How to Diagnose a Failing Antminer PSU: Symptoms, Tests & Replacement Guide

The #1 Preventable Killer of Antminer Hashboards

If you've spent any time repairing ASIC mining hardware, you've probably seen it: a perfectly healthy hashboard arrives at the workshop, dead. You open the unit, plug it into a test fixture, and the chips are fine. The real problem was upstream all along — the power supply was feeding dirty, drifting voltage into the board for weeks before it finally took out a chip.

A failing PSU is the single most preventable killer of Antminer hashboards. The good news: a 10-minute multimeter test is usually all you need to confirm whether the power supply is dying. This guide walks you through the symptoms to watch for, the exact test procedure our technicians use in our Shenzhen workshop, and how to pick a compatible replacement for every Antminer generation from the S19 through the S21 and T21.

Before You Touch Anything: A Safety Warning

Antminer PSUs run at grid voltage on the input side and hold large filter capacitors that can retain a lethal charge for several minutes after the unit is unplugged. This guide only covers diagnostic procedures performed on the DC output side of an unplugged, discharged PSU — never on live mains components.

If you are not trained to work on mains-side electronics, do not open the PSU cover. Every procedure below can be performed from the output connectors, without ever exposing yourself to AC hazards. When in doubt, the correct move is to replace the unit, not to dig into it.

The Four Most Common Antminer PSU Failure Symptoms

Before you reach for a multimeter, identify what you're actually seeing. Nearly every PSU failure we diagnose falls into one of four patterns.

Symptom 1: Miner Doesn't Power On At All

You plug the unit in, flip the breaker, and nothing happens. No fan, no lights, no beep. This is usually the input stage: a blown input fuse, a failed bridge rectifier, or a shorted primary-side transistor. Visually, you may see browning or scorching on the PCB near the input block, sometimes accompanied by a burnt smell.

If the miner plugs into a PDU shared with other working units, the PDU port is fine and the problem is in the PSU. If you're using a dedicated circuit, always verify the outlet with a known-good load before condemning the power supply.

Symptom 2: Fan Spins But There's No DC Output

This is the most common pattern we see, and the most diagnostic. The PSU fan starts — meaning the auxiliary power rail is alive — but the main DC bus never comes up, so the miner itself stays dark. The standby circuit is working, but the main converter stage has failed. Typical causes: a blown output MOSFET, a failed PWM controller, or a protection circuit that has latched itself off because of an internal short.

This is the case where a multimeter test confirms the diagnosis in under a minute.

Symptom 3: Miner Powers On, Then Cycles Repeatedly

The PSU comes up, the miner boots, hashrate starts climbing, and 30 seconds to 3 minutes later the whole unit resets. Then it happens again. This is almost always a thermal protection trip — the PSU is delivering power but overheating internally because the cooling fan has failed, dust has clogged the airflow path, or the heatsink-to-component thermal interface has degraded.

Less commonly, it's output current protection tripping because the hashboards are pulling more than the PSU can safely deliver — which usually means you also have a hashboard problem in parallel. If you see cycling on multiple miners that were recently moved to the same PDU leg, check your circuit loading before blaming the PSUs.

Symptom 4: Normal Output But Fan Not Spinning

Rarer but dangerous. The DC output is correct, the miner runs, but the PSU fan is dead. The unit will operate for a while, then either overheat into thermal shutdown or, worse, keep running long enough to cook its own components and destroy capacitors from the inside. Any PSU running with a silent fan should be pulled from service immediately — this is a fire risk on larger farms.

