A five-figure machine arrives in a plain cardboard box, and the first thought is rarely exciting. It is a worry. Did it survive the trip? Will every hashboard fire up? Is one of the chips already dead? That worry is not paranoia. Miners that fail almost always fail in the first hours or days, and the buyer who plugs one in cold is the one who finds out the hard way.
Burn-in testing exists to move that discovery to the seller’s bench instead of your living room. It is the difference between a miner that was verified before it shipped and one that was boxed straight off a pallet and hoped for the best. Here is what burn-in actually is, why new miners fail early, how a proper test runs, and how you confirm your own machine the day it lands.
What is burn-in testing for an ASIC miner?
Burn-in testing is running a miner at full load for an extended period, usually 24 to 72 hours, before it ships, so any early defect is forced to surface while the machine is still with the seller. It catches dead hashboards, cooling faults, and hashrate that falls short of the rated number, the failures that would otherwise land on the buyer. A quick power-on proves the lights come on. Burn-in proves the machine can hold full performance for days without faltering.
That distinction is the whole point. Every genuine seller should be able to say exactly how each unit is tested before it leaves, and that testing is a core part of
how we build ASIC miners. A miner is not a phone; it runs flat out, around the clock, generating heat that would cook lesser electronics. Verifying it can do that before you rely on it is not a luxury, it is the baseline.
Why do new ASIC miners fail early?
Because hardware that is going to fail usually fails early. Reliability engineers describe this with a bathtub curve: failure rates are high when a machine is brand new, drop to almost nothing through its useful life, then climb again years later as parts wear out. The early spike is the dangerous one for a buyer, and it comes from things no spec sheet shows.
The first hours are when problems show up
A weak solder joint that cracked in transit, a chip that was marginal from the factory, a fan that seized, a loose power connector: these do not announce themselves. They show up as a hashboard that never comes online, temperatures that climb past safe limits, or a hashrate that sits stubbornly below the rating. Most modern Bitcoin ASICs throttle above 85 to 90 degrees Celsius and risk permanent damage past 105, so a cooling fault is not cosmetic, it shortens the machine’s life. Left undetected,
a dead hashboard means paying full price for two thirds of a miner.
Shipping is a bigger culprit than most buyers expect. A miner is a dense, heavy stack of boards that spends days on trucks and in warehouses, absorbing vibration and temperature swings the whole way. A joint that was solid on the factory bench can fracture in transit, and a connector can work loose. This is why a machine that was tested before it was packed still deserves a check on arrival: the last leg of the journey is its own hazard.

How does burn-in testing actually work?
A proper burn-in is not one test, it is a sequence, and every unit runs the whole thing. Skipping a stage defeats the purpose, because each stage catches a different class of fault. The sequence runs like this:
- Incoming inspection. A visual and serial check for shipping damage, tampering, and that the unit matches what it claims to be.
- Power on and firmware. Boot the machine, confirm it reaches the dashboard, and verify it is running clean, official firmware, not mystery software.
- Extended full-load run. Run the miner at its rated load for 24 to 72 hours. This is the burn-in itself, the stretch that separates a healthy unit from a marginal one.
- Hashrate verification. Confirm the machine holds its rated terahash across the full window, checked against the pool, not just at the first boot.
- Thermal checks. Watch chip temperatures across every hashboard. They should stabilize within 30 to 60 minutes and stay even; a board running 10 degrees hotter than its siblings signals a cooling problem.
- Pass or fail. A unit with stable hashrate, clean error logs, and even thermal passes and ships. Anything that stumbles is pulled and routed to repair.

The hashrate step matters more than it sounds. A miner can boot, look healthy, and still deliver less than
its rated J/TH efficiency, which quietly costs you output for the life of the machine. Burn-in is where that shortfall gets caught, measured over hours rather than guessed at from a thirty-second boot.
What does a proper burn-in test check for?
