Five Signs a Hash Board Needs Service Before Complete Failure
Subtle indicators our technicians watch for during routine farm walkthroughs — before a board drops to zero hashrate.
A hash board rarely dies without warning. Over four years of bench assessments at Panel Sequoiacore, our technicians have catalogued the early signs that precede a full slot failure. Catching these during a routine walkthrough can mean the difference between a planned twenty-minute swap and an emergency shutdown.
1. Gradual Hashrate Decline on a Single Slot
When one board slot drops 5–15% below its neighbours while others remain stable, the ASIC chips on that board are likely developing bad cores. Check your pool’s per-board reporting if your firmware supports it. A slow decline over two weeks is more concerning than a sudden drop, which often indicates a connection issue rather than chip degradation.
2. Elevated Board Temperature With Normal Ambient
If a specific slot runs 8–12°C hotter than adjacent boards under identical airflow, the thermal interface between the hash board and heatsink may be failing. Peel back the shroud during a scheduled maintenance window and inspect the thermal pads for hardening or uneven compression.
3. Intermittent LED Fault Codes
Many Antminer and Whatsminer models flash specific LED patterns before a board goes offline entirely. Photograph the pattern and cross-reference the manufacturer error code list. An intermittent code that clears after reboot often means a failing voltage regulator rather than dead ASICs — but the board should still be pulled before the regulator takes adjacent components with it.
4. Increased Rejected Share Rate on One Board
Pool-side rejected shares climbing on a single board slot suggest the chips are producing invalid hashes intermittently. This is easy to miss if you only monitor total farm hashrate. Set an alert threshold at 2% rejected shares per board.
5. Visible Corrosion Near Power Connectors
Humidity in Korean summers causes oxidation on board-edge power connectors, especially in farms without dehumidification. A green or white residue on the gold fingers where the board meets the backplane is a service flag. Clean with isopropyl alcohol and inspect the connector on the backplane side as well.
When to Schedule an Assessment
If you observe two or more of these signs on the same board, consider sending the rig model for a Hash Board Serviceability Assessment before the failure happens. Knowing the swap procedure in advance — and having a spare board on the shelf — turns an emergency into a scheduled maintenance task.