IceRiver Warranty Void: Third-Party Firmware Risks (xyys/tswift)

Read IceRiver's full warranty FAQ at `iceriver.io/faq` before doing anything. The 180-day window, original-purchaser clause, third-party firmware void clause, mixed control board void clause, refund policy. All of it. Five minutes here is the highest-leverage warranty work you'll do. Screenshot the page in case the language changes — vendors update warranty wording, and the version that applied at your purchase date is the version that governs your unit.

Locate your IceRiver shipment date. Not the order date, not your purchase date — the date IceRiver shipped to either you or your reseller. That's the start of the `180 days`. Calculate end-of-warranty before you commit time to an RMA path; if you're past it, the firmware question is moot and bench repair is the only path. If you bought from a reseller, the warranty responsibility transferred — your counterparty is the reseller, not IceRiver.

Document the unit's current state with photos and screenshots. Web UI dashboard, version string, OC profile if visible, front panel LEDs, chassis label with serial / model / hardware revision, any visible damage. Save these somewhere you won't lose them. If you eventually escalate the warranty claim past first-line denial, this is the evidence trail. If you go to D-Central instead, this is what we'll ask for first.

Decide the path before you start any reflash. Three options: (a) IceRiver warranty round-trip (only viable if covered + clean stock reflash possible + failure unrelated to OC), (b) reseller warranty (if you bought through one), (c) D-Central or comparable Western bench repair. Picking the path determines what you do next — don't reflash blindly without knowing where the unit is going.

If you're researching before flashing: stop here. Don't flash. The 50% hashrate gain from `xyys` / `tswift` is real, but so is the brick rate, and so is the warranty void. For an in-warranty unit, the math almost never works in the first `180 days`. Wait until warranty expires, then flash if you still want to. That's the discipline that keeps your warranty intact and gives you the OC firmware path later.

Download the matching stock firmware image from `iceriver.io/firmware-download/`. Match exactly: model (KS0 / KS0 Pro / KS0 Ultra / KS1 / KS2 / KS3 / KS3L / KS3M / KS5 / KS5L / KS5M) and hardware revision (printed on the chassis label). Wrong model bricks the unit. Wrong hardware revision within the same model also bricks it. If you can't find the exact revision, contact IceRiver support before flashing — `support@iceriver.io` is the published contact, response time variable.

Verify line voltage stability before any flash retry. Multimeter on AC at the wall under load. KS-series PSUs expect clean `100-240 V` AC input. If line voltage sagged below `190 V` during your previous OTA attempts, that's why a flash bricked — and trying again on the same circuit will brick it again. Move to a stable circuit, ideally UPS-protected, before any reflash. Residential evening-peak voltage sag is the silent cause of half the 'mid-OTA brick' tickets we see.

Reflash via the stock web UI firmware-update tab. Power on the unit, wait for the web UI to load (this assumes the OC firmware is at least partially functional). Browse to the firmware-update tab, upload the stock image, commit. Do NOT power-cycle, close the browser, or interrupt the network during the flash. The unit reboots once after a successful flash and returns at DHCP within `90 seconds`. If the unit is unreachable after the flash, the active partition was too damaged — proceed to Tier 3.

Factory reset for `20 seconds` after stock reflash. Power on, wait `60 seconds` for boot, then hold the reset button until the red status LED starts flashing. Release. Wait `5 minutes` untouched. This wipes any residual config from the OC firmware and ensures the unit boots clean stock. Verify the web UI shows the IceRiver `iMiner` interface, not any OC-firmware UI. If the OC UI is still showing, the reflash didn't take and you're in Tier 3 territory.

Run for 1 hour at stock V/F and document. Record hashrate, chassis temp, error codes if any, fan RPM. Print or screenshot the dashboard at the 1-hour mark. This is the 'unit is back to stock and stable' evidence for the RMA conversation. If the unit holds clean stock for 1 hour with no errors that wouldn't be present on a never-OC'd unit, you have the strongest warranty position the firmware-modification path will ever give you.

