Mining Academy
Buying a used ASIC: how to inspect it before you pay
Before buying a used ASIC, check three things: real hashrate under load, hashboard condition (the number of working chips) and temperatures. The unit must be powered up on the spot, its logs reviewed, and any hidden faults ruled out, since these are what turn a bargain into a string of repairs.
Why buy used at all, and where the risk lies
A used ASIC is usually much cheaper than a new one per unit of hashrate, and with a low electricity tariff that can make sense. But the saving has a cost. The unit may have spent years in a hot room, survived overclocking, overheating, power surges and clumsy repairs. Some wear is invisible from outside, and sellers do not always describe condition honestly. Inspection at purchase is therefore not a formality; it is the main way to protect your money.
For SHA-256 and Scrypt the inspection principles are the same, only the benchmarks differ: for Bitcoin units compare TH/s and J/TH; for Scrypt units, which mine Litecoin and Dogecoin together, compare GH/s and power draw in watts. You do not need a separate device for Dogecoin: on RuPool a Scrypt ASIC connects to the LTC stratum and DOGE is credited automatically as an additional coin.
If you have not chosen a model yet, start with how to choose an ASIC: older generations depend heavily on cheap electricity, and savings on the device are easily lost on the power bill.
What to check before meeting the seller
- Exact model and revision. Names can be similar while hashrate and consumption differ significantly. Ask for a photo of the label and a screenshot of the status page.
- History. Where and how long it ran, whether it was overclocked or repaired, and why it is being sold. Answers like "just don't need it" deserve follow-up questions.
- Seller reputation. Reviews, listing history, the option to return the unit. You can browse listings on the used equipment market page, for example, but checking the seller and the unit is still up to you.
- Completeness. Power supply, power cables, mounting hardware. A missing original PSU is a cost to build into the price.
- Return terms. If you buy remotely, agree beforehand on an inspection period after delivery.
Inspection at startup, step by step
Ideally you inspect and start the unit together with the seller, or accept it with a right of return. The minimum procedure is:
- Visual inspection. Look for corrosion, oxidation, overheating marks (darkened PCB), liquid traces, dents in heatsinks, bent fins, non-standard screws and signs the case was opened. Ask the seller to open the cover and show the boards.
- Fans. Check for noise, vibration and play. Fans are relatively cheap, but replacing them means extra cost and downtime.
- Power-up and boot. Connect power and network, wait for boot and open the web interface. The unit should detect all hashboards and settle at a stable working hashrate.
- Pool connection and 1-2 hours of runtime. Do not settle for a minute on the bench. Real problems show up under load and as the unit warms up.
- Comparison with the datasheet. Real hashrate should be close to the model's nominal value.
Hashboards and chips: what matters most
An ASIC consists of several hashboards (usually three), each carrying dozens or hundreds of chips. If part of the chips are dead, the unit will deliver reduced hashrate while power draw stays almost the same. In the web interface (the board status page) look at:
- Number of detected boards. All must be present; a "missing" board is a common reason for a discount, but not always repairable.
- Chips per board. This must match the model's specification. Missing chips indicate damage or wear.
- Chip status. The log and status page often show dead or unstable chips.
- Frequency and voltage. A very low frequency on a board may mean the firmware lowered it because of problems.
- Per-board hashrate. Boards should give similar values; a big spread signals trouble.
Ask separately whether any boards were repaired. A repaired board is not necessarily bad, but the seller must say so and the price should reflect it. You can look for a workshop for diagnostics in the ASIC repair section.
Logs and temperatures
The event log (kernel or system log on the device page) says a lot about an ASIC's past. Look for:
- Repeated messages about chip errors, lost communication with a board, overheating or undervoltage.
- Frequent reboots and protection triggers.
- Errors related to fans or temperature sensors.
Check temperatures on both boards and chips. For a running ASIC in an ordinary room, values within the range given by the manufacturer are normal; if after warm-up temperatures approach the protection threshold and fans go to maximum, cooling or thermal paste may be in poor shape. A big temperature difference between boards is also a warning sign. If the unit went into protection during your check, do not accept "it's just hot in here" as an answer: check the air temperature first.
Runtime and remaining life
Exact runtime is hard to establish: ASICs generally have no odometer. But indirect signs help: production year and batch, condition of the boards and heatsinks. The uptime shown in the interface does not help here: it counts only since the last boot, not lifetime. Heat-cooling cycles also wear a device, and an ASIC that spent two years in a data centre with good cooling may be healthier than a "younger" one that lived in a hot garage. Ask where exactly the unit stood and how it was cooled.
Power supply, firmware and setup pitfalls
The power supply is a frequent weak point in used units. It runs near its limit almost constantly, and its failure causes random reboots and boards dropping offline. Ask whether the PSU is original, whether it has repair marks, and listen for humming or whining under load. If the PSU is separate, make sure its power and connectors fit your mains and PDU.
Check the firmware separately. Many sellers install third-party firmware that changes frequencies and operating modes. This is not necessarily bad, but you should know which version is installed, what its settings are and whether you can return to stock. Non-standard firmware can hide real wear if a board runs at a reduced frequency. After purchase, reset the unit to factory settings and update firmware from the manufacturer's official sources.
Finally, look at network behaviour: the unit should get an address by DHCP or accept a static one, open the web interface and keep a stable connection to the stratum. A flaky network port or weak control board is another hidden defect that only shows during long runs.
What to do right after purchase
- Connect the ASIC to stable power and network and change the web interface password.
- Enter the stratum address and, in the user field,
subaccount_name.worker_name; the password can be anything. Give each device its own worker name so you can see which one dropped. - Enable monitoring: the watcher link shows statistics without a password, and the mobile apps help you notice a hashrate drop quickly.
- For the first day or two, watch hashrate and temperatures and compare with what you saw during inspection. If it does not match what was promised, raise a claim before the return period ends.
How much to pay without miscalculating
Judge the price of a used device together with your electricity tariff and expected service life. Build the calculation in the profitability calculator: enter the unit's real hashrate (not the advertised one) and your price per kilowatt-hour. Add a reserve for repairs and for replacing fans and the power supply. If the gap to a new unit nearly disappears, buy new with a warranty. A calculator result is a calculation on current parameters, not a guarantee of income: price, network difficulty and electricity cost all change.
Deal safety
- Do not send a full prepayment to an unknown seller, especially in a remote deal.
- Sign a purchase agreement describing model, serial numbers and condition.
- Keep correspondence and photos of the device.
- If the purchase goes through an intermediary or platform, confirm the return terms.
- Before putting it to work, reset to factory settings and set your own passwords and pool details: the previous owner may have left their own settings.
FAQ
How can I tell an ASIC is "dead"?
Typical signs: hashrate well below nominal, missing boards or many missing chips, constant errors in the log, overheating at normal room temperature, frequent reboots.
Can I check an ASIC without powering it on?
Only visually. To assess real condition the unit must be powered on and run under load. If the seller refuses to power it on, be cautious.
Is it worth buying an ASIC after repair?
It can be acceptable if the repair is honestly disclosed, done by a qualified workshop and documented. The price should reflect it.
Do I need a warranty on a used ASIC?
Ideally at least a short warranty or an inspection period after delivery. It does not replace your own checks, but it protects against hidden defects.