A used ASIC needs a review of the actual machine, its operating costs and the sale terms. Comparing its advertised price with a new unit is insufficient. A high instantaneous hashrate can coexist with downtime, errors or site requirements you cannot meet. Connect evidence to the specific equipment and keep unknowns visible before making a financial commitment.
Illustrative calculations below do not value a real machine or guarantee future performance.
BTKWIN.COM · EDUCATIONAL EXAMPLE1. Match identity to the unit being offered
Record the complete model, revision, serial number and included components, including a power supply or cooling accessories where relevant. Compare them with the seller’s invoice or ownership documentation while excluding unnecessary personal information. A product-family name and a generic store photograph do not establish the identity of a particular unit.
Connect the serial number with dated test evidence or an appropriate independent inspection. A video showing a dashboard and then a similar machine leaves a gap if their identity cannot be linked. That gap does not establish fraud, but it means the claimed performance remains unverified for the offered unit.
Confirm the intended algorithm and network. An ASIC designed for one kind of computation does not become suitable for every cryptocurrency because it has networking and mining software. Hashrate figures across different algorithms are not a universal performance ranking. Asset publicity does not establish equipment compatibility or expected output.
2. Request sustained performance evidence
Preserve the test duration, operating mode, firmware, temperatures, power, hashrate, errors and interruptions. A screenshot taken shortly after startup may not describe settled operation. No single test duration proves every machine’s remaining life; the test should answer a defined technical question and disclose what was not examined.
Where available, compare local measurements with accepted pool contribution over the same period. Short estimates can fluctuate, and raw submission counts can differ in difficulty. Record rejected work, connection time and counter resets instead of interpreting the largest count as the strongest sustained contribution.
A fictional 150 TH/s machine may display that rate during filming but operate only eighteen hours of a full day. If it is inactive for the other six, its daily average before other factors is 112.5 TH/s. The instantaneous figure cannot support an uninterrupted-day estimate without disclosing the downtime.
3. Measure energy and review site requirements
Request a power reading with a defined boundary, such as input to the complete miner, plus units and operating conditions. An internal estimate or maximum power-supply rating does not necessarily describe wall consumption. Verify the model’s electrical and cooling requirements and arrange a suitable technical assessment of the intended site.
Consider two fictional devices each producing 150 TH/s, one at 3,300 W and another at 2,700 W. Their efficiency ratios are 22 and 18 J/TH. The 600 W difference becomes 14.4 kWh over 24 continuous hours. At a hypothetical 0.10 per kWh, it costs 1.44 daily or 43.2 over thirty days.
If the purchase-price gap is 400 currency units, dividing it by that daily energy difference gives approximately 278 constant-operation days. This is a comparison of one assumed saving, not project payback or proof of a better purchase. Repairs, availability, infrastructure, revenue and residual value may differ.
4. Check temperature and cooling under load
Identify each sensor and its meaning. A chip temperature is not interchangeable with an intake-air temperature limit. Review fan or liquid-system condition according to the actual design and preserve thermal errors or shutdowns from the test. A universal temperature threshold cannot replace model-specific documentation.
Include external cooling and ventilation in site cost. Their power and requirements can vary by season. A location suitable for a smaller load may not support another machine unchanged. Do not block airflow to reduce noise or bypass protective controls to keep a demonstration running. A successful display is not evidence of installation suitability.
5. Review maintenance history and condition
Ask for repair dates, replaced parts, reasons and known exposure to water, corrosion, dust or problematic transport. Missing records do not prove damage, but increase uncertainty. Distinguish seller statements from independent observations. A clean exterior does not demonstrate that components were never repaired or that internal condition is sound.
Internal or electrical inspection belongs with appropriately qualified personnel following relevant procedures. The inspection report should identify observations, test limits and unresolved items. It should not promise health through an assumed future lifetime. Include diagnosis, labour and transport rather than valuing an entire repair from one advertised spare part.
6. Verify software and ongoing support
Record installed firmware, its source and operating mode. Check the supported process for reconfiguration or updating when ownership changes. Do not assume an unknown seller’s file is appropriate, or that every software modification preserves compatibility and warranty conditions. A performance claim should identify the configuration that produced it.
Review repair availability and model support separately from present operation. BITMAIN publishes an end-of-service list, illustrating why support for older equipment needs checking. A machine working today does not establish that the manufacturer will repair it later.
An independent repair option also needs its own evidence of capability and compatible parts. Estimate shipping to a repair center, waiting time and lost productive time when relevant. A listed replacement component does not by itself establish authenticity, compatibility or the final cost of restoring the unit.
7. Put warranty, shipping and return terms in writing
Verify warranty status using the serial number and official channels, including when coverage began, what it includes and who can claim after resale. The BITMAIN warranty policy links coverage to official shipment and contractual terms. A used purchase does not automatically start a new manufacturer warranty.
Specify responsibility for packaging, shipping, insurance, transit damage and returns. Record the evidence required if the received machine differs or fails an agreed test. A bare word such as guaranteed does not identify the remedy, conditions or responsible party. The actual written transaction matters.
If a material item cannot be verified, preserve its uncertainty and the cost of obtaining independent evidence. Do not convert an unknown into a pass simply because the other checks look favorable. A low asking price can coexist with unresolved condition or expensive delivery requirements.
Build a cost sheet that follows the evidence
A hypothetical 1,000-unit advertisement plus 100 shipping, 150 preparation and 200 known repairs requires 1,450 before operation. If repairs are uncertain, show separate scenarios rather than treating an estimate as fact or excluding it entirely. Date each source and distinguish upfront, recurring and contingent spending.
Use the payback guide to test full costs and cash flow. Even a successful inspection does not guarantee profit: difficulty, asset price, pool rules, electricity and later faults remain relevant. The review should establish what is known about this unit and transaction, and what still needs evidence.
Continue reading
English content and examples reviewed: 27 October 2026. Read the Arabic edition.
