Diagnostics

Reading a Miner Status Page

Decode the live status dashboard field by field - the first screen on every service call - and read board-down, throttle, and fan-loss at a glance.

The first screen you open

Every service call starts on the miner's live status page (the web dashboard / cgminer-style summary). It shows what the unit is doing RIGHT NOW. Antminer, Whatsminer and Avalon label the fields a little differently, but the readings mean the same thing - learn the concepts, not the labels.

This is the LIVE dashboard, not the kernel log. The status page tells you the current state; kernel-log strings (covered in the reference tables) tell you what already went wrong. Read the dashboard first.

Field by field

A chain = one hashboard. Most units run 3 chains. Read every field top to bottom before you touch anything.

Field	What it tells you	Bad reading
RT vs avg hashrate	Instantaneous (RT) vs rolling average	Big gap between the two = unstable / bouncing
Chips per chain	ASIC count found on that hashboard	0/126 on a chain = that board is DOWN
Frequency (MHz)	The tuned chip clock	Dropped below its set point = thermal throttle
Chip temp / PCB temp	Two separate sensors per board	Watch the HOTTER of the two; near limit = throttle
HW-error count	Chip returned a bad nonce	Climbing steadily = weak/hot board (keep under ~1%)
LSTime (last share)	Seconds since last accepted share	Large or growing = not landing work (pool/net or dead boards)
Reject rate	Shares the pool refused	Rising % = overclock too far / bad flash / failing board
Fan RPM	Live fan speed	0 RPM on any fan = the miner will HALT on fan-loss

Three reads that tell you the fault instantly

0 chips on a chain -> board-down. That hashboard makes nothing; roughly a third of the unit's output gone.
0 RPM on a fan -> imminent halt. Fan-loss protection stops the miner; find the dead/unplugged fan now.
Full chip count but low AVERAGE hashrate -> degraded / throttled. All boards present, but heat or a dropped frequency is holding it back.

Never leave a miner running with a fan showing 0 RPM. It will trip fan-loss and halt - and if the guard ever fails, boards cook. A dead fan is a stop-and-fix, not a note-for-later.

RT vs average, and LSTime

These two catch problems the raw hashrate number hides.

RT (real-time) hashrate swings second to second; the rolling average smooths it. A steady unit shows both close together. A wide, jumpy gap = instability - chase temps, a weak board, or a flapping network.
LSTime is your pool-connection heartbeat. A few seconds is normal. Large or steadily climbing LSTime means the miner is not getting accepted shares landed - suspect the pool/network path first, then dead boards.
HW errors climb slowly on a healthy unit. Fast growth points at one weak or overheating board - cross-check it against the per-chain chip count and temps.

The first screen you open

Field by field

A chain = one hashboard. Most units run 3 chains. Read every field top to bottom before you touch anything.

Field	What it tells you	Bad reading
RT vs avg hashrate	Instantaneous (RT) vs rolling average	Big gap between the two = unstable / bouncing
Chips per chain	ASIC count found on that hashboard	0/126 on a chain = that board is DOWN
Frequency (MHz)	The tuned chip clock	Dropped below its set point = thermal throttle
Chip temp / PCB temp	Two separate sensors per board	Watch the HOTTER of the two; near limit = throttle
HW-error count	Chip returned a bad nonce	Climbing steadily = weak/hot board (keep under ~1%)
LSTime (last share)	Seconds since last accepted share	Large or growing = not landing work (pool/net or dead boards)
Reject rate	Shares the pool refused	Rising % = overclock too far / bad flash / failing board
Fan RPM	Live fan speed	0 RPM on any fan = the miner will HALT on fan-loss

Three reads that tell you the fault instantly

0 chips on a chain -> board-down. That hashboard makes nothing; roughly a third of the unit's output gone.

0 RPM on a fan -> imminent halt. Fan-loss protection stops the miner; find the dead/unplugged fan now.

Full chip count but low AVERAGE hashrate -> degraded / throttled. All boards present, but heat or a dropped frequency is holding it back.

Never leave a miner running with a fan showing 0 RPM. It will trip fan-loss and halt - and if the guard ever fails, boards cook. A dead fan is a stop-and-fix, not a note-for-later.

RT vs average, and LSTime

These two catch problems the raw hashrate number hides.

RT (real-time) hashrate swings second to second; the rolling average smooths it. A steady unit shows both close together. A wide, jumpy gap = instability - chase temps, a weak board, or a flapping network.

LSTime is your pool-connection heartbeat. A few seconds is normal. Large or steadily climbing LSTime means the miner is not getting accepted shares landed - suspect the pool/network path first, then dead boards.

HW errors climb slowly on a healthy unit. Fast growth points at one weak or overheating board - cross-check it against the per-chain chip count and temps.