Control Board vs Hashboard: Understanding ASIC Miner Components

Every ASIC miner — whether it is an Antminer S21, a Whatsminer M60, or an old S9 running as a space heater — is fundamentally built from two types of boards: the control board and the hashboard(s). Understanding what each does, how they communicate, and what happens when they fail is foundational knowledge for anyone who owns, operates, or repairs mining hardware.

This is not abstract theory. When your miner shows an error, knowing whether the problem is on the control board or the hashboard determines whether you are looking at a simple firmware flash or a complex BGA rework. It determines whether you can fix it at home or need to send it to D-Central for professional repair. It determines how much it will cost and how long your miner will be down.

At D-Central Technologies, we have been doing board-level ASIC repair since 2016. This guide distills what we have learned from thousands of repairs into a comprehensive reference for miners who want to understand their machines at a deeper level.

The Two Brains of an ASIC Miner

Think of an ASIC miner as having two distinct systems:

A typical ASIC miner has one control board and three to four hashboards (the S19 series uses three; the S21 uses four). The control board manages everything except the actual hashing, and the hashboards do nothing except hash. Neither can function without the other.

The Control Board: Command Center

What It Does

The control board is a small single-board computer that runs a Linux-based operating system. It is responsible for:

• Network connectivity: The Ethernet port on your miner connects to the control board. It handles DHCP, IP configuration, and all network communication.
• Pool communication: The control board connects to your mining pool using the Stratum protocol, receives work (block header templates), and submits shares (valid hashes found by the hashboards).
• Web interface: The configuration dashboard you access through your browser runs on the control board. Miner status, pool settings, fan controls, firmware updates — all served from this board.
• Hashboard management: The control board initializes each hashboard, distributes mining work to them, collects results, and monitors their health (chip counts, temperatures, error rates).
• Fan control: The control board reads temperature sensors and adjusts fan speeds through PWM (Pulse Width Modulation) signals to maintain optimal operating temperature.
• Firmware storage: The operating system, mining software (cgminer/bmminer), and configuration data are stored on the control board’s NAND flash memory or eMMC.
• Monitoring and protection: Over-temperature shutdown, fan failure detection, voltage anomaly detection, and error logging all happen on the control board.

Key Components on the Control Board

Common Control Board Failures

1. Corrupted Firmware / Bad NAND

Symptoms: Miner does not boot, boot loops, web interface shows errors or garbled text, kernel log shows filesystem errors.

Cause: Power loss during a firmware update is the classic scenario. The firmware write was interrupted mid-way, leaving the NAND in an inconsistent state. Over time, NAND flash cells also wear out from repeated write cycles, especially on older miners that have been reflashed many times.

Diagnosis: Attempt an SD card firmware recovery. If the miner boots and runs normally from the SD card but fails from internal NAND, the NAND is the problem. Check our SD card firmware recovery guide for step-by-step instructions.

Repair: SD card recovery resolves most firmware corruption issues. If the NAND flash chip itself is physically failing (SD card recovery works but the problem returns after flashing back to NAND), the NAND chip needs to be replaced — a professional repair requiring BGA rework.

2. Ethernet Port / Network Failure

Symptoms: No link light on the Ethernet port, miner does not appear on your network, no IP assigned via DHCP, ping fails.

Cause: The Ethernet PHY (Physical Layer) chip can be destroyed by ESD (Electrostatic Discharge) traveling through the Ethernet cable. This is more common in setups where the miner is connected to long cable runs, outdoor runs, or through non-grounded network equipment. Lightning strikes on nearby power or data lines can also kill the Ethernet PHY.

Diagnosis: Try a known-good Ethernet cable. Try a different port on your switch or router. If no link light appears with confirmed-working cable and switch, the PHY chip or the RJ45 connector itself is likely damaged. Inspect the RJ45 port for bent pins or physical damage.

Repair: Ethernet PHY chip replacement is a board-level soldering job requiring hot air rework. The RJ45 connector can also be replaced if physically damaged. Both are professional-level repairs.

3. Fan Controller Failure

Symptoms: All fans report 0 RPM or “fan error” even though the fans are physically spinning. Or fans run at full speed regardless of temperature settings.

Cause: The fan controller IC on the control board reads RPM signals from each fan and sends PWM signals to control speed. If this IC fails, the control board cannot read or control fan speeds. Sometimes only the reading fails (fans spin but report 0 RPM), causing the firmware to trigger a fan error shutdown even though cooling is actually fine.

