How we judge and rank

A verdict is only useful if you know what produced it. Here is ours.

What we have not done, first

This site does not run a test workshop. No machine here has been through a bench of ours, and you will not find a measured decibel figure or a dial-indicator reading presented as our own. That is a real limitation and we would rather state it in the second paragraph than bury it — a great many sites in this field imply hands-on testing they have not done, and it is the single most common dishonesty in the genre.

What we do instead is narrower and checkable: we read the specifications, the manuals and the standards, we weigh what long-term owners consistently report, and we name the source of every figure.

What we weigh

We rank on the specifications that decide the job and the ones that decide how long a machine keeps doing it. In practice that means five things:

• Duty cycle, stated honestly. The percentage, the amperage it applies to, and the ambient temperature it was measured at. A machine quoted at 60 % without those two figures has told you nothing, and we say so rather than repeating the number.
• Control where it is hardest. Thin sheet on MIG, arc stability under 20 A on TIG, restriking a basic electrode, dross on the underside of a plasma cut. Every category has one job that exposes a machine, and that is the one we describe.
• Fume and safety behaviour. Where the fume goes, whether the machine can be used with an extraction nozzle close in, guard and restart protection on grinders. This is weighted more heavily here than on most sites, deliberately.
• Consumables and serviceability. What tips, nozzles, discs and liners cost over a year; whether parts are sold separately; whether a torch is a standard fitment or a proprietary one. This is why a machine that welds slightly worse sometimes outranks one that welds better.
• Availability where you are. We verify each product reference market by market. A machine that is excellent but not distributed in your country is not a recommendation we can make to you.

Where the evidence comes from

Three sources, weighted differently and always named. Manufacturer specifications and manuals, cited as such rather than repeated as fact — they are reliable for dimensions and unreliable for performance. Published standards and independent measurement where they exist. And the weight of long-term owner reports, which is the only practical source for the question that matters most: what is this machine like in year ten, after the warranty and after the honeymoon?

Where a claim rests on one source rather than three, the page says so. Where we could not source a figure at all, the field is left empty and the gap is stated. We never fill one in by inference.

What the ordering means

We publish no scores. There is no number out of ten on this site, and there never has been — a single figure would have to compare a 200 A inverter with a packet of contact tips, and it cannot. What we publish instead is an order: within a category, the sequence is deliberate, and the reasons for it are in the text.

Hubs are sorted by price tier — entry, mid, then top — because a hub should read like a budget ladder, not a list with the most expensive machine nailed to the top.

The biases we designed against

Affiliate sites drift towards expensive products, towards whatever is newest, and towards whatever converts. Our defence is structural rather than a promise: no ranking on this site has ever been reordered because of what a product pays, several pages recommend building a jig instead of buying one, and no product has ever been placed in exchange for anything.

The one bias we cannot remove is availability. We cover what is genuinely purchasable across Amazon's US, UK, German, French, Italian and Spanish stores, which means excellent machinery sold only through specialist dealers — a good deal of the serious European market — is under-represented here. Where that matters to the recommendation, the page says so.

Safety is not a section, it is a constraint

These machines burn, blind and start fires, and welding fume has been classified as a Group 1 carcinogen by the IARC since 2017 — mild steel included, not only stainless. No page here ranks a machine higher for making a guard easier to remove, and no procedure on this site replaces the manufacturer's instructions for your specific machine. Where a technique is commonly done unsafely, we describe the safe version or we do not describe it at all.