How do you check ASIC chip health without custom firmware?

You read each miner through its manufacturer's own local API, which already reports per-board and per-chip data on stock firmware — no flashing, no BOS/Vnish required. MinersMe Cloud's per-site agent scans miners on the local network and pulls per-chip status, temperature and hashrate for every board, so chip diagnostics work on factory firmware.

← Blog · July 26, 2026 · Feature deep-dive

ASIC chip diagnostics: find the dead chip before the board dies

Q: What is ASIC chip diagnostics?

ASIC chip diagnostics is reading a miner's health down to the individual ASIC chip rather than just the whole machine: per-chip status, per-board chip count, per-board hashrate and temperature, and the miner's error log. It shows exactly which chip on which hashboard has failed, so a sagging-but-online miner can be diagnosed instead of guessed at. MinersMe Cloud does this on stock firmware across Antminer, WhatsMiner, IceRiver and Avalon.

Q: Why does a miner lose hashrate but still show online?

Almost always one or two dead or throttling chips on a single hashboard. The miner stays reachable and reports a status, so board-level monitoring calls it healthy — but its real hashrate has dropped. Only per-chip diagnostics reveal the dead chip and which board to pull.

Q: How many chips are on an ASIC hashboard?

It varies by model — an Antminer S19-class board carries on the order of 100+ chips (for example ~126), with three boards per miner. MinersMe Cloud reads the exact chip count and per-chip status for each board so a single failed chip out of hundreds is pinpointed.

A miner shows online. Its hashrate is down 9%. Nobody gets an alert, because board-level monitoring only knows "reachable" and "not reachable." The real story is one dead chip on one of its three hashboards — and until you can see that, you're swapping whole boards on a guess. ASIC chip diagnostics is what closes that gap.

What "per-chip" actually means

Board-level monitoring treats a miner as a single number. ASIC chip monitoring reads it the way the hardware is actually built — three hashboards, each carrying a hundred-plus chips — and reports:

Per-chip status — every chip on every board, so a single failed or throttling chip out of hundreds is pinpointed.
Per-board hashrate — which of the three boards is dragging the machine down.
ASIC chip temperature — per-board thermals, so a hot board is caught before it cooks its neighbours.
Live error-log analysis — the log signatures that precede a board failure, read continuously.

Why board-level monitoring misses it

A miner with one dead chip still answers the network, still reports a status, still looks green on a dashboard that only checks "up/down." Its hashrate has quietly dropped and stays dropped for weeks — a slow leak across a fleet of thousands. The only way to catch it is to look inside the machine, chip by chip, which is exactly what per-chip diagnostics do.

Chip diagnostics on stock firmware — no flashing

You don't need BOS, Vnish or any custom firmware to read chips. Every major vendor already exposes per-board and per-chip data through its own local API on factory firmware. MinersMe Cloud's per-site agent scans the local network and reads that data directly — brand-aware for Antminer (stock & Vnish), WhatsMiner, IceRiver and Avalon — so ASIC diagnostics work on the firmware your miners already run, with no warranty risk. (How the agent reads a whole facility without touching your switches is covered in how to monitor 1,000+ ASIC miners.)

From a dead chip to a fixed board

Chip diagnostics feed two things. First, failure prediction: a board losing chips one at a time is trending toward death, and the fleet-health dashboard ranks it before it fails outright. Second, repair: when a board crosses the line, the AI ticket bot opens a ticket that already names the miner, the board and the evidence, and it enters the RMA/repair pipeline while the machine keeps hashing on a spare. That's the whole point — a dead chip becomes a scheduled board swap instead of an emergency.

Deep test when you need certainty

Beyond passive monitoring, you can run a brand-aware deep hardware test on demand — a fuller diagnostic pass that confirms whether a suspect board is genuinely failing or just needs a reboot, before anyone drives to the site. It's the difference between dispatching a technician on evidence and dispatching one on a hunch.

See your own fleet chip-by-chip: create a free account, install the agent, and open any miner to its per-chip heatmap — included in Pro at $0.40/miner, everything on stock firmware.

FAQ

What is ASIC chip diagnostics? Reading a miner's health down to the individual chip — per-chip status, per-board chip count, hashrate and temperature, plus the error log — so you see exactly which chip on which board failed instead of guessing from the machine's total hashrate.

Can you monitor ASIC chips without custom firmware? Yes. Every major vendor reports per-chip data through its local API on stock firmware; MinersMe Cloud reads it directly, no flashing required.

Why does a miner lose hashrate but still show online? Usually one or two dead or throttling chips on a single hashboard — the miner stays reachable, so up/down monitoring calls it healthy while its real output has dropped.

How many chips are on a hashboard? Model-dependent — an S19-class board carries 100+ chips (about 126), three boards per miner. MinersMe Cloud reads the exact count and per-chip status for each.