What's inside
As an Amazon Associate I earn from qualifying purchases. This post may contain affiliate links at no extra cost to you.
A combination square can look square and still be out by enough to spoil a cabinet carcass, a machine setup, or a line of joinery. Checking it against a known straightedge is a useful shop test, but there is an important limitation: a straightedge checks whether the square’s blade is straight; by itself, it does not prove the blade is exactly 90 degrees to the stock. Use the checks below to separate those two problems.
What you need
Use a flat, stable surface, a sharp pencil or fine marker, and a reliable straightedge. A quality steel rule or known-straight metal edge is convenient; a verified machinist’s straightedge is better for precision work. A framing square or another combination square is not automatically a trustworthy reference just because it looks straight.
Clean sawdust, dried glue, and burrs from the combination square’s blade and stock. Check that the blade seats fully in the stock’s groove and that the locking screw holds it without allowing the blade to shift. If the tool has a reversible head, check that the blade is seated in the same orientation you normally use it.
If you need a reference edge, a precision straightedge is a sensible purchase for layout or machine setup. For general carpentry, a decent steel rule or a known-straight metal edge may be enough.
Check the blade against a straightedge
Lay the straightedge and combination-square blade side by side on the flat surface. Bring one long edge of each into contact, with their ends aligned. Hold them gently together and look for gaps along the length. Do not force the pieces together: pressure can hide a small bow or make a thin blade flex.
Repeat the check with the combination-square blade flipped over, so the opposite blade face contacts the straightedge. A gap that stays in the same place relative to the blade suggests the blade may be bowed, kinked, or burred. A gap that changes when you flip it can point to a damaged edge, dirt, or an unreliable reference. Sight along the contact line against a bright background, or shine a light from behind; a bright, continuous sliver is easier to spot than a dark gap.
For a quick shop check, a visible gap is the warning. For finer work, measure it with feeler gauges or compare against a straightedge with a stated accuracy. A ruler’s marked graduations do not establish that its edge is straight. If the blade is bent, do not try to correct it by levering against the straightedge. Replace the blade or the square; bending it can introduce a kink that is harder to detect.
Check whether the square is actually square
A straight blade can still sit at the wrong angle to the stock. To test the 90-degree relationship, place the square’s stock against a clean, straight reference edge on a board. Draw a fine line along the blade. Flip the square over, keeping the stock against the same reference edge, and draw a second line from the same starting point. If the two lines diverge, the square is out of square. The farther apart they are at the far end, the larger the error.
For a more controlled test, make the lines about 200 mm (8 in.) long. If the lines are 1 mm apart at that distance, the error is roughly 0.29 degrees. A small difference may be harmless for rough framing, but it is worth correcting or replacing the tool before using it to mark cabinet parts or machine setups. Keep the pencil sharp and use light pressure; a thick line can obscure a small error.
This test depends on the board edge being straight and the square’s stock sitting flat against it. If the result changes when you move the square to another section, the reference edge may be the problem. Check the board with the straightedge first, or use a verified square as a cross-check.
Read the result before replacing anything
| What you find | Likely cause | What to do |
|---|---|---|
| Gap between blade and straightedge | Bent blade, burr, dirt, or poor reference edge | Clean and repeat with the blade flipped; replace a bent blade |
| Blade is straight, but test lines diverge | Stock-to-blade angle is off, or the stock is not seated consistently | Repeat against a verified edge; replace or adjust only if the tool allows it |
| Results change along the board edge | Reference edge is bowed, rough, or damaged | Check with a straighter reference before judging the square |
| Blade shifts when the lock is tightened | Worn or poorly fitting locking mechanism | Retest after locking; replace if it will not hold a repeatable setting |
Adjust, repair, or replace?
Some combination squares have a fixed head and offer no dependable user adjustment. If one fails the flip test, replacing it is usually more practical than trying to bend the blade or file the stock. A bent blade can be difficult to straighten accurately, and removing material from the stock can make the tool worse or create a sharp edge.
On a square designed for adjustment, follow the maker’s procedure and make small changes. Retest after each one. Do not assume that tightening the lock harder will fix a bad fit; excessive force can shift the blade, damage the screw, or make the reading less repeatable. If you need a replacement, a combination square for carpentry with a blade that locks firmly is often adequate for site work. A cheaper square is fine for rough layout if it repeats reliably; it is a poor choice for fine marking if the blade rocks or the lock moves it.
Keep the check repeatable
Store the square where its blade will not be bent by other tools, and wipe metal surfaces dry after use. Before marking a run of parts, check that the blade has not shifted in the stock and that the reference edge is clean. Recheck after a drop, a hard knock, or any time measurements stop agreeing with another trusted tool.
For ordinary framing, a small error may not affect the finished work. For drawer parts, hand-plane setup, or machine alignment, even a modest angular error can compound across a longer cut or assembly. The practical standard is not that every square must meet a machinist’s tolerance; it is that the tool must suit the job, hold its setting, and produce the same result when you check it twice.