Whatsminer M30S – LED Slow Red Blink

Power-cycle the miner from the PDU or breaker, not a soft reboot. Leave it off for 60 seconds. Soft reboots don't clear the `udhcpc` retry timer; a hard cycle does. Watch the LED during power-up — green for ~45 seconds, then a short amber, then either solid green (healthy) or slow red (still broken). Note which transition happens.

Reboot your router. Sounds trivial, but exhausted DHCP pools and stuck ARP tables on residential routers cause roughly 30% of "miner dropped off the network" reports. Let the router fully come up (2–3 minutes) before power-cycling the miner again.

Check the miner's Ethernet cable visually. Unplug both ends. Look for bent pins in the RJ45, oxidation on the contacts, crimping damage from a door hinge or furniture leg. Re-seat both ends firmly — listen for the click. If the cable is >5 years old or has been stepped on, swap it with a known-good Cat6 (≤$10). This is one of the single highest ROI fixes in the Canadian home-mining queue.

Factory-reset via the RESET button. Hold the RESET button on the control board for ≥3 seconds until the LED flickers red-green. This clears the stored network config back to DHCP, purges the cached DNS, and restarts cgminer. Do NOT hold longer than 10 seconds — some firmware builds interpret longer presses as "firmware recovery request" and put the miner into a more fragile state. Re-apply pool config after reset.

Move the miner to a different switch port or a different switch entirely. Cheap unmanaged switches die silently — one port works, the next doesn't. If the LED clears on a different port, your switch is the problem, not the miner.

Run `IPFOUND` in WhatsminerTool and capture the miner's MAC + IP. If no IP is returned, check your router's DHCP lease table. Pre-emptively expand the lease pool if you're running more than ~20 devices on a /24 subnet — the miner's MAC won't get a lease if the pool is saturated. Set a static DHCP reservation for every ASIC you own; this eliminates lease-expiry drops entirely. MinerTool's identify function (traffic-light flash) helps correlate which physical unit holds which IP when you have multiple.

Switch DNS at the router to `1.1.1.1` + `9.9.9.9`. Residential ISPs in North America — Bell, Rogers, Videotron, Comcast, Xfinity — have intermittently failed to resolve mining-pool stratum hostnames on specific provincial or metro PoP nodes. Hard-coding public resolvers at the router eliminates this entire class of failure and speeds up every device on your LAN. While you're there, set the router to hand out those same DNS values via DHCP option 6.

Reflash MicroBT stock firmware for your exact hardware revision. Via WhatsminerTool: `Firmware Upgrade` → select the `.bin` for M30S vs M30S+ vs M30S++ (they are different) → click Start. If the upgrade fails and the miner bricks further, hold RESET during power-on to enter recovery mode (firmware ≥ `20210425`), re-attempt the flash. Stay on authorized MicroBT builds — custom firmware on M30S-class trips the `100001`–`100003` anti-tamper chain and produces the same slow red blink for a different reason.

Measure the Ethernet link at the miner's PHY. With the cable connected and the miner powered, verify the link LED on the switch side is solid (not blinking erratically). On a managed switch, pull the port statistics: look for `CRC errors`, `late collisions`, or `runts`. If any are non-zero and climbing, your cable or the miner's PHY has degraded. Re-cable first; if errors persist, Step 10.

SSH in and inspect logs. Default credentials and SSH availability vary by firmware revision — MicroBT does not publish them and changes them across builds. Check `/var/log/miner.log`, `/var/log/system.log`, and `dmesg` output for `eth0` link flaps, `udhcpc` retries, `stratum error`, `dns resolution failed`. The log schema is undocumented by MicroBT; cross-reference against D-Central's [Whatsminer log schema notes](https://d-central.tech/asic-troubleshooting/whatsminer-minertool-diagnostic-codes/) when decoding fields.

Bypass the network entirely with a direct laptop connection. Disconnect the miner from your LAN. Connect a laptop directly via a short Cat6 patch cable (no switch). Set the laptop NIC to `192.168.1.100/24`. Browse to `http://192.168.1.1`. If the miner's web UI answers, the miner is healthy and the fault is on your LAN/WAN side. If it doesn't answer, the control board or its PHY is the suspect — go to Tier 3.

Swap the control board with a known-good M30S control board. The control board is a single field-replaceable unit on the M30S — four screws, one ribbon to the backplane, one power connector. If the swap clears the slow red blink, your original control board is the fault (usually a dying PHY, a failed RTC chip, or NAND corruption). D-Central stocks M30S control boards as a replacement part — cheaper than shipping the whole miner for repair. Preserve the original for diagnostics.

Replace the RTC coin-cell battery. The Xilinx Zynq on the M30S control board uses a small coin-cell (CR1220 on most revisions, CR2032 on later ones — confirm before buying) to maintain real-time clock across power events. A dead battery means every power-up hands cgminer a 1970-01-01 timestamp, which stratum servers reject. Swap the cell (≤$5), power-cycle, verify `date` via SSH shows correct Eastern Time (or whichever zone you configured).

Verify firmware hash offline. Download the M30S firmware directly from Whatsminer's official repository. Hash it locally (SHA-256). Compare against the hash the miner reports in WhatsminerTool. If they diverge, the flash is corrupt — re-flash via WhatsminerTool's recovery mode, or JTAG-flash via the exposed header on the control board if recovery mode fails. JTAG flashing requires a Xilinx-compatible programmer and is a bench-only procedure.

Check the Ethernet magnetics / PHY. On the M30S control board, the Ethernet magnetics module sits inline with the RJ45 jack. Physical damage (cracked solder, corrosion from humidity) or a failed magnetics IC reports as intermittent link, constant DHCP retries, and — yes — slow red blink. Visual inspection under magnification; reflow if joints are suspect; replace the magnetics module if the part is available.

Stop DIY when: RTC replacement doesn't hold time across reboots (the RTC IC itself is failing, not just the battery), the PHY / magnetics replacement doesn't clear the slow red blink, or you've confirmed NAND corruption via `dmesg` and a recovery-mode reflash fails. You're now in chip-level rework. [Book a D-Central ASIC Repair slot](https://d-central.tech/services/asic-repair/) — M30S control-board rework is a standard bench job.

What D-Central does at the bench. Full M30S control-board diagnostic under test fixture: PHY / RTC / NAND isolation, JTAG-level firmware integrity verification, component-level rework on failed PHY or magnetics, NAND re-image with verified MicroBT factory firmware, post-repair 24-hour burn-in with a live pool connection and per-chain hashrate verification. Whole-miner turnaround or control-board-only service — your choice.

Ship safely. Anti-static bag for the control board (or the whole miner if shipping complete). Double-box, 5 cm foam all around, tape all seams. Include: miner serial, firmware version last seen, observed LED cadence, and a contact number. D-Central's Canadian shipping workflow cuts customs friction for US/international senders — ask about the cross-border ASIC repair channel when you book.