Introduction
The Canaan AvalonMiner 1066 is a SHA-256 Bitcoin miner from the company that literally invented ASIC mining. Canaan Creative — the team that shipped the Avalon 1, the world’s first commercial Bitcoin ASIC miner, back in January 2013 — has been refining their craft for over a decade. The AvalonMiner 1066, part of the A10 series (also known as the Avalon 10), delivers 50 TH/s of SHA-256 hashrate at 3,250W wall power, yielding an efficiency of approximately 65 J/TH. It is not the most efficient miner ever built, but it is a well-engineered workhorse with an architecture fundamentally different from anything Bitmain ships — and that difference matters when you open it up for maintenance.
If you have spent most of your time inside Antminers, the AvalonMiner 1066 will feel like visiting a different country. Canaan does not use a single integrated control board bolted to the miner chassis. Instead, the Avalon architecture separates the controller from the miner unit. The miner itself is essentially a headless hashing machine — three hashboards, fans, and power distribution — while an external AUC3 (Avalon USB Converter) module and a Raspberry Pi-based controller (or Canaan’s own controller box) handle pool communication, configuration, and management. Multiple miner units can daisy-chain to a single controller. This modular approach is elegant from a decentralization perspective: swap miners in and out without touching your controller configuration.
At the core of each hashboard sit 114 A3205 ASIC chips fabricated on a 16nm process — 342 chips total across three boards. These chips communicate through a serial chain carrying Clock (CK), Clock Transmit (C), Reset (R), and Data (D) signals. The power delivery system uses dual 12V DC connectors per hashboard, each board consuming roughly 1,083W. Four 12038-size fans push 250 CFM of airflow through the chassis, generating a typical noise level of 75 dB.
The AvalonMiner 1066 occupies an interesting position in the mining landscape. It is from the era when 16nm was competitive, sitting alongside machines like the Antminer S9 and T9+. While newer 7nm and 5nm machines have eclipsed it in raw efficiency, the 1066 still finds a place in home mining operations where electricity is cheap, in Bitcoin Space Heater conversions where its 3,250W of heat output is a feature rather than a cost, and in operations where paid-off hardware just needs to keep running. Every hash it produces strengthens the decentralization of the Bitcoin network — and that is what we are here for.
This guide is your complete field manual for maintaining, diagnosing, and repairing the AvalonMiner 1066. We cover everything from routine dust removal to advanced hashboard diagnostics, including the Avalon-specific controller architecture that trips up technicians who have only worked on Bitmain machines. Whether you run a single 1066 heating your garage or a rack of them daisy-chained to one controller, this guide gives you the knowledge to keep your hardware hashing.
D-Central Technologies has been repairing and servicing Canaan AvalonMiners since 2016 at our facility in Laval, Quebec. With 2,500+ miners repaired across all major manufacturers — Bitmain, Canaan, MicroBT, and Innosilicon — we stock replacement hashboards, fans, power supplies, AUC controllers, and every component needed to keep Avalon machines running. The Avalon architecture is less common than Antminers in North America, which means fewer technicians understand it. We do. If anything in this guide goes beyond your comfort zone: 1-855-753-9997.
This guide covers the Canaan AvalonMiner 1066 (A10 series, 50 TH/s). The A10 generation also includes the AvalonMiner 1047 (37 TH/s, lower power) and the AvalonMiner 1046 (36 TH/s). While they share the same A3205 ASIC platform and controller architecture, chip counts, hashboard power draw, and thermal profiles differ between models. The maintenance procedures in this guide apply broadly across the A10 family, but always verify your specific model’s parameters against its official specifications. Earlier Avalon generations (A8, A9) and later ones (A11, A12, A13) use different chip designs and may have different hashboard layouts.
Technical Specifications
Before you touch a single screw, internalize the specs. The AvalonMiner 1066 is architecturally distinct from Bitmain and MicroBT machines in several critical ways: modular controller design, three hashboards instead of four, different power connector types, and a completely different management interface. If you are coming from Antminer experience, leave your assumptions at the door. This is Canaan territory.
AvalonMiner 1066 Hardware Specifications
| Model | Canaan AvalonMiner 1066 (A10 series) |
|---|---|
| Manufacturer | Canaan Creative (Hangzhou, China) |
| Release Period | 2019 |
| Algorithm | SHA-256 (Bitcoin) |
| Hashrate | 50 TH/s (±3%) |
| Power Consumption | 3,250 W (±5% to +8%) |
| Power Efficiency | 65 J/TH (±5%) at 25°C — 63 J/TH achievable at cooler ambient |
| ASIC Chip | Canaan A3205 — 16nm process |
| Hashboards | 3 hashboards |
| Chips per Hashboard | 114 A3205 chips |
| Total Chip Count | 342 A3205 chips (114 × 3) |
| Power per Hashboard | ~1,083 W (3,250W ÷ 3) |
| Power Connectors | Dual 12V DC connectors per hashboard — 6 total for 3 boards |
| Controller | External — AUC3 (Avalon USB Converter) + Raspberry Pi or Canaan controller box with Kendryte K210 AI SoC |
| Communication | CAN bus via AUC3 — supports daisy-chain of multiple miner units to one controller |
| Network Interface | RJ45 Ethernet 10/100 Mbps (on controller, not on miner unit) |
| Cooling | 4 × 12038 fans (2 intake, 2 exhaust) |
| Airflow | 250 CFM |
| Noise Level | 75 dB typical |
| AC Input | 185–280V AC, 50-60 Hz, 16A max — requires C19 plug |
| Operating Temperature | -5°C to +35°C air intake |
| Operating Humidity | 5% to 90% (non-condensing) |
| Dimensions | 331 × 195 × 292 mm |
| Weight | 11.4 kg (miner unit only) |
This is the single most important thing to understand about every Canaan Avalon miner: the controller is external. An Antminer has its control board bolted inside the chassis — one brain, one body. An AvalonMiner is headless. It connects to an AUC3 module via a CAN bus cable, and the AUC3 connects to a Raspberry Pi (or Canaan’s standalone controller) running the Avalon management software (cgminer fork). One controller can manage multiple miner units in a daisy chain. This means network configuration, pool settings, and firmware live on the controller — not on the miner. When troubleshooting, you need to think about both the miner hardware AND the controller as separate diagnostic domains.
