Thread: Squares - I wish

I've got the Chris Vesper double square and it's a beauty, I can't fault it.
I also have a Crown CC knock off. It's terrible, I've now dismantled it for parts

When he ffffiiiiiiiiiiiiiiinnally releases them I'll be buying a CV 7" square

Originally Posted by Scribbly Gum

I'm guessing that 1mm over 360mm would be 1 degree so yours is close to 1 degree out. If my logic is right it would be 0.9 of a degree.

Cheers
SG

SG,

I think your recollection of trigonometry is a bit out. Unless I and my calculator are mistaken, the arc tan of 1/400 is 0.143 degrees, and as Derek points out, the actual error is .5 mm in 400 mm. So the error there is 0.07 degrees.

Rocker

I was going to post that Rocker.
Well I wanted a square square one day and didn't have one so I got a diamond stone and rubbed the inside base of the handle to widen it a little and after a few trys it was spot on (on the inside). This would only work if it is under 90 if over try another square.

I mostly use my engineers square to true up machines & Jigs & to check my Irwin square. My engineers square it my base reference so i look after it. Not so gentle with some of my other squares.

Great replies. A passionate subject by the looks.

It seems I need to get myself an engineers square for truing up blades, and just stick to what I have have for wood working.

If you buy two squares that are accurate to half a degree theoretically they could be a whole degree different from each other. One could be over, one could be under.
The whole of this thread shows why it's so important to measure diagonals in the final assembly of a piece and not to rely on one or more corners being 'square' according to a try-square.
Jim

Originally Posted by Rocker

SG,

I think your recollection of trigonometry is a bit out. Unless I and my calculator are mistaken, the arc tan of 1/400 is 0.143 degrees, and as Derek points out, the actual error is .5 mm in 400 mm. So the error there is 0.07 degrees.

Rocker

Correct - it looks like Scribbly has got his mm and º mixed up.

For todays text - I take, "Uncertainty analysis 101"

Another quick way to check squares is using a digital angle finder. Lightly clamp the handle on a surface, zero the angle finder and then measure the angle on the blade (see extra little trick below). A total uncertainty estimate is ~ +/- 0.2º (0.1º for each measurement). For a 300 mm square that is about +/- 1 mm.

It may appear possible to do better than this with the flipping the square on a straight edge method and a very fine pencil. But most people forget there are uncertainties in the base line comparison as well (ie getting the base line up exactly). The flipping method requires two comparisons, (1) a base line comparison - drawn line with base of flipped blade, and (2) drawn line with top end of flipped blade. I doubt whether each of these could be done in practice to within 0.5 mm each, so that's a total of +/- 1mm uncertainty. We can say that is a double angle measurement so the final limit of uncertainty is +/- 0.5 mm (yes we can legitimately halve the uncertainty). (now for the little trick) one can flip the square on the flat surface and use the digital angle finder too and claim the same thing!

Being a bit more technical we should add the uncertainties in what is called quadrature (square root of the sum of the square of the the uncertainties) so the final practical limit of uncertainties for both methods is ~ 0.35 mm which is about the same as practical human vision in most sheds.

On squares longer than 300mm the pencil method will have a lower uncertainty, on shorter squares the digital angle finder will be lower.

Here endth the lesson!

um... no fries today thanks.

How such a small ask as in Wondai's original question can spur so much passionate debate and not a little outright waffle. A few rave about the CC adjustable and CV tools but no one explained how they checked/adjusted same.
When I did my Basic Metalwork in the RAAF back in the early Sixties, we had to draw all the tools each day from the tool store. The squares we used were Starret Engineers 6 Inch, and before you were allowed to use same, they had to be checked (each day) against a Starret Master, which was a round steel machined bar about 300mm long by 100mm Dia. if the square was out, and most of the time they were, you just tapped the blade on a wooden bench until they read true.
So those of you who swear by an engineers square being square, think again.
To those who say wood moves etc. or measure diagonals, thats ok in the finished item, you know and expect it to move, its just a matter of how much, but when in the building/assembly stage its nice to know you can achieve some semblance of square.
I, like most here, use the mark and flip method to calibrate my square, but in my case I use a single bevel marking knife with the flat against the blade of the square, strike a mark at the base and end, flip knife and square, place knife in base cut line and move the square up to it and compare, you can be certain of better than 0.2mm any time, and the longer the square the easier to calibrate.
Cheers John M

Most engineering squares are very accurate and they don't have to be the expensive ones. I got 3 of them and they costed from $10 to $30.


https://www.woodworkforums.com/showpo...0&postcount=57

Same method as Woodwould and Ravna, I have a collection of old rosewood handled squares that are all hand me downs. I reference them against a brass set square and just tap the tips on a bench if they need adjustment. They seem to hold square and for timber work are more than accurate enough.

Hi all

The question of getting a square dead square is interesting, that's why Incra opted for a one piece laser cut version that they guarantee to be with in 1/1000th inch from any point to point measure e.g. tip of blade to tip of stock.

The flip over method with a scribing knife is IMHO the only true way of checking a square.

Like many things in woodwork there are many views and probably many correct solutions to this issue.

Regards

Grahame