Mean Well LRS-600 Foldback Current Limit on Bitaxe GT

Power off everything at the wall before touching screw terminals. LRS-600 has bare M4 lugs; a wedding ring across `12 V` and ground melts faster than you can react. Remove all conductive jewelry. Use insulated tools throughout. The brick stores residual charge in its bulk caps for several seconds after wall power is removed — do not assume it's cold the moment the lights go out.

Pull the side label and locate the CV/CC slide switch on the LRS-600 housing. Flip from `CV` to `CC`. Power on. If the miner now boots and stays running (even with slightly lower hashrate from rail droop), you have confirmed PSU sizing is the only fault and inrush OCP was the trigger. This is a diagnostic checkpoint, not a permanent fix — plan a PSU upgrade.

Stagger startup of multi-miner rigs. If you have two or three Bitaxes on one LRS-600, plug one in, wait 15 seconds, plug the next in. Manual stagger eliminates simultaneous inrush. If this stops cycling, you have confirmed cumulative inrush as the trigger. Buy a sequencer (Tier 3) or split miners across separate PSUs.

Trim output voltage up `0.3 V`. The trim pot under the side label adjusts output ±10%. Bumping `LRS-600-12` from `12.0 V` to `12.3 V` gives the miner a tiny headroom against rail droop during inrush. Band-Aid only — gets you through tonight's hashrate report while you wait for a bigger PSU.

Reduce miner peak load. On Bitaxe / NerdAxe / NerdQAxe variants, drop the operating frequency by `50 MHz` and re-flash the firmware. Lower frequency means lower peak inrush and a better chance of staying below the OCP threshold. This is a workaround for under-sized PSUs that are committed-deployed and you don't want to replace this week.

Measure peak boot inrush. DC clamp meter on the `12 V` lead between PSU and miner. Power up cold. Note the peak reading during the first 500 ms of boot — most clamp meters peak-hold this. Log `peak_amps × 12 V = peak_W`. Compare to PSU continuous rating. If peak > `1.0× rated`, sizing is the fault. This is the single most useful number to have before you upgrade.

Inspect cable connectors. XT30, XT60, barrel jack, screw terminal — all show degraded contact resistance after 100+ plug cycles. A `30 mΩ` connector dropping `0.6 V` at `20 A` looks identical to PSU sag at the miner end. Replace any connector that's been re-mated more than twice this year, particularly on a PSU you've been swapping during diagnosis.

Add an inrush limiter inline. A `5 Ω`, `10 W` NTC thermistor (`CL-90` or equivalent) on the `12 V` line softens the initial current spike from full-step to a 200-300 ms ramp. This is a `$3` fix that buys you `~30%` more headroom on inrush without changing the PSU. Solder it inline on the positive lead, heatshrink the joints, mount it where it can dissipate heat.

Verify wall voltage. A wall sagging from `120 V` to `108 V` (or `240 V` to `220 V`) under load reduces the LRS-600's input headroom by the same percentage. If wall voltage is borderline, the brick has less margin to handle output inrush. Move to a different circuit if possible, or address the wall-side undersize separately.

Confirm 12 V vs 24 V vs 48 V model. Read the LRS-600 model number on the white label one more time. `LRS-600-12` is `12 V × 50 A`, `LRS-600-24` is `24 V × 27 A`, `LRS-600-48` is `48 V × 12.5 A`. A miner-builder who grabbed the wrong variant is a weekly call to D-Central's bench. If you bought used, double-check before suspecting anything else.

Scope the rail during boot. Trigger on the rising edge of the `12 V` line. You should see clean rise to `12 V` in `< 100 ms`, no collapse, no oscillation. If you see 300-600 ms collapse cycles, that's hiccup-mode confirmed. If you see clean voltage but droop to `9-10 V` for 50-100 ms then recovery, that's sag — a different failure mode that needs different treatment.

Replace LRS-600 with `RSP-1000-12`. Mean Well `RSP-1000-12` (`1000 W × 80 A`) is the home miner's standard step-up: still hiccup-mode (Mean Well's house topology) but with `66%` more headroom and a much better soft-start ramp. `~$160 CAD`. Wires the same way. Drops in. The single most-reliable upgrade path.

Upgrade to a constant-current variant. Mean Well `HEP-600C-12` (`600 W` constant-current) costs more (`~$120 CAD`) but uses CC topology natively, so inrush sees voltage droop instead of cycling. Better for marginal-sized rigs that don't want to upsize physical brick footprint. Verify CC ceiling matches steady-state load + 30% margin before deploying.

Move to a server PSU breakout. A used `HP DPS-1200FB` (`1200 W × 100 A`) on a `$15` breakout board gives you `2×` the LRS-600 headroom for `$30 CAD` total in cost. Server PSUs use foldback or current-limited topology depending on model — `DPS-1200FB` is current-limited (good for inrush) and is the home-miner standard for Bitaxe Hex / NerdQAxe / multi-Bitaxe builds.

Build a startup sequencer. For multi-miner rigs that want to keep the LRS-600 because it's already installed, an Arduino + relay board can stagger power to each miner with 5-second delays. `~$25 CAD` in parts. Eliminates simultaneous inrush as a failure mode permanently. Wire the relay coils through a watchdog so a stuck Arduino doesn't leave a miner in unintended state.

When to stop DIY. If you've verified PSU sizing is correct (peak load × 1.5 with 20% derating), upgraded the wiring, replaced connectors, and the miner *still* cycles — the fault is not on the PSU side. It's on the miner. Investigate firmware corruption, hashboard fault, or — on Bitaxe — TPS546 / BM1366 / BM1368 / BM1370 silicon failure separately.

What D-Central does at the bench. Test fixture with a programmable load + a precision PSU lets us replicate hiccup-OCP on demand and definitively isolate PSU-side vs miner-side faults. Bitaxe and Antminer hashboards both get their power-side caps and PMIC inrush behaviour characterized. We document the failure profile, ship a written diagnosis, and quote any miner-side repair from there.

Ship safely. PSUs go in their own box. Miners ship in anti-static bags, double-boxed with `≥ 5 cm` foam on every side. Include a note describing the exact cycling pattern, the PSU you tested with, the current load profile, and what you've already ruled out — saves diagnostic time and your final repair cost.