A10 Series Variants
The AvalonMiner 1066 is the flagship of the A10 generation. Here is how it compares to its siblings:
A10 Series Comparison
| AvalonMiner 1066 | 50 TH/s @ 3,250W — A3205 — 114 chips/board × 3 boards — Flagship |
|---|---|
| AvalonMiner 1047 | 37 TH/s @ 2,380W — A3205 — Lower chip count — Mid-tier |
| AvalonMiner 1046 | 36 TH/s @ 2,100W — A3205 — Entry-level A10 |
All A10-generation machines share the same fundamental architecture: A3205 chips, three hashboards, external AUC3 controller, and the same power connector scheme. Chip counts and frequencies differ. The maintenance procedures in this guide are directly applicable across the A10 family.
Understanding the Hashboard Architecture
Each AvalonMiner 1066 hashboard contains 114 A3205 ASIC chips organized into groups connected in series, with multiple chips connected in parallel within each group. Understanding the signal architecture is essential for advanced diagnostics:
- Power delivery: Each hashboard receives 12V DC through dual power connectors. The board’s internal voltage regulation steps this down to three rails: Vcore (core ASIC voltage), VTOP, and VDDIO (I/O voltage). Each board consumes approximately 1,083W.
- CK (Clock) signal: The working clock for each chip, operating at 25 MHz. This signal maintains timing for all chip operations and coordinates data processing across the chain. A missing or weak clock signal means zero hashing from affected chips.
- C (Clock Transmit) signal: Operates at 5 MHz and transmits clock synchronization between chips in the chain. This is how the chips stay coordinated as a group.
- R (Reset) signal: The reset pin for each ASIC. Used during initialization and to recover from errors. A stuck reset line prevents chips from entering their hashing state.
- D (Data) signal: Carries mining work data between chips. This is the actual nonce distribution and result reporting path.
All four signals travel in a defined direction through the chip chain. When a chip fails or a solder joint cracks, the signal chain breaks at that point — and every chip downstream of the break goes dark. Knowing the signal direction on your specific board revision tells you exactly where to start probing when chips drop off.
Antminers use flat ribbon cables to connect hashboards to the control board. The AvalonMiner 1066 uses a completely different approach: a CAN bus interface connects the hashboards to the MM (Management Module) board, which then communicates with the external AUC3 controller. CAN bus is the same protocol used in automotive electronics — robust, noise-tolerant, and capable of long cable runs. If the MM board or its CAN bus connector is damaged, the controller loses communication with all hashboards in that miner unit even though the hashboards themselves may be perfectly functional.
Before You Begin
Safety Warnings
The AvalonMiner 1066 operates at 185–280V AC input and draws up to 16A. The PSU converts this to 12V DC at extremely high current — enough to weld metal. ALWAYS disconnect AC power and wait at least 60 seconds for capacitors to discharge before opening the miner or disconnecting any internal cables. Never work on the miner while it is powered. Never touch the AC side of the power supply. If you are not comfortable working with high-voltage electronics, stop and contact a professional.
The A3205 ASIC chips are sensitive to static electricity. A static discharge you cannot even feel (under 3,000V) can permanently damage a chip. Always wear an anti-static wrist strap grounded to the miner chassis when handling hashboards. Work on an ESD-safe surface. Do not wear synthetic fabrics. Do not work on carpet. One careless moment can turn a $200 hashboard into scrap.
Heatsinks and ASIC chips can reach temperatures above 80°C during operation. After powering down, wait at least 10 minutes for the miner to cool before touching internal components. If using a heat gun or soldering iron for repair, be aware of the thermal mass of the heatsinks — they retain heat longer than you expect.
At 75 dB, the AvalonMiner 1066 is loud enough to cause hearing damage with prolonged exposure. Wear hearing protection when working near a running miner. If you are running miners in a home mining setup, consider a Bitcoin Space Heater conversion with sound dampening.
Routine Maintenance
Preventive maintenance is the difference between a miner that runs for years and one that dies in months. The AvalonMiner 1066’s four high-speed fans pull air — and everything in it — through the chassis at 250 CFM. Dust, pet hair, and debris accumulate on heatsinks and fan blades, reducing cooling efficiency and driving up chip temperatures. Higher temperatures mean lower hashrate, higher error rates, and accelerated component wear. A 30-minute maintenance session every 4 to 8 weeks pays for itself many times over in extended hardware life.