Physical Inspection: What to Look For Before the Multimeter Comes Out

A five-second visual check catches half of the failures before you waste time on electrical testing. With the PSU unplugged, look for:

• Swollen or leaking capacitors. Any capacitor with a bulged top or any sign of dried electrolyte around its base is end-of-life. A PSU with swollen caps cannot be trusted on the DC rail — replace it.
• Burnt smell. Distinct, acrid, chemical. If you smell it when you open the miner cover, the PSU is the first suspect.
• Browning or discoloration on the PCB. Brown halos around components indicate sustained over-temperature events. The component may still function but has been stressed beyond its design envelope.
• Melted or discolored DC output connectors. The PSU-to-hashboard power cables run at very high current. If the contact points at either end show browning or melted insulation, the connector has been running hot — reseat, inspect for pin damage, and replace the cable if there's any doubt.
• Scorching at the fan housing. Indicates the fan failed and the PSU cooked itself around that area.

If any of the above are present, the decision is already made: replace the unit. Do not attempt to run it.

The Multimeter Test Procedure

If the PSU looks clean but you suspect it, this is the exact procedure we run on the bench.

Step 1: Prepare the PSU for Testing

Remove the PSU from the miner. Disconnect all DC output cables from the hashboards. Set your multimeter to DC volts on the 20V range. Reconnect the PSU to mains power. The PSU fan should spin within 2–5 seconds — if it doesn't, you're dealing with Symptom 1 and no further electrical testing is useful. Replace the unit.

Step 2: Measure Idle DC Output at J6 (APW12 Series)

With the PSU powered but unloaded, probe the main DC output terminal — on the APW12 series this is the J6 block. Red probe to the positive pin, black to negative. You should read between 12.1V and 12.50V at idle on a healthy APW12. Anything below 11.5V or above 13.0V is a red flag. A deviation greater than 0.5V from the nominal 12.1V–12.50V window means the output regulator has drifted and the PSU is degrading.

If you get no reading at all — zero volts or a flickering, collapsing value — the main converter stage is dead. This confirms Symptom 2.

Step 3: Measure Under Load

Idle readings lie. A PSU with degraded filter capacitors will often produce a clean 12.2V with no load, then sag dramatically the moment the hashboards start drawing current. To test properly, reconnect the PSU to the miner, boot the unit, and probe the DC output again while the miner is hashing. Healthy APW12 output under load should stay rock-stable around its rated set point. If you see voltage sagging more than 0.3V under load, or wandering up and down visibly, the PSU is on its way out.

Use backprobing on the PSU-to-hashboard connector or, safer, clip your probes onto the bus-bar screws with insulated leads before powering the unit on.

Step 4: Check the Adjustable Voltage Rail (Applicable Variants)

The APW12 series is adjustable between 12V and 15V through its PIC-controlled feedback loop. Maximum rated output is 12V/300A or 15V/240A. If your miner's control board is commanding a higher voltage (e.g., for aggressive overclock profiles) and the PSU is not responding, the PIC feedback signal path has failed. This is diagnosable only with the unit installed in a miner — if your bench test showed a healthy static 12V but the miner still shows "PSU undervolt" in its logs, the adjustable rail has failed while the static rail is still alive.

Step 5: Cross-Check Against a Known-Good PSU

If the test procedure is ambiguous, swap in a known-good PSU of the same model and run the miner for 15 minutes. If the miner behaves correctly, the original PSU is the problem. If the miner still misbehaves, you're looking at a hashboard or control board fault and the PSU was innocent all along. This swap test is the cheapest way to save yourself hours of diagnostic time.

PSU or Hashboard? How to Tell Them Apart

The two most common misdiagnoses in our workshop are calling a dead PSU a dead hashboard, and vice versa. Here's the quick decision rule we teach new technicians:

• If multiple hashboards fail simultaneously on the same miner, suspect the PSU first. Three chips blowing at once across three independent boards is almost never a coincidence — it's a voltage spike coming from upstream.
• If one hashboard fails while the other two are healthy, suspect the hashboard. Each hashboard has its own power distribution, and an isolated failure points to a local problem.
• If the miner cycles without ever reaching stable hashrate, suspect the PSU. A failing hashboard usually still lets the miner boot and run degraded.
• If hashrate is stable but one board shows reduced chip count, it's the hashboard. The PSU is delivering clean power — the problem is downstream.

