oldschoolcraft
Well-known member
I independently came up with something similar. My context is that I could be working on anything, but the probability if I'm working on cars, is that they are metric. Also the probability is the tolerances on hex fasteners are going to be tighter and more important in cars, than anything else I might be doing (like household appliances) or Ikea furniture.
So what I have decided to do is go with all metric sizes, and add in the non-overlapping SAE.
That generally means I will get 10mm to 19mm sets of things, and add in 3/8", 1/2", 9/16", 11/16"
I think a good argument could be made that if you were streamlining down to a single Metric or SAE equivalent size, you should pick the bigger size. 7/16" is 11.1mm so in theory, a 7/16" socket would fit with a bit of slop on an 11mm fastener. Compared to an 11mm socket being too tight on a 7/16" fastener. But, I'm more likely to be using metric fasteners that require a higher degree of tolerance, so I go with 11mm and skip 7/16"
In my version of this, I only go to one decimal point after the metric size. AdAstra goes to two decimal places, which to me seems excessive. He has 5/16" = 1.98mm but on mine I round that to 2.0 because I can't imagine 0.02mm being measurably different in reality. The first time the fastener is installed, it's going to deform by 0.02mm. A human hair is 0.1mm, so we're saying that 5/16" is 1/5th of a human hair larger than 2mm?
From a practical standpoint, the tool manufacturers have to build tooling to create the tools. I can't possibly imagine they are using different tooling for 5/16" and 2mm tools. They're almost certainly using the identical tooling and stamping/engraving a different number on it.
I'd be really curious if anyone has worked in tool manufacturing, to tell what the tolerances of various tools are that they produce. For example, SnapOn wrenches, the 10mm wrench probably has a 10.10mm opening. They must do QA. If they found a wrench was 9.9mm, they'd reject and fail it. But how about 10.05? Is that allowed?
Another thing to consider is stacking tolerance issues. From the shooting guns world, if you put a red dot scope on a gun, and the reticle has a tolerance of 2 Minutes of Angle (MOA) - meaning 2" at 100 yards, and the ammo you're using is 2 MOA, and your gun itself is 2 MOA, then each round you fire is 6 MOA accuracy (6 inches at 100yards) at best because you compound the factors. It's possible at any point in time, the red dot is off by 2 inches high at 100 yards, and if that particular round of ammo happened to be a little hot, it might be 2 inches high, but it's 2 inches high from the aiming reticle point, which was also 2 inches high, and the gun itself might be cheap and loose and it might be shooting 2 inches high. Now you're 6 inches high.
So you have the tolerances of the wrench itself, which might be +/- X, because it's impossible to make the wrenches all identical down to the individual atomic level. You're going to have to +/- 0.0X MM, you can't get down to 0.000X MM tolerances and even if you could, it would be a billion times the price with no practical benefit.
But then the fastener itself has some tolerances from manufacture that make it +/- Y.
And due to the nature of reality, some slight deformation that occurred the first time the fastener was put on, or if it got banged or dinged or has any corrosion. Add another +/- Z
And probably a smaller confounding variable but maybe instead of holding the wrench exactly perfectly on the fastener maybe you're holding it 0.3 degrees which would be imperceptible to your hand anyway. That angular shift will add or subtract a few hundredths of a millimeter.
So you add up all of these possible factors and I will argue that 0.02mm doesnt make a difference and might as well be exactly the same. The question becomes, at what point is it not exactly the same? 0.05mm? 0.1mm? 0.2mm?
Personally I landed at 0.1mm when I did this exercise, but I like how AdAstra uses angular differences, which is a smarter way of saying that 0.1mm makes far less difference to a 30mm fastener than it does to a 0.7mm fastener.
Great work, I love this!
So what I have decided to do is go with all metric sizes, and add in the non-overlapping SAE.
That generally means I will get 10mm to 19mm sets of things, and add in 3/8", 1/2", 9/16", 11/16"
I think a good argument could be made that if you were streamlining down to a single Metric or SAE equivalent size, you should pick the bigger size. 7/16" is 11.1mm so in theory, a 7/16" socket would fit with a bit of slop on an 11mm fastener. Compared to an 11mm socket being too tight on a 7/16" fastener. But, I'm more likely to be using metric fasteners that require a higher degree of tolerance, so I go with 11mm and skip 7/16"
In my version of this, I only go to one decimal point after the metric size. AdAstra goes to two decimal places, which to me seems excessive. He has 5/16" = 1.98mm but on mine I round that to 2.0 because I can't imagine 0.02mm being measurably different in reality. The first time the fastener is installed, it's going to deform by 0.02mm. A human hair is 0.1mm, so we're saying that 5/16" is 1/5th of a human hair larger than 2mm?
From a practical standpoint, the tool manufacturers have to build tooling to create the tools. I can't possibly imagine they are using different tooling for 5/16" and 2mm tools. They're almost certainly using the identical tooling and stamping/engraving a different number on it.
I'd be really curious if anyone has worked in tool manufacturing, to tell what the tolerances of various tools are that they produce. For example, SnapOn wrenches, the 10mm wrench probably has a 10.10mm opening. They must do QA. If they found a wrench was 9.9mm, they'd reject and fail it. But how about 10.05? Is that allowed?
Another thing to consider is stacking tolerance issues. From the shooting guns world, if you put a red dot scope on a gun, and the reticle has a tolerance of 2 Minutes of Angle (MOA) - meaning 2" at 100 yards, and the ammo you're using is 2 MOA, and your gun itself is 2 MOA, then each round you fire is 6 MOA accuracy (6 inches at 100yards) at best because you compound the factors. It's possible at any point in time, the red dot is off by 2 inches high at 100 yards, and if that particular round of ammo happened to be a little hot, it might be 2 inches high, but it's 2 inches high from the aiming reticle point, which was also 2 inches high, and the gun itself might be cheap and loose and it might be shooting 2 inches high. Now you're 6 inches high.
So you have the tolerances of the wrench itself, which might be +/- X, because it's impossible to make the wrenches all identical down to the individual atomic level. You're going to have to +/- 0.0X MM, you can't get down to 0.000X MM tolerances and even if you could, it would be a billion times the price with no practical benefit.
But then the fastener itself has some tolerances from manufacture that make it +/- Y.
And due to the nature of reality, some slight deformation that occurred the first time the fastener was put on, or if it got banged or dinged or has any corrosion. Add another +/- Z
And probably a smaller confounding variable but maybe instead of holding the wrench exactly perfectly on the fastener maybe you're holding it 0.3 degrees which would be imperceptible to your hand anyway. That angular shift will add or subtract a few hundredths of a millimeter.
So you add up all of these possible factors and I will argue that 0.02mm doesnt make a difference and might as well be exactly the same. The question becomes, at what point is it not exactly the same? 0.05mm? 0.1mm? 0.2mm?
Personally I landed at 0.1mm when I did this exercise, but I like how AdAstra uses angular differences, which is a smarter way of saying that 0.1mm makes far less difference to a 30mm fastener than it does to a 0.7mm fastener.
Great work, I love this!