Diagnosis: Swap a “failing” fan to a different port. If all ports report 0 RPM for any fan connected, the fan controller IC is the likely culprit, not the fans themselves. You can also bypass the fan error check temporarily in some firmware versions to confirm the fans are physically working, but this is not recommended for extended operation.

Repair: Fan controller IC replacement requires SMD soldering equipment and a replacement chip. This is a professional repair.

4. RAM or SoC Failure

Symptoms: Random crashes, boot loops, kernel panics in the log, mining software crashes repeatedly, web interface freezes.

Cause: RAM failures are typically caused by heat stress (the control board sits in the airflow path but generates its own heat), power quality issues, or manufacturing defects. SoC failures are rare but can result from extreme voltage events.

Diagnosis: If an SD card recovery and full firmware reflash does not resolve random crashes and boot loops, suspect hardware failure. Kernel log entries showing “segfault,” “kernel panic,” or memory allocation errors point to RAM issues.

Repair: RAM chip replacement requires professional BGA rework. SoC failure typically means control board replacement, as reflowing or replacing the main processor is not economically practical in most cases.

The Hashboard: The Workhorse

What It Does

The hashboard is where mining actually happens. It is a long PCB (printed circuit board) densely packed with ASIC chips — application-specific integrated circuits designed to do one thing: compute SHA-256 hashes as fast as possible.

When the control board sends a block header template to the hashboard, each ASIC chip on the board iterates through billions of nonce values per second, computing SHA-256(SHA-256(block_header + nonce)) and checking whether the result meets the difficulty target. When a chip finds a valid hash, it sends the result back to the control board, which submits it to the mining pool as a share.

A single hashboard on a modern miner like the Antminer S21 contains approximately 110 BM1688 chips, each contributing a portion of the board’s total hashrate. All chips on a board share a common power rail (stepped down from 12V to approximately 0.3V through voltage regulators) and communicate through a serial daisy-chain.

Key Components on the Hashboard

How Voltage Domains Work

ASIC chips on a hashboard are organized into voltage domains — groups of chips that share a common voltage regulator. For example, a hashboard with 110 chips might have 5 voltage domains of 22 chips each. Each domain has its own buck converter stepping 12V down to the precise voltage the chips require (typically around 0.3V for modern chips).

This is important for diagnostics because a failure in one voltage domain affects all chips in that domain simultaneously. If you see exactly 22 chips missing from a 110-chip hashboard, you almost certainly have a voltage domain failure — either the regulator itself has failed, or a single shorted chip has pulled down the voltage for the entire domain.

Common Hashboard Failures

1. Dead ASIC Chips

Symptoms: Reduced chip count on one or more hashboards. The kernel log shows fewer chips found during initialization than expected. Hashrate is proportionally reduced.

Cause: Individual chips can die from ESD, thermal cycling stress, power surges, or manufacturing defects. A single dead chip can also take out its entire voltage domain if it shorts, making it appear as though many chips are dead when only one is the root cause.

Diagnosis: Compare the number of missing chips to the voltage domain size. If the missing count matches a domain size (or a multiple), suspect a single-point failure (one chip or regulator) taking out the domain. Thermal imaging can reveal a hot chip that has shorted, as it will still draw current and generate heat even though it is not hashing.

Repair: Requires identifying the specific failed chip(s) and replacing them using a BGA rework station. This is professional-level repair. The 5nm BGA packages on modern chips (BM1688) require extremely precise temperature profiles during reflow.

2. Burned Traces / Physical Board Damage

Symptoms: Visible charring or discoloration on the PCB. One or more voltage domains are dead. The board may draw excessive current or blow fuses immediately on power-up.

Cause: A short circuit (from a failed chip, liquid intrusion, or physical damage) can cause enough current to flow through a PCB trace to literally burn it open. High-current traces (the power rails) are most vulnerable. Liquid damage from condensation or coolant leaks is also a common cause — water bridges between traces create short circuits that burn the board.

Diagnosis: Visual inspection is usually sufficient — burned traces are visible as dark brown or black marks on the PCB. Use a multimeter in continuity mode to check whether traces are intact.

Repair: Minor trace damage can sometimes be repaired with jumper wires — a skilled technician solders thin wire to bridge the burned-out trace. Extensive trace damage may render the board unrepairable, depending on how many layers are affected (modern hashboards use multi-layer PCBs).