Visual Inspection
Start every maintenance session with a thorough visual check. Power down the miner, disconnect AC power, and wait for it to cool. Then inspect:
- External chassis: Look for dents, bent fan grilles, or physical damage that could restrict airflow. Check that all screws are present and tight.
- Fan blades: Inspect both intake and exhaust fans for cracked, chipped, or warped blades. A damaged fan blade creates vibration that accelerates bearing wear and can damage neighboring components.
- Power cables: Check all 12V DC power cables for burns, melted insulation, loose connectors, or discoloration at the connector pins. Canaan’s dual-connector-per-board design means six power cables total — check every one.
- CAN bus cable: Inspect the CAN bus communication cable between the miner and the AUC3 controller for damage, loose connectors, or bent pins.
- Heatsinks: Look for dust accumulation between heatsink fins. Heavy dust buildup acts like insulation, trapping heat against the chips.
- PCB surfaces: With good lighting, check visible PCB areas for corrosion (green or white deposits), burn marks, cracked solder joints, or bulging capacitors.
- Thermal paste residue: If thermal paste has leaked or dried out (appears cracked and chalky), it is time for a repaste.
Cleaning Procedures
Dust is the number one killer of mining hardware. Here is how to do it right:
- Power down completely. Disconnect AC power. Disconnect the CAN bus cable from the AUC3. Wait 10 minutes for cooling.
- Remove the top cover by unscrewing the chassis screws. Set screws aside in an organized manner — Canaan uses different screw sizes in different positions.
- Blow out dust using compressed air or an electric blower. Work from the intake side to the exhaust side, following the natural airflow direction. Hold the can upright (if using compressed air) to avoid spraying propellant liquid onto components.
- Clean fan blades individually. Use a soft brush or compressed air to remove caked-on dust from each blade. Hold the fan hub to prevent it from spinning while blowing — allowing the fan to freewheel at high speed can damage its bearings.
- Clean heatsink fins. Use narrow blasts of compressed air between heatsink fins. If dust is heavily caked, use a soft brush to loosen it first, then blow it out. For stubborn deposits, a cotton swab dampened with 99% IPA works well.
- Inspect and clean power connectors. Use IPA and a lint-free cloth to clean any oxidation or residue from the 12V DC power connector pins. Corroded connectors create resistance, generate heat, and can cause voltage drop that starves the hashboard.
- Clean the MM board area. The Management Module board and its CAN bus connector should be dust-free. Use compressed air gently — the MM board has smaller, more delicate components than the hashboards.
- Reassemble and reconnect. Ensure all power cables are firmly seated before reconnecting AC power.
If you run your AvalonMiner 1066 in a Canadian home during winter (and you should — free heating), be aware that spring thaw brings humidity spikes that can cause condensation inside cold miner chassis. When transitioning from heating season to warmer weather, do a thorough cleaning and inspection. Check for any signs of corrosion from moisture exposure. Our cold climate is an advantage for mining, but the humidity swings between seasons demand attention.
Thermal Paste Replacement
Thermal paste degrades over time. On a miner running 24/7 at high temperatures, you can expect thermal paste to lose effectiveness after 12 to 18 months. Dried, cracked, or chalky thermal paste creates an insulating barrier between the chip and its heatsink, causing chip temperatures to climb even when fans and airflow are adequate.
- Remove the hashboard from the chassis by disconnecting both 12V power connectors and the hashboard-to-MM-board connection.
- Remove heatsinks by carefully detaching the heatsink retention clips or screws. Apply gentle, even pressure — do not pry at an angle, which can crack the chip die. If the heatsink is stuck, apply gentle heat with a heat gun (150°C, 30 seconds) to soften the old paste.
- Clean old paste from both the chip surface and the heatsink base using 99% IPA and lint-free cloths. Ensure both surfaces are perfectly clean and dry before applying new paste.
- Apply new thermal paste using the X-pattern method: draw a thin X across the chip surface. Do not over-apply — a thin, even layer provides better thermal transfer than a thick glob. Non-conductive paste is mandatory (Arctic MX-5, Noctua NT-H2, or Thermal Grizzly Kryonaut).
- Reattach heatsinks firmly. Ensure even pressure across the chip surface. Uneven mounting creates hot spots.
- Repeat for all 114 chips on the board (or at minimum, for any chips showing elevated temperatures).
Some thermal pastes (liquid metal, silver-based compounds) are electrically conductive. If they squeeze out from under a heatsink and contact nearby SMD components or traces, they will create short circuits that can permanently damage the hashboard. Stick with non-conductive ceramic or carbon-based pastes exclusively. This is non-negotiable.
Fan Maintenance
The AvalonMiner 1066 uses four 12038-size fans — two intake, two exhaust. Fan failure is one of the most common issues on any ASIC miner, and it is usually the easiest to fix. Signs of fan problems include:
- Unusual noise: Grinding, clicking, or rattling sounds indicate bearing wear. A healthy fan produces a consistent, smooth whine.
- Vibration: Excessive vibration at the chassis indicates an unbalanced or damaged fan blade.
- Reduced airflow: If chip temperatures are climbing but the fans appear to be spinning, one or more fans may have degraded bearings that reduce RPM under load.
- Complete fan stop: The controller monitors fan speed. If a fan stops, the miner should trigger a temperature-based shutdown or error alert through the management interface.
To replace a fan: power down, disconnect AC, remove the fan guard screws, disconnect the fan power cable from the board, and install the replacement. Ensure the replacement fan matches the original specifications (120mm x 38mm, 12V DC, correct airflow direction). Reverse-installed fans create dead zones in the airflow path that cause localized overheating.