Choosing the Right Replacement PSU

Once you've confirmed the PSU is the problem, the next question is which replacement to stock. Antminer PSUs are not fully cross-compatible across generations, and even within the APW12 family there are sub-variants that matter.

Within the APW12 12V–15V family, there are two sub-groups that are commonly confused: APW121215a/b/c (no voltage feedback) and APW121215d/e/f (with voltage feedback). These are not fully interchangeable. A control board that expects voltage feedback will report "PSU comm error" if you install a no-feedback variant — and vice versa, a non-feedback control board will ignore a feedback-capable PSU's signaling entirely. Always match the sub-variant to your specific Antminer revision before ordering.

We stock the APW12 12V–15V for S19/S19 Pro/T19, the full APW12, APW17 and APW11 PSU collection, and the Avalon 2-in-1 PSU for Canaan hardware. Every PSU in our inventory is bench-tested under load before it ships.

When to Repair vs When to Replace

The honest answer for most operators: replace the PSU, don't repair it. Unlike hashboards, where a $30 chip swap can bring a $1,000 board back to life, PSUs rarely justify the labor. Component-level PSU repair requires mains-side safety training, an isolation transformer, and specific replacement parts that are hard to source reliably. Unless you run a repair workshop with a dedicated PSU bench, you'll spend more in technician time than a new unit costs.

The exceptions: a dedicated PSU repair bench with volume (50+ units per month makes the economics work), a unit with a trivially identifiable failure like a blown input fuse on an otherwise cosmetically perfect PSU, or a rare / discontinued model where replacements are genuinely unavailable.

For everything else, the correct answer is to pull the failing unit, swap in a tested replacement, and send the miner back to work in under an hour.

Frequently Asked Questions

What voltage should an Antminer APW12 PSU output?

A healthy APW12 should output between 12.1V and 12.50V at idle on the DC output side (J6 terminal). Under load, the voltage should stay stable within 0.3V of its set point. A deviation greater than 0.5V from the nominal range indicates degradation and the PSU should be replaced.

Is APW12 interchangeable with APW17?

No. The APW17 is designed for the S21 series, S19j XP and newer platforms and has different connectors, output capacity, and communication signaling than the APW12. Always match the PSU model to your specific Antminer generation.

Why does my Antminer keep restarting every few minutes?

Repeated cycling is almost always a thermal protection trip — the PSU is delivering power but overheating internally because the cooling fan has failed or airflow is obstructed. Less commonly, it's output current protection caused by hashboards drawing excess current. Check fan operation and dust buildup first.

Can I test an Antminer PSU with a standard multimeter?

Yes. Set the multimeter to DC volts (20V range), probe the J6 output terminal on an APW12 with the PSU powered but no load, and verify 12.1V–12.50V. Then repeat the test under load by running the miner and measuring at the DC bus while it hashes. A standard digital multimeter is all you need — never probe the AC input side unless trained.

How do I know if the problem is my PSU or my hashboard?

Multiple hashboards failing simultaneously on the same miner points to the PSU. A single hashboard failing while the others remain healthy points to the hashboard. If the miner cycles without ever reaching stable hashrate, suspect the PSU first. The cheapest confirmation is to swap in a known-good PSU of the same model and observe the result.

Order a Tested Replacement PSU from LYS Shenzhen

Every power supply in our catalog is bench-tested under load before shipping. We stock the APW12 12V–15V for S19/S19 Pro/T19, the complete APW12, APW17 and APW11 PSU collection for every Antminer generation, and the Avalon 2-in-1 PSU for Canaan hardware — all with worldwide DDP shipping.

Browse the full Antminer replacement parts catalog, or contact our team directly at contact@lys-sz.com or via WhatsApp for bulk pricing, sub-variant matching on APW12 revisions, and custom quotes on hard-to-find PSUs.