How to Measure Steel Box Section Thickness Before Welding
TL;DR: You cannot get a caliper inside a closed tube, and guessing steel box section thickness is how burn-through happens. An ultrasonic gauge solves the geometry: one flat probe on the outside face, a blob of couplant gel, and the wall thickness reads out in millimetres from one side. With the InfuriPaint Pro-T1 (£104.87, metal mode 1.0–300.0 mm, ±0.05 mm) a five-minute check of every length in the rack tells you what you actually have — not what the label, or the bloke at the scrapyard, said you have.
Why nominal wall thickness lies to you
Three ways the "3 mm box" on your bench turns out not to be:
- Reclaimed and second-hand stock. Salvaged bed frames, old racking, offcuts from a job lot — the wall could be anything, and corrosion inside the tube (where you cannot see it) may have thinned it further.
- Mixed batches. A rack that holds both 2 mm and 3 mm lengths of the same outside dimension is indistinguishable by eye. The two need noticeably different welding settings.
- Tolerance and wear. Even new ERW box carries manufacturing tolerance; used structural steel may have lost material to rust scale exactly where you plan to weld.
The failure mode is familiar: set the machine for the wall you assumed, strike an arc on the wall you actually have, and either blow through thin stock or stack a cold, under-penetrated bead on thick stock. Both mean grinding it out and starting again — the weekend gone on rework.
The one-side measurement, step by step
- Zero the gauge. Check the Pro-T1 against its included calibration block first. Seconds now, confidence in every number after.
- Prep a test spot. On painted, powder-coated or rusty stock, take a flap disc or sanding block to a coin-sized patch down to bright metal. Loose scale, flaking paint and pitting scatter the echo — a gauge reading through crud reads wrong or not at all. Clean mill-finish steel needs only a wipe.
- Apply couplant. A pea-sized blob of the included gel on the spot. Ultrasound will not cross an air gap; the gel is not optional. (Running low, any standard ultrasonic couplant works — even a thick water-based gel in a pinch on flat spots.)
- Probe flat, read, repeat. Hold the transducer square to the surface, let the reading settle, then check two more spots along the length. Consistent numbers = trustworthy numbers. On round tube, probe on the crown and keep it square — curvature punishes sloppy contact.
- Mark the stock. Paint-pen the wall thickness on each length as you verify it. Your future self, mid-project, will thank you.
From millimetres to machine settings
With a real wall figure you can set the welder from a chart instead of a guess. The old MIG rule of thumb — roughly one amp per thousandth of an inch, so about 40 A per millimetre of steel — gets you to a starting point (a rule of thumb, not gospel: joint type, wire size, gas and machine all shift it, and thin-wall box likes a downhill technique and shorter arc time regardless).
| Measured wall | Practical read for hobby MIG |
|---|---|
| 1.5 mm | Delicate territory — low power, tack-and-move, expect distortion management |
| 2 mm | Comfortable hobby-machine range with care on gaps |
| 3 mm | Sweet spot for garage box-section projects; single pass fine |
| 4–5 mm+ | Bevel or multi-pass thinking; check your machine's duty cycle honestly |
For load-bearing work — trailer frames, loft ladders, anything with people under or on it — the measurement is also your material record: if a wall reads 40% under nominal, that length gets demoted to bench-frame duty, not tow-bar duty.
Rusty reclaimed steel: how far to clean before trusting a number
Surface rust that wipes to a powder measures fine after a quick sand. Flaking laminated rust means grinding to bright metal and expecting the number to drop — you are measuring what is left, which is the point. If a reclaimed length reads wildly different spot to spot (say 2.9 mm, then 2.1 mm a hand-width away), corrosion has been eating from the inside; that variability, not any single reading, is the reason to relegate the length to non-structural jobs. This one-sided honesty is exactly what a Birmingham owner's verified review of the Pro-T1 describes — using it for checking steel box section thickness before welding, with an ultrasonic mode that is "spot on".
FAQ
Will it measure through paint or powder coat?
Coatings distort single-echo ultrasonic readings (they add apparent thickness or kill the echo). Sand a test spot to bare metal — thirty seconds with a flap disc — and measure there. The same honesty applies to thick rust scale.
Does it work on aluminium box and stainless?
Yes — ultrasonic metal mode measures common metals including aluminium and stainless from one side with couplant, in millimetres. (Different alloys have slightly different sound velocities; check readings against a known sample when precision matters.)
What is the thinnest wall it can read?
Metal mode spans 1.0–300.0 mm with 0.01 mm resolution, which covers everything from thin-wall furniture tube to serious structural section. Below ~1 mm (car body skins), measure a folded edge with calipers instead.
The included gel bottle is small — what do heavy users do?
Exactly what the reviews suggest: buy a larger bottle of standard ultrasonic couplant when the sample runs out. The gauge is not fussy about brand. USB-C charging means the battery, at least, never sends you hunting — as one owner put it, no more nicking AA batteries from the TV remote.
Next step: the same metal mode monitors steel hull plate — see narrowboat hull thickness testing — or if you are costing up instruments for a workshop, our budget ultrasonic thickness gauge guide covers QC use. Full metal-mode specs: Pro-T1, £104.87 with free UK delivery.