The value of burn-in is in what it measures over time, not in a single reading. A thorough test watches for all of the following, and a fault in any one is grounds to pull the unit:
- Sustained hashrate. The machine holds its rated terahash across the full run, not just at the first boot. On an Antminer S21 XP that means a steady 270 TH/s, not a figure that sags as the unit heats up.
- Even thermals. Every hashboard reaches a stable temperature within the first hour and stays close to its neighbours, with no board running hot.
- Fan and cooling health. All fans spin at the right speed and respond to load, and on hydro units the coolant loop holds pressure and temperature.
- Clean error logs. No repeated hardware errors, chip dropouts, or restarts over the test window.
- Stable power draw. The unit pulls the expected wattage steadily, without spikes or sag that point to a failing power supply.
None of these checks means much in isolation or for a single moment. A miner can pass a five-second glance and still fail on the third day, which is the entire reason the test runs for hours rather than minutes. The record of how a unit behaved over a long, hot run is the real product of burn-in, and it is what lets a seller stand behind the machine rather than simply hope it holds.
Does skipping burn-in make grey-market miners risky?
Yes, and this is the clearest line between a trustworthy supply chain and a gamble. Grey-market and unverified sellers move machines fast and cheap precisely because they skip the slow, expensive step of testing every unit. The savings are real, and so is the reason for them: you are buying an untested machine and taking on the early-failure risk yourself. This is the same risk covered in
authorized versus grey-market miners, and burn-in is the single practice that most separates the two. A seller who tests every unit has already absorbed the failures you would otherwise inherit.How do you check if your new ASIC miner is faulty on arrival?Even with a tested machine, you should verify it yourself when it lands, both to confirm it survived the last leg of shipping and to start your own record. You do not need a lab. Before you commit it to a pool or a rack, run through this on arrival:
- Inspect the box and unit for shipping damage, dents, rattling, or loose parts inside.
- Power it on a suitable circuit and confirm it boots to its dashboard without errors.
- Check that every hashboard is detected. A machine reporting two boards when it should have three has a dead board.
- Let it run for at least an hour and confirm the hashrate climbs to its rated figure and holds.
- Watch the temperatures level off and stay even across boards, with all fans clearly spinning.
If all of that holds, the machine is almost certainly sound and you can
set up your ASIC miner for real. If something is off, catching it in the first hour, while it is clearly a hardware fault and not something you caused, is exactly what protects your warranty claim.
Ready to Start Mining?
Free worldwide DDP shipping. Professional hosting from $0.055/kWh.
What are the signs of a faulty ASIC miner?
A faulty unit usually gives itself away quickly if you know what to watch. The common signs are a hashrate that sits well below the rating and will not climb, a hashboard that is missing from the dashboard entirely, temperatures that keep rising past the point where a healthy machine levels off, fans that do not spin or run at full scream constantly, and a miner that restarts on its own every few minutes. Any one of these on a new machine points to a hardware problem rather than a setup mistake, and the sooner it is documented, the stronger the case for a replacement or repair.
The timing is what tells you the most. A fault that appears in the first hour, before you have changed any settings or joined a pool, is almost certainly a hardware problem the seller or manufacturer should own. A problem that surfaces weeks later, after configuration changes and real running, is murkier. That is why the arrival check carries so much weight: it timestamps the machine’s condition the moment it reaches you, while responsibility for any fault is still clearly upstream.
When a miner fails burn-in testing
On our testing floor, a unit that fails does not get quietly shipped with a discount. It is pulled and diagnosed: the specific hashboard, chip, fan, or power stage at fault is identified, and the machine is repaired and re-tested, or set aside. That is the entire point of testing before shipping, the failures stay with the people equipped to fix them. For a machine already in the field, the same logic applies through the manufacturer or supplier, which usually means
opening a repair ticket rather than living with a weak unit. A miner that clears the process, by contrast, is ready to be
deployed and brought online with confidence that it will hold.
Should you burn-in test a used or grey-market miner yourself?