Diagnostics & Troubleshooting
When your AvalonMiner 1066 is not performing as expected — low hashrate, dropped boards, overheating, or complete failure — a systematic diagnostic approach saves you hours of guesswork. The Avalon architecture means you need to diagnose two separate systems: the miner hardware and the controller stack.
LED Status Indicators
The AvalonMiner 1066 has a front-panel LED indicator that communicates the miner’s operational state through colors and patterns. Learn to read it:
LED Status Reference
| Flashing Red | Miner is booting up. Normal during initial power-on. Should transition to green within 2–3 minutes. |
|---|---|
| Solid/Flashing Green | Normal operation — mining actively. Everything is working. |
| Flashing White | Configuration mode — entered by holding FUNC button 5 seconds. Web interface accessible. |
| Flashing Yellow | Exiting configuration mode, returning to normal mode. |
| Flashing Blue | Hash board LED inspection mode — individual hashboard LEDs visible. Toggle with FUNC button. |
| Persistent Red | Error state — firmware error, hashboard failure, or communication failure. Requires diagnostics. |
| No Light | No power to the control panel — check power supply, cables, and fuses. |
The FUNC button on the AvalonMiner 1066 cycles through modes with each 5-second hold: Normal (green) → Configuration (white) → Normal (yellow transition) → LED Inspection (blue) → Normal (green). The IP Reporter button gives you the miner’s IP address with a short press. The RESET button (held 5 seconds, LED flashes red) restores factory settings — use with caution as it erases all custom configuration.
Controller Web Interface Diagnostics
The Avalon management web interface runs on the controller (Raspberry Pi or Canaan controller box), not on the miner itself. Access it by navigating to the controller’s IP address in your web browser. Default credentials are typically root / root — change them immediately after setup.
Key diagnostic information available in the web interface:
- Hashrate per board: Each of the three hashboards should report approximately 16.7 TH/s (50 ÷ 3). Significant deviation indicates a problem on that board.
- Chip count: Each board should report 114 chips. Missing chips mean dead or unresponsive ASICs.
- Temperature readings: Per-board temperature sensors. Normal operating range is 55°C to 75°C. Above 85°C indicates cooling problems.
- Fan speed: Reported in RPM. All four fans should show similar speeds. One fan at zero or significantly lower RPM means a failed fan.
- Hardware errors (HW): A rising HW error count indicates chip-level problems. Occasional HW errors are normal; a rapidly climbing count points to a failing chip or bad solder joint.
- Pool connection status: Verify the controller is connected to your configured pools and submitting shares.
- Firmware version: Note the current version for troubleshooting and to determine if an update is available.
Common Error Codes & Symptoms
Troubleshooting Quick Reference
| Fewer than 3 hashboards detected | Check power cables (both connectors per board), hashboard-to-MM-board connection, and hashboard LED indicators. Reseat cables. Test with known-good power supply. |
|---|---|
| 0 TH/s — no hashing | Verify controller is running, AUC3 is connected and powered, CAN bus cable is intact, and pool configuration is correct. Check if issue is controller-side or miner-side. |
| Low hashrate (below 45 TH/s) | Check for missing chips (chip count below 114 on any board), high temperatures causing throttling, or high HW error rates. Clean dust, check fans, consider thermal paste replacement. |
| High chip temperatures (>85°C) | Check fan operation (all 4 spinning), dust buildup, ambient temperature, thermal paste condition, and airflow path for obstructions. |
| Persistent red LED after boot | Firmware corruption or hashboard communication failure. Try firmware reflash. If problem persists, test hashboards individually by disconnecting two and connecting one at a time. |
| No LED, no fan spin on power-up | PSU failure, blown fuse, bad AC cable, or failed power connector. Test PSU output with multimeter: should read 12V DC (±5%). |
| Controller cannot find miner | AUC3 failure, CAN bus cable damage, or MM board failure. Test with replacement AUC3. Check CAN bus cable continuity. Verify AUC3 firmware is compatible. |
| Frequent pool disconnections | Network issue (controller-side). Check Ethernet cable, router, DNS settings. Not a miner hardware issue — this is controller/network territory. |
| Rising HW error rate | Failing chip or degraded solder joint. Note which board reports errors. If one board has significantly more errors, that board needs inspection. Clean, repaste, and retest. If errors persist, chip-level repair or board replacement is needed. |
Diagnostic Commands via Controller
If your controller runs a Linux-based system (Raspberry Pi with Avalon management software), you can access deeper diagnostics via SSH. Connect to the controller (not the miner — remember, the miner has no network interface):
ssh root@<controller-ip-address>Once connected, you can query cgminer (the Avalon fork) directly for real-time diagnostics:
# Check overall status and hashrate
echo '{"command":"summary"}' | nc localhost 4028
# List all detected devices (each miner unit)
echo '{"command":"devs"}' | nc localhost 4028
# Get detailed stats per device
echo '{"command":"estats"}' | nc localhost 4028
# Check pool connection status
echo '{"command":"pools"}' | nc localhost 4028
# Get per-device temperature and fan data
echo '{"command":"devdetails"}' | nc localhost 4028The estats command is the most useful for diagnostics — it returns per-board chip counts, temperatures, voltages, fan speeds, and error rates. Parse this output to identify which board is underperforming.
