How to fix my Antminer if one hashboard shows significantly lower hashrate?

Troubleshooting Low Hashrate on a Single Hashboard

1. Verify physical connections between the hashboard and control board using a multimeter to check for continuity across all signal pins and power rails.

2. Inspect solder joints on the BM1387 ASIC chips for cold joints, bridging, or oxidation—especially around the clock input and voltage regulation circuitry.

3. Measure VDD and VDDQ voltages at test points on the affected board; deviation beyond ±3% from nominal 0.8V indicates regulator failure or trace damage.

4. Swap the suspected hashboard with another identical unit from the same miner to isolate whether the issue follows the board or remains at the slot location.

5. Run bmminer --test-hw --board 3 via SSH to force hardware-level diagnostics and capture raw chip response timing logs.

Firmware and Configuration Diagnostics

1. Confirm firmware version matches the exact hardware revision by checking output of cat /etc/bmminer.conf | grep firmware and cross-referencing with Bitmain’s S9 firmware compatibility matrix.

2. Disable AsicBoost temporarily by setting asicboost=off in bmminer.conf and rebooting to rule out timing mismatches under boosted clock domains.

3. Adjust frequency scaling manually: set freq=500 in the board-specific section and observe stability over 30 minutes before incrementally raising to target.

4. Check thermal throttling indicators via cat /tmp/thermal.log; sustained junction temperatures above 85°C on any chip trigger automatic frequency rollback.

Power Delivery Path Analysis

1. Monitor rail ripple on +12V input using an oscilloscope—excessive noise (>150mVpp) correlates strongly with intermittent hash loss on edge-mounted boards.

2. Replace the 12V power cable with a known low-impedance variant rated for ≥60A continuous draw to eliminate voltage sag during high-load transitions.

3. Test each MOSFET in the board’s local DC-DC converter stage for gate threshold consistency; mismatched thresholds cause asymmetric current sharing among ASIC clusters.

4. Inspect PCB vias connecting the power plane to the hashboard edge connector for microfractures using 20x magnification and oblique lighting.

ASIC-Level Failure Identification

1. Execute chip-level burn-in using bmminer --scan-chips --board 3 --cycles 50 to identify failing dies that respond only intermittently.

2. Cross-check serial numbers etched on BM1387 packages against known batch recalls—certain 2018 Q2 wafers exhibited elevated defect rates in cluster #7 logic units.

3. Probe test points TP12–TP15 on the hashboard while running full-load hashing to detect missing clock edges or stalled reset signals.

4. Compare thermal camera readings across all 64 ASICs; a single cooler die amid uniformly hot neighbors suggests internal short or open bond wire.

Common Questions and Direct Answers

Q: Can I replace just one BM1387 chip on an S9 hashboard?A: No. Individual ASIC replacement is not viable due to proprietary die bonding, underfill epoxy adhesion, and uncalibrated thermal interface material redistribution.

Q: Does cleaning the hashboard with isopropyl alcohol affect performance?A: Yes—if concentration exceeds 90%, residual solvent can degrade conformal coating on oscillator circuits, leading to clock jitter and hash instability.

Q: Why does the low-hash board report correct temperature but erratic fan speed?A: That indicates faulty tachometer signal routing on the board’s sensor bus—not fan hardware failure—often caused by cracked traces near the JST connector.

Q: Is it safe to run the miner with one hashboard disabled?A: Yes, provided the remaining boards operate within total power envelope limits; however, firmware may reduce global frequency to maintain thermal headroom, lowering overall efficiency.