3. Bad Voltage Regulators

Symptoms: An entire voltage domain of chips is missing from the chip count. The board may draw abnormal current (too high or too low). In some cases, the voltage regulator area runs noticeably hotter or cooler than surrounding areas.

Cause: Voltage regulators endure significant thermal stress — they convert 12V to 0.3V, and the energy difference is dissipated as heat. Over time, the components degrade, especially the MOSFETs and inductors. A regulator can also fail catastrophically from a power surge or from a downstream short (a shorted ASIC chip pulling too much current).

Diagnosis: Use a multimeter to measure the output voltage of each voltage domain. A dead domain will show 0V or a voltage far outside the expected range. Thermal imaging during power-up reveals abnormal heat patterns around the failed regulator.

Repair: Voltage regulator replacement requires identifying the failed component within the regulator circuit (MOSFET, driver IC, inductor, or capacitor) and replacing it. This is a professional repair requiring knowledge of the specific regulator circuit topology used on that hashboard model.

4. Power Connector Damage

Symptoms: Hashboard does not power on, or powers on intermittently. Visible melting, discoloration, or carbon deposits on the 6-pin power connector. Burning smell when the miner runs.

Cause: The 6-pin connectors carry significant current (sometimes 30A+ per connector). If the connector is not fully seated, or if a pin has poor contact, the increased resistance generates heat. Over time, this creates a positive feedback loop: heat increases resistance, which generates more heat, which eventually melts the connector. This is one of the most common and most preventable failures in ASIC mining.

Diagnosis: Visual inspection. Pull the power connector and look at both the male pins (on the cable) and the female receptacle (on the hashboard). Any discoloration, melting, or carbon deposits indicate a damaged connector.

Repair: The board-side connector must be desoldered and replaced with a new one. The cable-side connector should also be replaced. This is a medium-difficulty repair that requires competent soldering skills. If you are comfortable with a soldering iron, this is a feasible DIY repair — but get the right replacement connector (correct pinout and current rating).

How the Control Board and Hashboards Communicate

Understanding the communication between these two systems helps explain many failure modes:

1. Initialization: When the miner powers on, the control board sends an initialization sequence to each hashboard through the data connector (typically SPI or UART). Each ASIC chip on the hashboard responds with its address and status. The control board counts the responding chips and reports the result in the kernel log (e.g., chain[0] find 110 chips).
2. Work distribution: The control board receives block header templates from the mining pool and distributes work to each hashboard. Each hashboard then subdivides the work among its ASIC chips, assigning different nonce ranges to each chip.
3. Result collection: When an ASIC chip finds a hash that meets the share difficulty, it sends the result back through the data chain to the control board. The control board validates the result and submits it to the pool.
4. Health monitoring: The control board periodically polls each hashboard for temperature readings, chip status, and error counts. This data drives fan speed control, thermal protection, and the status information you see in the web interface.

The data connector between control board and hashboard is the critical communication link. It carries all of the above — initialization commands, work data, results, and health data. A single broken wire in this ribbon cable can take an entire hashboard offline.

What Happens When Each Component Fails

When the Control Board Fails

• Complete failure: The entire miner is dead. No network, no web interface, no hashing. The hashboards are physically fine but have no “brain” to operate them.
• Partial failure (network): The miner may be hashing internally but you cannot access it, monitor it, or change settings. Some miners will continue hashing using their last known configuration even if the web interface is down, but you have no visibility or control.
• Partial failure (fan control): Fans may run at full speed (safe but loud) or fail to report speed (triggering shutdown even if cooling is adequate).
• Firmware corruption: The miner may boot-loop, show error messages, or partially initialize. Mining may start and stop erratically.

Key point: A control board failure does not damage the hashboards. The hashboards are passive until instructed by the control board. If you replace or repair the control board, the hashboards should work exactly as before.

When a Hashboard Fails

• Single hashboard failure: The other hashboards continue to operate normally. You lose approximately one-third (3-hashboard miner) or one-quarter (4-hashboard miner) of your total hashrate. The miner remains operational at reduced performance.
• Partial hashboard failure (dead chips): The hashboard operates at reduced hashrate proportional to the number of working chips. If an entire voltage domain is dead, you lose those chips but the remaining domains keep hashing.
• Hashboard short circuit: In rare cases, a severe hashboard fault can trip the PSU’s over-current protection, shutting down the entire miner. This is the PSU protecting itself, not a control board failure.

