Bitaxe Hex – XT30 12V Rail Droop Under Load

Visually inspect the XT30 cable and Hex board socket. Disconnect cable, look at male pins under bright light: discolouration, pitting, gold flaking, sprung-out feel on insertion. Look down board-side female sockets: visible debris, sprung leaves, oxidation. Photograph anything that looks suspect - those photos are useful in any support ticket. A clear visual inspection before troubleshooting saves you from chasing electrical ghosts when the fault is mechanical. About 30% of XT30 droop tickets D-Central sees on the bench have visual giveaways at this step alone (counterfeit Amass logos, magnetic pins, or sprung-leaf socket fatigue).

Wipe the XT30 male pins with 99% isopropyl alcohol. Lint-free wipe, IPA, gentle wipe each pin face, let dry 30 seconds. This breaks any factory passivation oxide and removes finger oils. Do NOT use any abrasive (steel wool, scotch-brite, sandpaper) on the male pins - gold plating is microns thick and you will strip it instantly. Reinsert into the board-side socket firmly. Power up and test. This single step recovers a meaningful fraction of 'new Hex pulled from box' droop reports where the cause is just factory storage oxide rather than counterfeit metallurgy.

Verify PSU spec is correct. 12.0 V (NOT 5 V, NOT 9 V, NOT 19 V), 5 A minimum continuous current rating, and barrel polarity center-positive if the cable adapts to a barrel jack. Read the PSU label, confirm specs, and confirm it is a name-brand regulated supply (Mean Well, CUI, Mornsun, or a D-Central-validated Bitaxe bundle - NOT a generic AliExpress wall-wart). A wrong-spec PSU produces droop symptoms identical to XT30 contact-resistance droop; rule it out before fixating on the connector itself.

Power-cycle and re-measure Vin in AxeOS. Open the Hex web dashboard, look at the live Vin reading. Compare against the open-circuit DMM reading at the PSU output. If AxeOS Vin reads >200 mV lower than DMM PSU output under hashing load, you have a droop problem somewhere on the path PSU -> cable -> XT30 -> board -> TPS546. Note: AxeOS reports the board-side Vin as sampled by the ESP32-S3 ADC, accurate to ~50 mV when the divider resistors have not drifted - close enough for triage but do not trust it for absolute calibration.

Wiggle test under live measurement. With the Hex running and a DMM probe pair on the XT30 board-side terminals (or a DMM clipped via test leads), gently wiggle the XT30 cable side-to-side and pull-and-release. A connector with sprung leaves or partial contact will show the voltage reading jumping 50-200 mV during the wiggle. A solid connector reads rock-stable. This is the single fastest diagnostic for 'is the connector mechanically intermittent' without disassembly.

Swap to a different XT30 cable as a quick A/B. If you have a second known-good XT30 cable (or a backup Hex with its own cable, or a D-Central PSU bundle), swap and re-test. Droop disappears = original cable was the problem; replace it. Droop persists = the board-side female socket is the problem (Tier 2 work below). Cable swap is the cheapest decisive test in your bench arsenal.

Four-point voltage measurement under steady-state load. Hex running at full hash for 5+ minutes. DMM on mV DC scale. Measure: (a) PSU output terminals, (b) XT30 male plug pins on cable side, (c) XT30 female socket pins on board side, (d) TPS546 input pad. Document all four readings. Healthy: within ~15 mV of each other. Compute R_contact = (b - c) / I_load. The ratio of R_contact to your load decides which Tier 2/3 step to pursue. This four-point method is the gold standard for diagnosing where on the input chain the droop lives.

Thermal IR scan of XT30 connector body under load. IR thermometer or thermal camera on the connector body after 15+ minutes of full-load hashing. Healthy: within 5-8 C of ambient. Marginal: 15-25 C over. Bad: 30+ C over - replace urgently or solder-direct, that is contact-resistance heating in a positive-feedback oxidation loop. The IR scan also catches asymmetric heating (one pin hotter than the other) which signals a single-point contact failure that DMM averaging misses.

Clean the XT30 female socket cavities on the Hex PCB. Power off, fully disconnect. Spray 99% IPA into each socket cavity. Work a small clean brass brush (or folded 2000-grit sandpaper sliver) gently into each cavity to abrade oxide off the leaf contact surfaces. Compressed air or canned dust-off to blow out residue. Let dry 60 seconds. Reinsert the cable firmly. Re-run Step 7. About half the time, this restores a marginal connector to spec without further mechanical work.

Retension the XT30 female socket leaves. Under magnification (jeweller's loupe 10x minimum, USB microscope better), look down each socket cavity. Three brass leaves per cavity. If sprung outward, gently pinch each leaf inward 0.2-0.3 mm with a precision flat-blade jeweller's screwdriver or fine tweezers. Test fit: male pin should require firm-but-not-violent pressure to seat, with audible / tactile click. Re-run the four-point measurement. Aggressive over-tensioning damages leaves permanently - small movement only.

Replace the cable with a verified-genuine Amass XT30 or D-Central PSU bundle. If clean + retension did not resolve and the cable looks counterfeit, swap to a verified-genuine cable. D-Central's Hex PSU bundle ships with a tested-genuine XT30 + correctly-gauged 16 AWG wire + name-brand regulated 12 V / 5 A brick. Buying counterfeit twice is more expensive than buying genuine once. Reputable RC retailers (Amain Hobbies and similar) document Amass authenticity.

