How We Pick
This page explains how WeldGearLab arrives at a recommendation, and just as importantly, what our process does not include. Read it before you trust anything else on this site.
What our process is
Our picks are research based. For every product we consider, we work through the published specifications, the product listing, the safety listing, the warranty terms, and what comes in the box. We compare those against the other products in the same category and price range, and against the work our readers actually do.
Where a number can be worked out rather than looked up, we work it out and show the basis. Duty cycle is the clearest example. A rating of 20 percent at 90 amps means about 2 minutes of welding in every 10 at that output before the machine needs to cool. We restate ratings that way so two machines rated at different amperages can be compared fairly.
We also check what a machine asks of your shop. A welder that needs a 240V circuit is a different purchase from one that runs on a standard 120V outlet, and a gas setup needs a cylinder and regulator that may not be in the box. We say so in every pick it applies to.
What our process is not
We do not run a test shop and we do not take physical measurements of products. We have not welded with the machines on this site. When you read that a welder is rated for 1/4 inch steel in a single pass or that a plasma cutter severs 5/8 inch, that figure comes from the manufacturer, and we label it that way every time.
The Lab in WeldGearLab describes how we organize and compare published data. It is not a claim that we operate a laboratory.
We say this openly because a few numbers in this category are stretched often enough that a buyer should read them with care: maximum thickness, which may describe a multi-pass weld or a sever cut rather than a clean cut; duty cycle, which brands rate at different amperages and different ambient temperatures; and “up to” amperage, which may only be reachable on 240V input.
The Gear Score
Every product we recommend carries a Gear Score from 0 to 100. The same five criteria apply to every product in a category, and the weights are fixed and published here rather than adjusted to suit a conclusion.
| Criterion | Weight | What it covers |
|---|---|---|
| Performance | 30 | The specifications that decide whether the product does the job: amperage range, duty cycle at a stated amperage, input voltage options, supported processes and rated thickness for welders; rated cut and sever thickness for plasma cutters; viewing area, shade range, sensors and optical class for helmets. |
| Build | 20 | Materials, construction, what is included in the box, serviceability, and whether consumables and replacement parts are easy to find. |
| Value | 20 | What you get for the money within the price range of the guide, including the extras you will need to buy before you can start. |
| Safety listing | 15 | For machines and tools on mains power, a stated CSA, UL or ETL listing. For helmets and eye protection, a stated ANSI Z87.1 or CSA Z94.3 rating. |
| Warranty | 15 | Length and scope of the warranty, what it excludes, who pays return freight, and how reachable the company is. |
We do not skip a criterion because a manufacturer stayed quiet. Where a brand does not publish a warranty term, or does not state a safety listing or rating, that criterion scores zero. Non-disclosure is not neutral. A buyer cannot compare what is not published, and a score that ignored it would reward the least forthcoming brands.
Warranty points scale with what is put in writing. A one year term scores 7 of 15, two years 9, three years or longer 10, and a term that also states what it covers can earn up to 11. A promise of customer support without a repair or replacement commitment earns 2. No product gets full warranty marks unless the listing states the term, the coverage and who pays return freight. Where a listing states a safety listing for one component only, such as a power supply, we give partial credit and name the component.
Where a performance figure itself is missing, we say the information is unavailable rather than guessing at it.
How we choose which products to include
We start from what is actually available to buy in the United States in the category and price range of the guide, then narrow it down. Products are left out when the listing publishes no meaningful specifications or when the seller cannot be identified. Each guide uses its own set of products, so a machine you see recommended in one guide is not simply repeated in the next.
We are not trying to list every product. A roundup of seven picks where we can explain each choice is more useful than a list of thirty.
Language we avoid
You will not read the words tested, lab tested, or hands-on on this site describing our own work, because that is not what we do. We review, evaluate, compare, rate and recommend. The distinction matters both to you and to the Federal Trade Commission.
Updates and corrections
Products change. Manufacturers revise specifications, discontinue models and change warranty terms. Every article shows its publication date and the date it was last updated, and we revisit our guides when a product in them changes generation.
If you find something wrong, write to [email protected]. We correct errors and note what changed. Our full approach to accuracy and sourcing is in the Editorial Policy.
A note on safety
Welding and cutting involve arc radiation, fumes, hot metal, fire and electrical hazards. WeldGearLab reviews equipment and is not a substitute for training. Please read our Safety Disclaimer before acting on anything you read here.