After a Budget square that’s accurate

[johnnyb]

another vote for empire blue. reasonable quality. but definitely square from the factory. my Moore and Wright is a delight though.

 

[Phil Pascoe]

Yes obviously. Hence the need to double check - first that the edge is straight and second by repeating the exercise on other parts of the same edge, or another one.
It's not rocket science.
I've used lengths of MFC 6" shelving as straight edges, with no problems.

It's not rocket science? No, of course not, but the straightness of the edge although presumed hadn't been mentioned at the time.

 

[TheUnicorn]

its a fair point about presuming an edge is straight, I must remember that in the future and check in a couple of locations along an edge

 

[steve355]

I had a faithfull “engineering square” that was so out of square it was shocking. Bin.

I have several vintage combination squares that are perfectly good for woodworking purposes. I have a few vintage engineering squares that aren’t square enough for engineering. All depends on your mission.

I also have a granite surface plate and granite square, which I used when I was refurbishing my surface grinder, all accurate to <0.0001”. It isn’t much use for marking tenons. But I can check other squares with it (plus some feeler strips).

Chronos do good quality unbranded squares at a sensible price, to engineering specifications.

Steve

 

[Harold Hill]

Either buy a square that actually quotes an accuracy standard e.g. BS939 or DIN875, or go for a speed square where the accuracy of mass production works in your favour. Their construction means they hold their squareness well too.

Hi guys

Was wondering about Squares, combination/ carpenters/ engineering/ etc.
What company makes the most accurate (at a DIY,ers budget) I tried bacho, Presh, parkside, all are accurate in small size but the longer the rule the less accurate they become.

Can’t justify the 3 sum figure of a starrett, but is there any company with good accuracy?

Cheers

Mitutoyo very good.

 

[ChaiLatte]

You know a square is no good when you have two back to back and can see daylight down the middle.

If you look at your statement logically, it is not so.

It reminds me of the old Chinese proverb: "man with one watch always knows the time; man with two watches never sure".

Holding two squares back to back and seeing daylight only tells you that one or both is out. Holding two squares back to back and not seeing daylight could mean that both are out but in opposite directions.

The method posted above by MikeK is reliable and deterministic.

 

[Daniel.l]

I bought a Magnusson square when I first started and it is actually square and has stayed square for a good few years now. It could be a one off but you could buy a small engineers square and test them before you buy

 

[Spectric]

Holding two squares back to back and seeing daylight only tells you that one or both is out. Holding two squares back to back and not seeing daylight could mean that both are out but in opposite directions.

You have overlooked the laws of probability.

 

[Jorny]

I have been very pleased with Nobex Octo squares.

 

[HamsterJam]

I have a Marples square that I inherited from my late father.
It is accurate enough for my needs and is my reference for other squares.
They aren’t too expensive and quote 0.01mm/10mm blade length…
https://www.flinn-garlick-saws.co.uk/product/joseph-marples-trial-1-try-squares/

 

[Phil Pascoe]

I've read good reports of these Presch ones -
https://www.amazon.co.uk/gp/product/B08NVQ52VB?ref=em_1p_0_ti&ref_=pe_2443691_780776981
An interesting comment on the J. Marples Trial 1 squares on their accuracy on the INSIDE edge. Many people don't realise that the standard applies to the inside edge not the outside, which we tend to use as much.

 

[ChaiLatte]

You have overlooked the laws of probability.

So do feel free to enlighten me.

If you pick up two squares at random, each has three possibilities: acute, correct or obtuse.

Thus, there are nine combinations (of combination square).

Three of those nine will show no daylight. One of those three has both squares correct but there is no way of determining which one.

Hence, if you use "lack of daylight" as your assessment criteria, you will select a bad square 2/3 of the time.

Compare to the other method suggested above, which selects a good square every time.

 

[ajs]

So do feel free to enlighten me.

If you pick up two squares at random, each has three possibilities: acute, correct or obtuse.

Thus, there are nine combinations (of combination square).

Three of those nine will show no daylight. One of those three has both squares correct but there is no way of determining which one.

Hence, if you use "lack of daylight" as your assessment criteria, you will select a bad square 2/3 of the time.

Compare to the other method suggested above, which selects a good square every time.

It's not only the direction of error but also the amount, i.e. if one square is 89° and the other 92° there's still a 1° apparent error since the two angles are not supplementary.

 

[Jacob]

......

An interesting comment on the J. Marples Trial 1 squares on their accuracy on the INSIDE edge. Many people don't realise that the standard applies to the inside edge not the outside, which we tend to use as much.

I've done many hours of marking, over many years, with similar Marples style squares and I doubt I've used the inside edge even once.

 

[Jacob]

.....

I also have a granite surface plate and granite square, which I used when I was refurbishing my surface grinder, all accurate to <0.0001”.

Yebbut how would you know?

 

[ajs]

Yebbut how would you know?

Marking blue and a reference surface, a random pair of parallels do if another surface plate isn't to hand. I'd suggest reading up on Whitworth's method of hand scraping a set of cast iron surface plates, even if you never use it the underlying principles are often useful to bear in mind.

 

[steve355]

Yebbut how would you know?

Cos it has a certificate wot says so. It was inspected by an inspector.

The surface plate is the reference surface in my workshop and if that isn’t flat, nothing is.

As the previous poster says, you can ensure a surface is flat by comparing it with 2 other candidate flat surfaces, in a process invented by Mr Whitworth, one of our greatest Victorian engineers who also invented the Whitworth rifle, and the Whitworth thread, vastly superior to M-threads.

Slightly off topic I know but here’s an interesting video on Whitworth, the father of precision.

 

[Spectric]

Hence, if you use "lack of daylight" as your assessment criteria, you will select a bad square 2/3 of the time.

Assuming the OEM is such a bad manufacturer that there is not a common error, you are more likely to find if there is an error that it has been repeated as part of the manufacturing process and so when put back to back it is amplified so obvious. It is an easy method you can use just to see the obvious, it is not a method used in an inspection or calibration lab.

 

[throbscottle]

The Record one I got 20 years ago is still accurate despite having been dropped a few times. I test it regularly. Got a set of 3 Parkside ones a couple of years back. Not used much but I think they might be slightly better than the Record (so far)

 

[Kev.k]

I’m about to buy a combination square. The cheap junk I’ve bought in the past was useless.

Thinking of going with Teng SQAC300

​