The cgminer API listens on port 4028 by default. If the API is not responding, check that cgminer is running (ps aux | grep cgminer) and that the API is enabled in the cgminer configuration. Some Canaan controller firmware builds restrict API access by default — you may need to add –api-listen –api-network –api-allow W:0/0 to the cgminer startup parameters (adjust for your security requirements — this example allows all IPs).
Individual Hashboard Testing
When a hashboard is suspect, isolate it for testing:
- Power down the miner completely. Disconnect AC power.
- Disconnect all three hashboards from the MM board and power supply.
- Connect only the suspect board — both 12V power connectors and the hashboard-to-MM-board data connection.
- Power up and check the controller interface. The miner should detect one hashboard with 114 chips and approximately 16.7 TH/s.
- If the board fails to detect or reports low chip count: the problem is on that board. Proceed to hashboard repair section.
- If the board works fine alone: the problem may be with the PSU (insufficient power for all three boards simultaneously), the MM board, or inter-board interference. Test each board individually, then add boards one at a time to identify the combination that triggers the fault.
If all three boards test fine individually but fail when all are connected, the PSU is the most likely culprit — it cannot deliver sufficient current under full load. Test with a known-good PSU.
Common Repairs
Fan Replacement
Fan replacement on the AvalonMiner 1066 is the most straightforward repair and requires no soldering:
- Power down and disconnect AC power.
- Remove the fan guard screws (typically 4 per fan, Phillips head).
- Disconnect the fan power cable from the board connector.
- Remove the old fan and note the airflow direction arrow on the side of the fan housing.
- Install the replacement fan with the same airflow direction. The intake fans (front) should blow air into the chassis. The exhaust fans (rear) should pull air out.
- Reconnect the fan power cable. Ensure it is fully seated in the connector.
- Replace the fan guard and screws.
- Power up and verify the new fan is reported in the controller interface at expected RPM.
The AvalonMiner 1066 uses a push-pull airflow configuration: front fans push air in, rear fans pull air out. Installing a fan backwards creates a dead zone in the airflow path where hot air recirculates instead of being exhausted. This can cause localized chip temperatures to spike by 15–25°C even though overall fan RPM looks normal. Always match the airflow arrow on the replacement fan to the original orientation.
Power Supply Troubleshooting
The AvalonMiner 1066 requires a PSU that can deliver sustained 3,250W+ at 12V DC. Common PSU-related symptoms and fixes:
- Miner does not power on at all: Test the AC outlet with another device. Test the AC power cord. Check for a blown fuse on the PSU. Measure PSU output with a multimeter — should read 12V DC (±5%). If no output, the PSU has failed and needs replacement.
- Miner boots but hashboards drop under load: The PSU cannot sustain full load current. This is common with aging PSUs or undersized units. Measure voltage under load at the hashboard power connectors — if it drops below 11.4V, the PSU is sagging. Replace with a unit rated for at least 3,500W continuous.
- Burnt or discolored power connectors: High-resistance connections generate heat, cause voltage drop, and accelerate further degradation. Replace the affected cable and connector. Clean the mating connector on the hashboard with IPA. If the hashboard-side connector is damaged, that is a more involved repair.
- PSU fan not spinning: Most PSU fans are thermally controlled and only spin under load. If the PSU fan never spins even under full load, the PSU is at risk of thermal shutdown. Replace the PSU.
The AvalonMiner 1066 uses a C19 plug for AC input. Ensure your power cord, breaker, and outlet can handle the full 16A draw. A 20A circuit with 12-gauge wiring is recommended. Running a 3,250W miner on a 15A circuit shared with other loads is asking for breaker trips and potential fire hazards.
Hashboard Issues
Hashboard problems range from simple (loose cable) to complex (dead ASIC chip). Work through this decision tree:
- Board not detected at all:
- Reseat both 12V power connectors — firmly.
- Reseat the data cable between the hashboard and the MM board.
- Inspect connectors for bent pins or corrosion.
- Test the board in a different slot (if possible) to rule out an MM board port issue.
- Measure voltage at the power connector under load — confirm 12V is reaching the board.
- If the board is still not detected: the onboard voltage regulators, the CAN bus interface, or a critical chip in the chain may be damaged. This requires component-level diagnosis.
- Board detected with low chip count:
- Some chips are not responding. This typically means a chip has failed or a solder joint has cracked, breaking the serial signal chain.
- The number of missing chips helps localize the fault. If the last 20 chips in the chain are missing, the break is at chip ~94. If only 1–2 chips are missing, those specific chips have failed independently.
- Chip-level repair requires removing the heatsink over the suspect area, identifying the failed chip (visually or with an oscilloscope tracing signal flow), and replacing it with a known-good A3205. This is advanced work — if you do not have hot air rework experience, send the board to a professional.
- Board detected but low hashrate / high HW errors:
- Often a thermal issue. Clean, repaste, and retest.
- If errors persist after thermal maintenance, a chip is partially failing — producing hashes but at high error rates. Use the estats API output to identify temperature anomalies that point to the failing chip’s location.
ASIC Replacement Parts
D-Central stocks replacement hashboards, ASIC chips, fans, power cables, and components for Canaan AvalonMiners and other major manufacturers. All parts tested before shipping from our Laval, Quebec facility.
Controller & AUC3 Issues
Because the Avalon architecture separates the controller from the miner, controller-side problems are a distinct diagnostic category that does not exist with Antminers:
- AUC3 not detected by controller: Check USB connection between AUC3 and Raspberry Pi / controller box. Try a different USB port. Test with a replacement AUC3 module. Update the AUC3 firmware if possible.
