The Bitmain APW3, APW3+, APW3++, APW5 and APW7 are the fixed-12V PSUs that powered the S9 generation and its siblings (S9i, S9j, T9+, L3+, D3, Z9). They are “dumb” supplies — no I2C, no firmware handshake, just AC in and a bank of 6-pin PCIe cables out at a nominal 12V. That simplicity means most faults come down to four things: input AC, protection latch, thermal shutdown, or a dead fan. This guide walks each symptom with the actual measurements and safety steps a self-repairing miner needs, then tells you when a unit is beyond a bench fix.
Know your PSU before you probe it
These are the older per-board-connector supplies — each hash board is fed by its own pair of 6-pin cables, unlike the bolted copper bus-bar on later APW12-era units. Nominal DC output is roughly 11.6–13.0V across the range. Rated output depends on input:
- APW3 / APW3+ / APW3++ — wide input (100–264V AC); ~1600W at 200–240V, derated to ~1200W on North American 120V; up to ~133A at high line, ~100A at 120V; ~93% efficient.
- APW5 — ~1500W, ~125A; happiest on 200–240V but will run derated on 120V.
- APW7 — ~1800W at 220V, derated to ~1000W on 120V; up to ~150A; ~95% efficient.
The takeaway for home miners: these supplies do run on a standard 120V outlet, they just deliver less power. A wide-input APW3++ on 120V simply throttles the miner to what it can pull from the wall — it does not fault on low voltage the way a strict 200–240V-only supply (APW12/APW17) would. Match your expected wattage to your input voltage before you diagnose a “weak” PSU.
Safety first: what actually shocks you
The 12V DC output side is not a shock hazard — you can handle the output cables and busbar all day. The danger lives on the primary side: the mains input and the large PFC/bulk capacitors, which hold roughly 400–420V DC for minutes after you pull the plug. Before opening a PSU shell: unplug AC, wait several minutes, and confirm the bulk caps are discharged with a meter before touching the primary board. Never probe the primary with the lid off and AC live.
Symptom 1: Fan does not spin, no 12V output
A completely dead unit almost always means the AC side is not being powered.
- Confirm the C13/C19 AC cord is fully seated at both ends and the outlet is live. Swap the cord — kinked or under-gauge cords are a frequent culprit.
- Measure wall voltage with a multimeter. Feed the PSU a stable source within its input range.
- With AC applied, measure DC across a 6-pin output (pin-to-pin, +12V to ground). Zero volts with a dead fan and good AC input points to an internal primary fault — open a repair ticket rather than chasing it blind.
Symptom 2: Fan spins normally, but no 12V output
Here the primary is alive but the secondary is not delivering. Two common causes:
- Input voltage below what the mode needs. If you are expecting full ~1600–1800W output, the supply wants line voltage in the 200–240V band; sagging or brown-out line can leave the secondary unable to regulate. Verify input is stable under load.
- Protection mode latched. A short or overload on the output trips protection. Disconnect the miner, inspect output cables and the hash-board connectors for scorching or bent pins, then power the PSU down for a full 10 minutes to let protection reset before retrying.
Re-test DC output with the miner disconnected. Nominal ~12V with no load but nothing under load usually means the fault is downstream (a shorted board), not the PSU.
Symptom 3: PSU cycles — runs a few seconds, stops, resumes
This stutter is classic high-temperature protection.
- Watch the fan. If it is not spinning while the PSU tries to run, the fan has failed and the unit overheats within seconds — treat it as a fan fault (see Symptom 4).
- Clear the cooling path: blow out dust from the intake, exhaust, and internal air duct. Caked dust is the number-one cause of thermal cycling on these supplies.
- Check ambient. These PSUs derate with heat; keep intake air within the temperature/power curve in the unit’s documentation. A hot, enclosed hashcenter corner will cycle a healthy PSU.
Symptom 4: Output is normal, but the fan does not spin
If DC output is fine yet the fan is dead, the fan itself has failed — and a PSU running without airflow will cook itself and trip Symptom 3 shortly after.
- Power down and check for debris jamming the blades; clear it and spin the fan by hand.
- If it still will not run, the fan (or its drive circuit) needs replacement. Do not run the PSU as a stopgap without cooling — open a repair ticket.
Common mistakes
- Blaming the PSU for a shorted hash board. A “no output under load” fault is often the miner pulling the supply into protection. Test the PSU unloaded before condemning it.
- Expecting 220V wattage from a 120V outlet. Reduced output on 120V is normal, not a fault.
- Probing the primary with caps charged. Discharge and verify first.
- Ignoring dust. Most thermal cycling is a cleaning job, not a board failure.
When to stop and send it in
A dead primary (no fan, good AC), no output with the load removed, or repeated protection trips after cleaning and a full reset all indicate an internal fault — blown input fuse, failed PFC stage, or a bad secondary. These are component-level repairs, not field swaps. Bench-test a known-good PSU on the same miner to isolate PSU vs. hash board before committing to a repair.
Related: Match symptoms to the whole miner with the ASIC fault finder, source a replacement fan or bench-test supply from ASIC repair parts, download the original PSU documentation in our manuals library, or hand the unit to our bench via start a repair.
