A small mistake.

I made a small (75mm x 13mm) brass flywheel both faces to be identical but later noticed I had made a mistake. The rim on one side was 8mm wide but 7.95mm on the other. I was quite surprised that this small discrepancy was noticeable to me by eye considering you can't see both at once as they are on opposite faces of the wheel.

So how small difference can the human eye/brain discriminate?

Edited By Roger Provins 2 on 13/01/2016 16:51:39

It depends a lot on the circumstances, and it would be impossible to give a simple answer. We are generally very good at noticing pictures that are out of vertical, for example, but we are easily confused by 'optical illusions'. I like to think that I can distinguish a 1/4 Whit bolt head from a 7/16 UNC or M8, all of which differ by a few thou. But I expect I am also influenced by other clues at the same time.

The human finger-nail can detect a step of 0.0005 inches - is that a good start at an answer, I wonder?

Cheers, Tim

Hello Roger,

When I made my living industrially, part of my work was in viewing radiographic images of metallic structures for hidden faults and material defects.

It was always taken as a general case that two small inclusions, showing up as black on the X ray film, could be discriminated by the unaided eye satisfactorily when they were 0.010 inches [1/4 mm] apart. Magnifying aids helped a little but the definition and contrast then began to 'bleed' and deciding what constituted an edge became more difficult..

For the sort of difference you are talking about here with sharper edges, the eye is very good at spotting such differences and I would imagine as little as 0.005 inches might be detectable.

Regards

Brian

We use 300 dpi for print media as the eye's resolving power is about 0.003" at normal reading distance. So to see two separate, tiny, dots they need to be at least 0.003" diameter and at least 0.003" apart - which is as near as spit your 0.010".

Neil

Roger ... an interesting question ... I assume you have dozens, or even maybe more than a hundred small twist drills in your "shed" ... take the range from 1.5mm up to about 2.5mm ... try putting the shanks of 2 nearly similar bits in line with each other and touching end to end ... with a good light and a light background you may surprise yourself ... you may be able to differentiate "one thou" difference in diameter.

martin

Martin, indeed but in this case it isn't a case of physically comparing items while both are visible but rather holding one image in memory and comparing it with similar visible item. I'm sure it's something we all do subconsciously all the time. It just surprised me how small a difference can be noticed.

Roger

Edited By Roger Provins 2 on 13/01/2016 20:34:15

I can prove this.

It's down to the power of the MKI eyeball.

Take centering a shaft on the mill for cutting a keyway as an example. Many people have many different ways and gadgets to do it but the MKi eyeball still comes off the best.

Take a good look at this picture.

This shaft was lined up exactly central by touching on both sides with a dowel pin and splitting the difference on a DRO so backlash wasn't an issue.

A milling cutter was just allowed to kiss the surface on the left, you can see the 'butterfly' is equal.

It was then moved along at the same zero setting, then moved sideways by one thou, 0.001" and a second mark made, the one on the right.

It's a mile out, so no need for fancy gizmo's, just touch down on the work, look at the spot and trust the MKI eyeball, unless you need to go to specsavers.

I was after a bit of 1/4 stainless rod once, and a bloke at a fabricators gave me a length, but I told him I thought it was 6 mm. I measured it when I got home and it was 6 mm. I had no reference. I am continually amazed in the workshop what the eye can discriminate.

I can seem to tell Metric threads from Infernal as the pitch is not right for the diameter. (Am a Yank.)

Roger,

As per Tim Stevens' first response: "It depends a lot on the circumstances"

It would be interesting if you could kindly post a matched-pair of photographs, so that we can see for ourselves.

MichaelG.

Michael,

Sorry it's too late I've corrected it now.

Roger

Females are better at detecting small errors, at least my wife is.

They remember them better too.

Good at detecting unposted parcels by the front door

I can see a 0.04mm or greater run-out in the lathe by eyesight.