- Controller finds AUC3 but not the miner: CAN bus cable problem. Check both ends of the CAN bus cable for damage or loose connections. Try a replacement cable. If the cable tests good, the MM board on the miner may be faulty.
- Multiple miners in daisy chain, one drops out: Check the CAN bus cable to the affected unit. Ensure the daisy chain is not too long (CAN bus has distance limitations). Try connecting the problem miner directly to the AUC3 to rule out daisy-chain issues.
- Controller web interface not accessible: Verify the controller’s IP address has not changed (check your router’s DHCP lease table). Try accessing via direct Ethernet cable. SSH into the controller and check if the web server process is running.
- cgminer crashing or not starting: Check controller disk space (SD card full is common on Raspberry Pi). Check cgminer log files for error messages. Reinstall the controller firmware from Canaan’s release files.
Firmware & Software
Firmware Updates
Firmware updates on the Avalon platform apply to two separate systems:
- Controller firmware: The management software running on the Raspberry Pi or Canaan controller box. This controls pool configuration, monitoring, and communication with the AUC3/miners. Updates are typically distributed as SD card images that you flash to the controller’s storage.
- Miner firmware: The firmware on the miner’s MM board. This controls hashboard initialization, chip configuration, and local diagnostics. Updates are pushed through the controller’s web interface or via the FMS (Firmware Management System) for batch operations.
To update miner firmware through the web interface:
- Download the appropriate firmware file from Canaan’s official website. Verify the file is specifically for the AvalonMiner 1066 (A10 series). Using firmware for a different Avalon generation will either fail to flash or, worse, brick the MM board.
- Access the controller web interface and navigate to the firmware upgrade section.
- Upload the firmware file and initiate the upgrade.
- Wait for the process to complete — do NOT power off the miner during a firmware update. Interrupting a flash can corrupt the MM board firmware and require a recovery procedure.
- The miner will reboot automatically. Verify that all three hashboards are detected and hashrate returns to expected levels.
Only download firmware from Canaan’s official website or verified mirrors. Third-party firmware modifications exist in the Avalon ecosystem, but using untrusted firmware risks: (1) malware that redirects a percentage of your hashrate to the attacker’s pool address, (2) overclocking settings that damage your hardware, or (3) bricked devices. In the Bitcoin Mining Hackers spirit, we trust open-source and verifiable code — not random downloads from Telegram groups.
Batch Upgrades Using FMS
If you manage multiple Avalon miners, the Firmware Management System (FMS) allows batch firmware deployment:
- Install FMS on a Linux machine or Raspberry Pi on the same network as your controllers.
- Add your controller(s) to the FMS dashboard.
- Select the miner units to upgrade.
- Upload the firmware file and initiate the batch upgrade.
- FMS displays per-device progress and completion status.
FMS is also useful for monitoring firmware version consistency across a fleet — it can identify miners running outdated firmware at a glance.
Configuration Best Practices
- Configure at least three mining pools in the controller settings. The miner will failover to pool 2 if pool 1 goes down, and to pool 3 if pool 2 also fails. This ensures maximum uptime and hashrate utilization.
- Set a strong password on the controller web interface. The default root/root is widely known. Anyone on your local network can access and reconfigure your miner if you leave defaults in place.
- Use static IP or DHCP reservation for the controller. If the controller’s IP changes, you lose access to the web interface until you find the new address.
- Monitor firmware versions regularly. Canaan occasionally releases updates that improve hashrate consistency, reduce error rates, or fix security vulnerabilities.
- Do not overclock beyond factory settings unless you understand the thermal and power implications. Overclocking increases chip voltage, which increases power consumption quadratically (P = V²/R). A 10% frequency increase can produce a 20%+ power increase, pushing your PSU and thermal solution beyond their safe limits.
Restoring Factory Settings
If your miner’s configuration is corrupted or you need a clean start, here is the factory reset procedure:
- Power off the miner completely. Disconnect AC power. Wait 60 seconds.
- Reconnect AC power and turn on the miner.
- Watch the front-panel LED — it will flash red during boot.
- While the LED is still red, press and hold the FUNC button for 5 seconds. The LED will begin flashing white, indicating configuration mode.
- Press and hold the RESET button for 5 seconds. The LED will flash red, confirming the factory reset has been initiated.
- The miner will restart and boot with default factory settings. The LED will turn green when normal operation resumes.
A factory reset on the miner unit resets the miner-side configuration (network mode, hashboard parameters). It does NOT reset the controller. Your pool configurations, controller password, and management software settings live on the controller (Raspberry Pi / controller box) and are unaffected by resetting the miner unit. To reset the controller, you would need to reflash the controller’s SD card or use the controller’s own reset procedure.
Advanced Hashboard Diagnostics
This section is for technicians comfortable with a multimeter and oscilloscope. If you have never probed a PCB before, this is where we recommend calling a professional — specifically, us.
