and Perpetual Tune is important to understand. A one-time tuning pass is a snapshot: it finds a stable operating point at the moment it runs and sets the miner to those parameters. That snapshot becomes less accurate over time as conditions change.
Perpetual Tune is not a snapshot. It is an ongoing process that continuously re-evaluates and adjusts operating parameters as conditions change, keeping every miner in the fleet working at its current optimal point.
What Perpetual Tune Actually Does
Perpetual Tune operates as a closed-loop control system. It continuously reads performance data — temperature, hashrate output, and voltage response — and uses this data to make incremental adjustments to operating frequency and voltage across each miner. The adjustments are not dramatic or disruptive. They are small, ongoing corrections that keep each unit operating at the intersection of its performance target and its current efficiency threshold.
There are several hardware and environmental changes that Perpetual Tune responds to in real time:
Thermal variation.
The front of a miner facing the cold aisle operates at lower temperatures than the back of the miner facing the hot aisle of a data hall. Perpetual Tune adjusts chip frequency upward in cooler-running side, where chips can sustain higher clocks without thermal risk, and pulls back in hotter-running side, preventing thermal throttling before it occurs. This means every miner in a diverse thermal environment is operating at its actual capacity rather than the same conservative setting applied uniformly.
Load fluctuation.
At peak network difficulty, maximizing hashrate per watt is the priority. During periods of lower hashprice, efficiency per watt matters more. Perpetual Tune can be configured to track a hashrate target rather than a fixed power draw, which means it naturally shifts toward efficiency-focused operation when hash conditions are less favorable — without the operator needing to manually change settings across the fleet.
Hardware aging.
Over time, mining rigs can experience reduced performance and efficiency due to normal wear and tear, electrical stress, fan degradation, thermal paste aging, and prolonged exposure to heat and operating loads.Perpetual Tune detects this drift and adjusts frequency and voltage to compensate, partially recovering efficiency that would otherwise be permanently lost with static firmware settings.
Post-maintenance recalibration.
After a miner is serviced, cleaned, or has a component replaced, its operating characteristics change. Perpetual Tune picks up the new baseline automatically on restart, without requiring the operator to schedule a new calibration run.
Voltage Optimizer, Board Tune, and Chip Tune: When to Use Each
The three one-time tuning algorithms in UMC OS and Perpetual Tune are not competing approaches — they serve different purposes.
Voltage Optimizer
is the right starting point for healthy miners that have never been tuned. It runs a fast voltage sweep, finds the optimal voltage for the current clock settings, and applies it. The process takes under 30 minutes and is appropriate for large fleets being configured for the first time or after a firmware update.
Board Tune
goes a step further, independently adjusting clock speeds for each hashboard while simultaneously optimizing voltage. It is appropriate for fleets where miners exhibit uneven performance across their three hashboards — a common occurrence after months of operation or in facilities with inconsistent thermal conditions.
Chip Tune
is the most thorough one-time calibration, adjusting settings at the individual chip level. It is the right choice for aging hardware, refurbished units, or any miner showing significant hashrate instability. It takes up to 60 minutes to complete.
Once initial calibration is complete with one of these algorithms, Perpetual Tune takes over as the continuous maintenance layer. Its job is to prevent the calibration from going stale.
| Voltage Optimizer | Board Tune | Chip Tune | |
| Tuning depth | Voltage only | Per-board clock + voltage | Per-chip clock + voltage |
| Completion time | Under 30 min | ~45 min | ~60 min |
| Best for | Fast initial setup | Uneven boards | Aging/repaired hardware |
What Perpetual Tune Means for a Fleet at Scale
The operational value of Perpetual Tune compounds with fleet size. A single miner running at 2% below its optimal efficiency is a marginal issue. A fleet of 500 miners running at 2% below optimal is a meaningful and sustained profit leak.
Consider a fleet of 200 Antminer S19j Pro units where Perpetual Tune maintains an average efficiency improvement of 8% over what the fleet would achieve with static settings six months post-deployment. At an average power draw of 3,050W per unit and an electricity rate of $0.07/kWh, the fleet’s daily power cost is approximately $1,025. An 8% efficiency improvement reduces effective power cost by roughly $82 per day — approximately $30,000 per year — with no hardware change and no operator input required.
The other dimension is uptime. Manual tuning requires scheduled downtime for calibration runs, and fleets that go untuned for months accumulate efficiency losses that are never recovered. Perpetual Tune eliminates the need for periodic recalibration windows, which matters for large operations where taking miners offline for maintenance across the whole fleet is logistically difficult.
Perpetual Tune and Curtailment Operations
One important application of Perpetual Tune is its value in curtailment-managed facilities. Enterprise miners in markets like ERCOT or Alberta may participate in demand response programs, where they need to reduce load quickly and predictably in exchange for power credits.
In a curtailment event, the goal is not always to shut machines off. Often, the facility needs the fleet to reduce performance uniformly to meet a specific site-level power target. Static settings can be imprecise because individual miners respond differently depending on chip quality, temperature, board condition, and operating environment.
Perpetual Tune helps manage this automatically. When a curtailment target is applied, the system adjusts miner performance across the fleet toward the required operating level while continuing to optimize each machine’s efficiency. This allows operators to meet curtailment requirements with less manual intervention and more consistent fleet behavior.
Thermal Protection Built Into Every Tuning Mode
A common concern with continuous automated optimization is thermal risk — if the firmware is pushing chips harder in response to favorable conditions, what prevents it from pushing too hard?
UMC OS includes automatic thermal protection across all tuning modes. When the miner chip temperature approaches 2°C below the configured shutdown threshold, the firmware throttles performance automatically. Once temperatures drop 7°C below the shutdown threshold, performance is gradually restored. This protection applies during Perpetual Tune and cannot be disabled, which means the system’s pursuit of efficiency is always bounded by a thermal ceiling that protects hardware longevity.
For operators focused on extending the service life of their hardware investment — particularly relevant for S19x fleets where replacement costs are significant — this thermal ceiling is an important part of the equation.
Getting Started with Perpetual Tune on UMC OS
Perpetual Tune is available on UMC OS for compatible Antminer hardware, including the S19j and newer S19x and S21x variants. Enabling it requires UMC OS to be installed and running on the target hardware. For operators deploying across a large fleet, RigRunner — ePIC’s network-based firmware deployment tool — handles the installation process without requiring command-line access.
The recommended deployment sequence for a new fleet is:
- Install UMC OS using RigRunner across all target miners
- Run Voltage Optimizer on newly deployed, healthy hardware for initial calibration
- Enable Perpetual Tune for ongoing continuous optimization
- For older or inconsistent hardware, run Board Tune or Chip Tune before enabling Perpetual Tune
We have support documentation for each step available, and the full list of supported Antminer variants is maintained in the UMC OS GitHub repository.