SD-card recovery boot when stock reflash via web UI fails. Download the official KS-series recovery image from `iceriver.io/firmware-download/` for your exact model and hardware revision. Write to a microSD card with `balenaEtcher`. Power off, insert SD, hold the reset button while powering on, keep holding for `5 seconds` after power-on, release. Bootloader detects the SD and reflashes from it over `3-8 minutes`. Don't power-cycle during this. See the IceRiver bricked-firmware recovery page for the full procedure.

Verify clean boot post-SD-recovery. Power off, remove SD, power back on, wait 2 minutes for full boot, check DHCP for the unit's MAC, log into the web UI, confirm IceRiver `iMiner` interface (not any third-party UI). Run for `30 minutes` at stock V/F. If clean, proceed. If the unit immediately shows OC-firmware UI again, the eMMC has corrupted partitions that survived the SD recovery and the unit is now bench-only — UART-level eMMC reflash is the next tier and that's a D-Central job.

Document the recovery attempt for the RMA file. Photo of the IceRiver `iMiner` UI post-recovery, hashrate stability log, factory reset confirmation. Even if IceRiver receiving fingerprints prior third-party firmware in the eMMC, your documented good-faith effort to restore stock is part of any escalation conversation. It rarely flips a denial alone, but it's not nothing.

Decide: warranty path or bench path, with the stock-reflash data in hand. A unit that boots clean stock after SD recovery, with no visible damage, with a failure that didn't reproduce on stock, and that's still inside the `180-day` window from a direct IceRiver purchase — file the warranty ticket. Anything missing from that list, the bench is the more reliable path. Don't sink another `$80-200 CAD` of round-trip shipping into a unit that's going to come back denied at receiving.

If SD recovery fails entirely and the unit is bench-bound either way, stop further DIY. UART-level eMMC reflash, control-board swap, or PMIC component-level repair are all bench operations. Document everything you've done so far, photograph the boards, ship to D-Central or comparable Western repair shop with a printed history.

Stop DIY and ship to D-Central when (a) you've flashed third-party firmware and the unit has bricked or developed a fault you can't recover at home, (b) visible PMIC damage or capacitor bulging is present, (c) SD recovery has failed twice and the unit still shows OC-firmware UI or won't boot at all, (d) you've cross-flashed (e.g. `xyys` for KS0 onto a KS0 Ultra) and the unit is now in bench-only territory, (e) the IceRiver warranty path is closed, or (f) the round-trip math to China doesn't work for the unit value.

D-Central bench process for warranty-voided KS units: UART access at `115200 8N1`, full eMMC dump and analysis, eMMC chip desolder + external programmer reflash for irrecoverable cases, control-board swap with image transfer when chip replacement is uneconomic, full PMIC continuity testing for over-voltage damage, hashboard component-level repair if the brick chained into a hashboard fault, and post-repair `24-hour` burn-in at stock hashrate before sign-off. Western retail bench, Canadian quality control, no shipping to China.

Ship the unit (or just the control board if you can pull it cleanly) in anti-static bags, double-boxed with `≥5 cm` foam on every side. Include a printed history: original firmware (stock, `xyys`, `tswift`, or fork + version), OC settings at time of brick if known (V/F values), failure trigger (mid-flash / commit hang / load reboot / progressive), recovery steps already attempted, highest tier reached, IceRiver warranty status (in-warranty / out-of-warranty / denied / not pursued). Saves D-Central 30+ minutes of diagnostic time per unit and that savings passes back directly in the repair quote.

D-Central can also handle in-warranty units the operator chooses not to round-trip. This is a real path — operators with KS5 / KS5L / KS5M units inside the `180-day` window who calculate the `2-4 week` round-trip + customs is worse than a 5-7 day Canadian bench repair at fixed CAD pricing. Both paths are legitimate. We don't ask why; we just fix the unit and burn it in.