Voltage Rail Testing
Each hashboard has three critical voltage rails. Test these with the hashboard connected to the PSU but disconnected from the MM board (power only, no data):
Hashboard Voltage Rails
| Input (12V DC) | Measured at power connector pins. Expected: 11.4V – 12.6V. Below 11.4V = PSU sagging. Above 12.6V = PSU over-voltage (replace PSU immediately). |
|---|---|
| Vcore | Core voltage for ASIC chips. Varies by group/domain. Measure at test points near chip groups. Expected range depends on frequency setting. A dead group will show 0V where it should show regulated voltage. |
| VDDIO (I/O voltage) | Powers the chip I/O and communication logic. Typically 1.8V. A missing VDDIO rail means the chips cannot communicate even if Vcore is present — the board appears dead despite power being applied. |
Signal Chain Tracing
When chips drop off the chain, the fault is at the boundary between the last responding chip and the first missing chip. To trace the signal chain:
- From the controller’s estats output, determine how many chips are detected on the problem board. Example: 80 out of 114 detected means chips 81–114 are dark.
- Identify the physical location of chip 80 and chip 81 on the PCB (refer to the board’s silk-screen chip numbering or Canaan’s hashboard layout documentation).
- Using an oscilloscope, probe the CK (25 MHz clock) signal at chip 80’s output and chip 81’s input. If the signal is present at chip 80’s output but missing at chip 81’s input, the fault is in the trace or solder joint between them.
- If the signal is present at chip 81’s input but the chip does not respond, chip 81 itself has failed and needs replacement.
- For the Data (D) signal, trace it in the same direction as the chip chain. A missing D signal prevents work distribution to all downstream chips.
Chip replacement on the A3205 requires hot air rework equipment and experience with BGA/QFN packages. The chip must be precisely positioned and reflowed at the correct temperature profile to avoid damaging neighboring components. If you do not have this equipment and experience, professional repair is the correct path.
Preventive Maintenance Schedule
Consistency beats intensity. A regular maintenance schedule prevents the cascading failures that turn a minor dust issue into a dead hashboard. Here is our recommended schedule based on years of servicing Avalon miners:
Maintenance Schedule
| Daily | Check controller dashboard: hashrate, temperatures, fan speeds, pool connectivity. Takes 2 minutes. Set up email/push alerts for anomalies. |
|---|---|
| Weekly | Review HW error trends. Compare current hashrate to baseline. Listen for unusual fan noise. Check ambient temperature and humidity in the mining area. |
| Monthly | Visual inspection of chassis, cables, and connectors. Blow out dust with compressed air. Check for signs of overheating (discolored connectors, melted insulation). |
| Quarterly | Deep clean: remove top cover, clean heatsink fins, clean fans individually, inspect PCB surfaces. Check CAN bus cable and AUC3 connections. Verify firmware is current. |
| Annually | Full thermal paste replacement on all hashboards. Replace fans proactively (fan bearings degrade with continuous use). Test PSU output voltage under load. Update controller firmware. |
Frequently Asked Questions
What is the AUC3 and why does the AvalonMiner 1066 need an external controller?
The AUC3 (Avalon USB Converter 3) is a bridge module that connects one or more AvalonMiner units to a controller (typically a Raspberry Pi or Canaan’s standalone controller box). Unlike Antminers, which have an integrated control board, Canaan designed the Avalon architecture with a modular philosophy: the miner is a pure hashing unit, and the controller handles pool communication, configuration, and monitoring. This means one controller can manage multiple miners via daisy chain, and you can swap miners in and out without reconfiguring the controller. It is a more decentralized approach to hardware architecture — fitting for a Bitcoin miner.
How many AvalonMiner 1066 units can I daisy-chain to one controller?
Canaan’s official recommendation is up to 20 miner units per AUC3 module, depending on the CAN bus cable length and quality. In practice, keeping the chain under 10–15 units with short, high-quality cables produces the most reliable results. Longer chains increase communication latency and the probability of CAN bus errors. If you are running more than 20 units, use multiple AUC3 modules connected to the same controller via USB.
Can I use a Bitmain APW power supply with the AvalonMiner 1066?
Yes, in many cases. The AvalonMiner 1066 requires 12V DC at sufficient amperage. Bitmain APW series PSUs (APW3++, APW7, APW9+, APW12) output 12V DC and can physically connect to the Avalon hashboard power connectors if the connector type matches or you use adapters. The critical requirement is that the PSU can sustain 3,250W+ continuous output. An APW7 (1800W) is insufficient; an APW12 (3600W) works well. Always verify connector compatibility and polarity before connecting. An incorrect polarity connection will instantly destroy the hashboard.
Why does my AvalonMiner 1066 only show 2 hashboards instead of 3?
A missing hashboard usually means: (1) loose or disconnected power cable on that board — check both 12V connectors, (2) loose or damaged data cable between the hashboard and the MM board, (3) failed voltage regulator on the hashboard preventing it from powering up, or (4) a catastrophic chip failure that crashes the entire board. Start by reseating all cables. If the board is still missing, test it individually (disconnect the other two boards) to determine if the board is dead or if there is a power delivery issue when all three boards draw simultaneously.
Is the AvalonMiner 1066 still profitable to mine with?
At 65 J/TH, the AvalonMiner 1066 is not competitive with current-generation miners on pure efficiency. Profitability depends entirely on your electricity cost. At $0.05/kWh or below, the 1066 can still be marginal or profitable depending on Bitcoin price and network difficulty. But profitability is not the only metric that matters. If you are using the 1066 as a Bitcoin Space Heater, the electricity cost is offset by heating value — you were going to spend that energy on heat anyway. Every hash the 1066 produces strengthens Bitcoin’s decentralization, and that has value beyond the satoshis it earns. Technology-first, not investment-first.
