New square required!

[Trafalgar]

Lines will meet even if the square or work piece isn't perfect:

Register square on face edge, mark the true face. Then register square on true face and drop line down the face edge. Register square on face edge and mark the back face. Then register square on true face and mark edge to meet the back face.

 

[Peter Sefton]

Lines will meet even if the square or work piece isn't perfect:

Register square on face edge, mark the true face. Then register square on true face and drop line down the face edge. Register square on face edge and mark the back face. Then register square on true face and mark edge to meet the back face.

Indeed, thus using an inaccurate square leaves you believing you are marking out correctly but you will have accuracy issues later in the making process.

Use an accurate square for piece of mind.

Cheers Peter

 

[Trafalgar]

A sanity check is to see if smaller workpieces will stand upright on the bench by themselves, and larger carcase pieces upright on their mate with no gaps showing, near where the two will be joined - the 'where the rubber meets the road' test. Rail ends should always be offered up to their mating stiles to see if there are gaps before tenons are cut. If the stile has a little belly or the rail end isn't a gap-free fit then twist can be introduced at glue up. This is part of the process of getting over subordinating the accuracy of a work piece to the accuracy or supposed accuracy of a machine tool, or even a square. Test parts against their mate to the extent possible. Don't trust anything but test fits.

 

[John Brown]

"The Starrett 20-24 is a particularly attractive bargain in a 24" square for the low low price of $1800 or so."

In my naivety, I presumed this was a typo. Maybe $180, I thought, or possibly "low" was a mis-typed "ow!".

So I googled it.

Suffice it to say that I will never need a square of that accuracy, unless I'm building an apparatus for detecting gravitational waves. And I doubt that wood would be the best choice of materials for such a device.

 

[Trafalgar]

Large rectangular work pieces can of course be tested for square by measuring the diagonals. Edges square to the face can be tested with even a somewhat out-of-square, square, since you're only measuring squareness over a distance of an inch or less (the thickness of the component) and most squares are dead accurate over such a short distance. Slight dips and bumps at panel ends, which won't affect the diagonal test one way or the other, are best evaluated against the top of the workbench or the mating component. You don't need a large expensive square to make furniture. Some 'errors' between components do cancel. Be happy when that happens.

Some 'errors' that appear relatively huge when measured with a tool room grade square completely disappear or are clearly inconsequential when checked against its mate in the article of furniture being made.

 

[Peter Sefton]

I am afraid measuring diagonals does not guarantee squareness, trapezoids have the same diagonal lengths but are not square.

Cheers Peter

 

[Trafalgar]

A perfect one does and it's rare that you could manage to produce such a thing by accident. It doesn't matter because you're still going to test the panel against it's mate and there surely will be a gap unless you've managed to produce a mirror image in the mating piece - essentially an impossibility. See my previous posts.

The notion of working to a square, as if working a single project component in total isolation, is antithetical to more things than I have time to list. It is certainly not hand craftsmanship.

I'm not mounting an argument against owning a decent square. You have to have something, and it might as well at least start out accurate. If it stays that way so much the better. One would still be foolish not to test work pieces as I have described in my other posts. This is essential for work pieces larger than the capacity of one's square.

 

[thick_mike]

I am afraid measuring diagonals does not guarantee squareness, trapezoids have the same diagonal lengths but are not square.

Cheers Peter

Can you not measure 3 units one way and 4 the other? If the diagonal between the two points is 5 units, the angle must be 90 degrees.

 

[Trafalgar]

Yes, you can.

 

[Jacob]

I am afraid measuring diagonals does not guarantee squareness, trapezoids have the same diagonal lengths but are not square.

Cheers Peter

Obviously not. There's no way to guarantee the squareness of a circle either.
Making rails and stiles in pairs of the same length, with joints in the right places,and measuring the diagonals, does guarantee squareness, assuming they are all in the same flat plane

 

[Peter Sefton]

I am afraid measuring diagonals does not guarantee squareness, trapezoids have the same diagonal lengths but are not square.

Cheers Peter

Can you not measure 3 units one way and 4 the other? If the diagonal between the two points is 5 units, the angle must be 90 degrees.

Yes you are correct this does work but I have used this method more in laying out footings and brickwork than in furniture.

Cheers Peter