Absolutely, and more carefully than you would a new one. A used or grey-market machine has no seller guarantee behind it and may have been run hard, so the arrival checks above become non-negotiable, and you should extend the run to a full day or more rather than a single hour. Watch closely for a hashrate that drifts down as the machine heats up, which points to aging chips, and for any board that runs hot. The reassurance you get for free from
a seller you can actually trust is exactly the work you have to do yourself when buying from one you cannot.
The bottom line
Burn-in testing is the difference between a miner that was proven to work and one that was merely shipped. It moves the early-failure risk, which is where almost all miner failures live, from your hands back to the seller’s, and it is the practice that most cleanly separates a genuine supplier from a grey-market flipper. When you buy, ask how every unit is tested before it leaves, and when a machine arrives, verify it yourself in the first hour. A miner that clears burn-in and passes your arrival check is one you can rack, or hand to
ASIC miner hosting, and then mostly forget about, which is exactly what a reliable machine should let you do.
Frequently asked questions
What is burn-in testing for a Bitcoin miner?
Burn-in testing is running a miner at full load for an extended period, usually 24 to 72 hours, before it ships, to force any early defect to surface. It catches dead hashboards, cooling faults, and hashrate that falls short of the rating while the machine is still with the seller, not after it reaches the buyer.
How long should an ASIC miner burn-in test run?
Typically 24 to 72 hours. A single hour confirms the machine boots and hashes, but early failures often appear only after the unit has been under sustained heat and load for many hours. For used or grey-market machines, a full day or more is wiser, since there is no seller guarantee behind them.
How do I know if my new ASIC miner is faulty?
Power it on, confirm it boots without errors, and check that every hashboard is detected. Let it run at least an hour and confirm the hashrate reaches and holds its rated figure, that temperatures level off evenly across boards, and that all fans spin. A missing board, a low hashrate, or climbing temps signal a fault.
What temperature is too hot for an ASIC miner?
Most modern Bitcoin ASICs begin throttling performance above roughly 85 to 90 degrees Celsius and risk permanent damage above about 105. In a healthy machine, chip temperatures stabilize within 30 to 60 minutes and stay even across hashboards. A board running about 10 degrees hotter than its neighbours points to a cooling problem.
Why do new ASIC miners fail in the first few days?
Hardware failures follow a bathtub curve: they are most likely when a machine is brand new, then rare through its useful life. Early failures come from factory defects, marginal chips, or shipping damage such as cracked solder joints and loose connectors. Burn-in is designed to trigger these before the machine reaches you.
Is burn-in testing the same as a quick power-on check?
No. A power-on check confirms the machine turns on and reaches its dashboard, which takes seconds. Burn-in runs the miner at full load for many hours to prove it can sustain performance and stay cool over time. Many faults only appear under sustained load, so a power-on check misses them entirely.
Do grey-market miners get burn-in tested?
Usually not, and that is a large part of why they are cheaper. Skipping testing lets sellers move machines fast, but it hands the early-failure risk to the buyer. A seller who burn-in tests every unit has already absorbed the failures an untested machine would pass on to you.
What happens if a miner fails a burn-in test?
A responsible seller pulls the unit, diagnoses the faulty hashboard, chip, fan, or power stage, and repairs and re-tests it or sets it aside, rather than shipping it. The failure stays with the people equipped to fix it. For a machine already in the field, the equivalent is a warranty or repair claim through the supplier.
Can I burn-in test a miner at home?
Yes, at a basic level. Run the machine at full load and watch its dashboard for sustained hashrate, even and stable temperatures, all fans spinning, and clean error logs over several hours. You will not have lab equipment, but the miner’s own web interface reports enough to catch the common faults.
Does burn-in testing reduce a miner’s lifespan?
No, in any meaningful sense. A day or two of running is negligible against a lifespan of five to seven years with proper cooling and maintenance. The small amount of runtime is a worthwhile trade for confirming the machine is sound before you depend on it, and it does not shorten a healthy unit’s life.
Images: technician photo CC0 (Derrick Coetzee), circuit-board photo public domain (Leon Brooks), both graded by MillionMiner. Diagrams original.