How loud is the AvalonMiner 1066 and can I reduce the noise?
The AvalonMiner 1066 produces approximately 75 dB at full speed — comparable to a vacuum cleaner running continuously. For home mining, this is uncomfortably loud for living spaces. Options to reduce noise: (1) replace stock fans with quieter aftermarket fans (at the cost of reduced airflow — monitor temperatures carefully), (2) build or purchase a sound-dampening enclosure with proper ventilation, (3) use a fan shroud with ducting to route exhaust outside or to another room, or (4) convert to a Bitcoin Space Heater configuration with integrated sound dampening. D-Central builds custom space heater editions specifically designed for home environments.
Can I run the AvalonMiner 1066 on 120V power?
No. The AvalonMiner 1066 requires 185–280V AC input. Standard North American 120V outlets cannot power this miner. You need a 240V outlet (like a dryer outlet, NEMA 14-30 or NEMA 6-20) with an appropriate circuit breaker. Running a 3,250W miner requires at least a 20A 240V circuit. Have a licensed electrician install the circuit if you do not already have one. Do not use voltage step-up transformers as a workaround — they add resistance, heat, and a fire hazard at this power level.
How do I find the IP address of my AvalonMiner 1066?
The IP address you need is the controller’s IP address, not the miner’s. The miner itself does not have a direct network connection — it communicates through the AUC3 to the controller. Find the controller’s IP by: (1) checking your router’s DHCP lease table for the controller (Raspberry Pi), (2) pressing the IP Reporter button on the miner briefly — if your miner unit has this feature, it may announce the controller’s IP, or (3) using a network scanner like Advanced IP Scanner or nmap to find devices on your network. If using a Raspberry Pi controller, you can also connect a monitor and keyboard to the Pi to check its IP directly.
What is the expected lifespan of an AvalonMiner 1066?
With proper maintenance, the hardware can physically run for 5+ years. The A3205 ASIC chips themselves rarely fail spontaneously — it is the supporting components (fans, power connectors, thermal interface, capacitors) that degrade over time. Fans typically need replacement every 1–2 years. Thermal paste should be refreshed annually. Power connectors develop resistance over thousands of hours of high-current operation. The “economic lifespan” (when electricity cost exceeds mining revenue) is a separate question that depends on Bitcoin price, network difficulty, and your electricity rate. But remember: economically “unprofitable” miners still produce heat. In a Canadian winter, a 3,250W heater that earns even a few sats per day is better than a 3,250W heater that earns nothing.
Where can I get replacement parts for the AvalonMiner 1066?
D-Central Technologies stocks replacement components for Canaan AvalonMiners, including hashboards, fans, power cables, AUC3 controllers, and CAN bus cables. We ship from our facility in Laval, Quebec, across Canada and worldwide. Browse our parts catalog or call 1-855-753-9997 to confirm availability for your specific component needs. We also carry A3205 ASIC chips for technicians performing chip-level repairs.
When to Call a Professional
This guide gives you the knowledge to handle routine maintenance, basic diagnostics, and straightforward repairs like fan replacement. But some situations demand professional equipment, professional skills, and professional accountability:
- Chip-level repairs — replacing a dead A3205 requires BGA/QFN rework equipment and technique. A bad reflow can damage neighboring chips and turn a one-chip repair into a board replacement.
- Voltage regulator failures — the power delivery circuitry uses precision components. Incorrect replacement can over-voltage and kill every chip on the board.
- MM board / CAN bus interface damage — the Management Module is a specialized communication board. Repair requires understanding the CAN bus protocol and the specific Canaan firmware interface.
- Multiple boards failing simultaneously — this usually indicates a systemic issue (PSU, environment, or power quality) that requires holistic diagnosis, not just component swapping.
- Firmware recovery from a bricked state — if a failed firmware update has corrupted the MM board, recovery requires direct serial/JTAG access and specialized tools.
- Any repair you are not 100% confident performing — there is no shame in knowing your limits. A miner sent to a professional for a $150 repair is better than a miner destroyed by an amateur attempt at a $150 repair.
D-Central ASIC Repair Service
D-Central Technologies has been repairing ASIC miners since 2016 — Bitmain, Canaan, MicroBT, and Innosilicon. We have repaired 2,500+ miners at our facility in Laval, Quebec. We diagnose at the component level, stock replacement parts, and stand behind our work. Avalon miners are less common in North American repair shops, but we have deep experience with the entire Canaan lineup. Ship your miner to us, and we will get it hashing again.
We accept repair shipments from across Canada and internationally. Package your miner securely (original box or double-boxed with foam padding — hashboards are heavy and fragile). Include a note describing the symptoms. Ship to: 1325 Rue Bergar, Laval, QC H7L 4Z7, Canada. We will diagnose, quote, and await your approval before performing any paid repair work. Typical turnaround is 5–10 business days depending on parts availability.
Every AvalonMiner 1066 you keep hashing is another independent node contributing to Bitcoin’s decentralization. Canaan built the first commercial Bitcoin ASIC — they started this industry. Maintaining their hardware is maintaining Bitcoin history. Keep it running. Keep it hashing. Keep it decentralized.
Interactive Hashboard Schematic
Explore the AVALONMINER 1066 hashboard layout below. Toggle layers to isolate voltage domains, signal chains, test points, key components, and thermal zones. Hover over any region for quick specs — click for detailed diagnostics, failure modes, and repair guidance.