Key point: A hashboard failure is isolated. It does not affect the control board or other hashboards (except in the rare short-circuit scenario). You can remove a failed hashboard and continue mining with the remaining boards while you arrange for repair.

Repairability Comparison

Diagnostic Decision Tree: Control Board or Hashboard?

When something goes wrong with your miner, use this decision tree to determine where the problem lies:

1. Does the miner power on at all?
   • No → Check PSU first, then control board power circuit
   • Yes → Continue to step 2
2. Does the miner appear on your network?
   • No → Control board issue (Ethernet PHY, firmware, network config)
   • Yes → Continue to step 3
3. Can you access the web interface?
   • No / very slow / errors → Control board firmware or RAM issue
   • Yes → Continue to step 4
4. How many hashboards are detected?
   • Zero → Control board issue (data connectors, SPI/UART bus) OR all hashboards happen to be bad (less likely)
   • Some but not all → The missing board(s) likely have hashboard-level issues (connector, power, chips)
   • All detected → Continue to step 5
5. Are all chips detected on each hashboard?
   • Some missing → Hashboard issue (dead chips, voltage domain failure)
   • All present → Continue to step 6
6. Is hashrate normal and stable?
   • Low or unstable → Could be either: check temperatures (hashboard thermal issue), check error codes (control board or hashboard), check power delivery
   • Normal → Your miner is probably fine

When to Replace vs. Repair

Control Board

Control board replacement is often the most practical option because:

• Control boards are less expensive than hashboards
• They are somewhat standardized within a model series (e.g., S19 control boards work across S19 variants)
• A replacement board gets you back online faster than a board-level repair

Repair makes sense when: the issue is firmware-only (SD card fix), when replacement boards are not available, or when the specific board has custom firmware (Braiins OS+, VNish, LuxOS) that is not easily transferable.

Hashboard

Hashboard repair is usually the better option because:

• Replacement hashboards are expensive (they contain the most valuable components — the ASIC chips)
• Hashboards are batch-specific — not all S21 hashboards are interchangeable between production batches
• Many hashboard failures involve one or a few chips, making repair significantly cheaper than replacement

Replacement makes sense when: the hashboard has extensive physical damage (burned traces across multiple layers), when repair costs approach replacement cost, or when a compatible replacement is available at a good price.

Protecting Both Components

Many failures that affect both control boards and hashboards share common preventable causes:

• Use proper surge protection. A quality surge protector or UPS protects both the control board’s sensitive logic circuitry and the hashboards’ power delivery from voltage spikes.
• Ground your equipment properly. ESD is a primary killer of both Ethernet PHY chips (control board) and ASIC chips (hashboard). Proper electrical grounding reduces ESD risk.
• Maintain clean, stable power. Voltage fluctuations stress voltage regulators on both boards. Overloaded circuits, long extension cords, and generators with dirty power all increase failure risk.
• Control your environment. Keep humidity below 65% to prevent condensation and connector corrosion. Maintain ambient temperature below 35C. Keep dust under control with regular cleaning.
• Handle with care. When swapping hashboards, use an ESD wrist strap. Do not flex hashboards. Seat connectors straight and fully. These machines are precision equipment despite their industrial appearance.

Frequently Asked Questions

When to Call the Professionals

Understanding the difference between control boards and hashboards empowers you to diagnose problems faster, communicate more effectively with repair services, and make informed decisions about whether to repair or replace.

If you have identified the failing component and it is beyond a firmware flash or connector reseat, D-Central Technologies has the equipment, expertise, and parts inventory to handle both control board and hashboard repairs for all major ASIC miner models. See our complete repair process guide for details on what to expect when you send your miner in. We have been doing board-level ASIC repair since 2016 — thousands of boards have crossed our bench.

Submit a repair request →

Related Resources

• D-Central ASIC Repair Services — Professional repair for control boards and hashboards
• Kernel Log Reading Guide — Understand what your miner’s diagnostic logs are telling you
• Antminer Error Code Guide — Decode every Antminer error message
• Whatsminer Error Code Reference — Decode every Whatsminer error message
• Multimeter Guide for ASIC Repair — Essential electrical testing for board-level diagnostics
• SD Card Firmware Recovery Guide — Step-by-step firmware flashing for control board recovery