Verify wire gauge in the cable. Strip back 1 cm of the cable's outer insulation near the XT30 plug and look at the conductor: 16 AWG minimum for Hex (5 A continuous + headroom). If you see 20 AWG or smaller - the marketing-mistake 'Bitaxe-compatible' cables that are actually scrap-spec - replace the cable. 20 AWG over 30 cm adds ~10 mOhm per leg = 20 mOhm round-trip = ~50 mV droop at 2.5 A before the connector even contributes.

Solder-direct mod: bypass the XT30 entirely. For permanent stacks. Desolder the female XT30 socket from the Hex PCB: hot air at ~330 C, slow heating across all three solder pads (the XT30 has substantial copper plane connection - it sinks heat, you may need a preheater on the bottom side at ~150 C to bring the joint to flow temperature). Lift cleanly. Strip cable wires 5 mm, tin them with 60/40 or lead-free solder + flux. Solder directly to the through-hole pads where the XT30 sat: positive to the 12 V pad, negative to GND. Hot-glue strain relief at the cable entry to prevent flex stress on the new joint. Heat-shrink over the joint. Re-run four-point measurement: should be flat <5 mOhm PSU-to-TPS546.

Replace the female XT30 socket with a fresh genuine Amass part. If you want to keep the connector for future cable swaps, desolder the old (bad) female socket and solder a fresh genuine Amass XT30-PB through-hole part in its place. Verify alignment before reflow - the pin spacing is precise and a misaligned solder joint will lift on first cable insertion. Genuine Amass XT30-PB through-hole sockets are available from reputable component distributors (Digi-Key, Mouser) - beware of counterfeit listings on AliExpress / generic marketplaces.

Add a bulk capacitor at the TPS546 input. If your Hex sees frequent droop transients during chain init (chain-init pulls 5 A for ~50 ms), adding ~470 uF of low-ESR aluminium polymer or solid tantalum at the TPS546 input pad provides local reservoir charge that masks short droops without changing average behaviour. This is a band-aid, not a fix - it reduces symptoms but does not address the connector resistance. Use only as a temporary measure while a replacement cable is on order.

Scope the Vin rail at the TPS546 input during chain init. A 100 MHz handheld scope on Vin while the Hex boots will show the 5 A chain-init transient as a 100-300 mV notch. Healthy: notch recovers within ~10 ms and Vin settles to 12.00 V. Failing: notch is >500 mV deep and recovery is slow / oscillatory. The scope captures the connector failure mode that DMMs (averaged readings) miss. If the notch is bad, retension or replace the connector regardless of the steady-state DMM reading - chain init is the worst-case load and is where one chip drops off the chain.

Re-flash ESP-Miner-multichip after any rework. Hot-air work near the ESP32-S3 can briefly raise SPI flash temperature and corrupt config silently. After any Tier-3 rework, re-flash the correct multichip image via the Bitaxe Web Flasher and verify all six chips enumerate before declaring the repair complete. Verify Vin reads clean in AxeOS after the flash. Common failure mode: the hardware repair is good but leftover rework-induced flash corruption makes it look like the repair did not take.

Stop DIY and ship to D-Central if any of: solder-direct mod is outside your soldering comfort range, the female XT30 socket is mechanically damaged (cracked plastic body, lifted pad, snapped leaf), repeat counterfeit-cable failures suggest a deeper environmental problem (ground loop, shared PSU draw, electrical noise), or you want a warranty-documented mod. D-Central has done hundreds of XT30 reworks on Bitaxe Hex - we manufactured the original Hex heatsinks and we know this hardware better than anyone. Solder-direct mod with proper strain relief and post-mod four-point validation is CAD $35-75 on our bench, plus parts. Faster than ordering tools you will use once.

Document and ship to D-Central. Pack the Hex (in anti-static bag), include the XT30 cable that was on it, attach a note with: AxeOS firmware version, your tune (frequency / Vcore setpoint), measured droop values from your four-point measurement, IR scan results if available, and what Tier 1/2 steps you tried. The more diagnostic data you include, the faster we can confirm the root cause and the cheaper the repair. Canada-wide shipping; US/international welcomed. Turnaround 3-7 business days for a mod-only ticket, 5-10 business days for a full Hex diagnostic + mod.

Consider the D-Central Hex PSU + cable bundle for the replacement. If you have fought XT30 droop on multiple cables, the issue is upstream - the no-name supply chain you have been buying from. The D-Central bundle ships a pre-tested genuine Amass cable, name-brand regulated 12 V / 5 A brick, and matched-spec connectors that will not develop droop in their first six months on shelf. Pioneers of the Bitaxe Hex ecosystem; we would rather sell you the right cable once than rework counterfeit failures on bench three times.

Log the failure in your mining records. Note in your Hex log: date droop appeared, measured R_contact value, what fixed it (clean / retension / replace / solder-direct), what cable / PSU was involved, ambient conditions. If you run multiple Hex units, the pattern across the fleet tells you whether you have a systemic supply-chain problem or random one-off failures - and whether to standardize on a single validated cable + PSU spec going forward. Solo-mining lottery stacks deserve a logged maintenance trail.