Thread: Recommended reading please : flatness - accuracy - straightness

Hi Bill,
Basically the way I see it and with the odd generalization thrown in.
If you want two surfaces to slide over each other (without moving in another plan) they have to be flat, the flatter the better*. Now there are other ways to get there but the easiest is to use a plate that you know(or at least for most of us "assume") is flat. As you cant make a surface flatter than the one you started with, you reach a point where you are just chasing your tail, so the flatter the better.

For the most part all this talk about flatness isn't really about metal working as such(though it can be). Its more about the machines themselves.

You can start out as agricultural and inaccurate as you like. If you're making hinges for a gate +/- 1mm is unlikely to matter. If you're making a piston for a diesel engine you're going to want to be a lot closer than that.

Flat surface is also handy for measuring.

Hope at least some of that makes sense.

Stuart

*Up to the point to YOU choose not to worry about it any more, be that 1mm or 0.001mm(or in Ray and Josh's case some faction of a 0.001mm if they so choose). Even then you are talking about flat generally not locally. if things are to flat locally they wont slide well either.......... fun isnt it?

Originally Posted by steamingbill

Can I start off by being fairly agricultural and innaccurate ?

Why not, I did.

Standard reading on the subject is Connelly's Machine Tool Reconditioning. And Moore's Foundations of Mechanical Accuracy.

But you would have to have a real interest, as they are the most boring books know to man. That's sort of my industry, but I cant stomach them.

Regards Phil.

Hi Bill,

Just to add a little to what Stuart and Phil have already said, as far as references go there are a number of good ones, but I'd start with "Foundations of Mechanical Accuracy" by Wayne R Moore

There's a PDF floating about somewhere? RC might know...

As to why bother about surface plate flatness? The surface plate is used as the primary reference for scraping machine parts, so any errors in the surface plate will transfer to other parts. You blue the part to be scraped against the surface plate. You really want the surface plate to be verified to be flat..

Machine rebuilding and reconditioning is a big part of the various forum activities, finding old quality machine tools and resurrecting them to build up a cheap and efficient home workshop is almost a full time job. Scraping worn ways and machine parts is often needed.

I don't think I've ever made anything on the lathe or mill to better than 0.01mm but to restore machines that can do that accurately needs you have to measure to at least the next digit of accuracy.

So to measure something accurately to 0.01mm, ideally you should be accurate to +-0.005 or better..


Regards
Ray

PS
At one of the scraping classes someone asked Marco, what **really** is 5 microns, Marko with his German accent said, "dip you finger in sh*t and then wash it
and you can still smell it.. that's 5 microns"

No need to jump in the deep end, just start where you are comfortable. We are all here because we like metal work and what we can build and there are guys just starting out like yourself to full professionals.

When they are talking about microns and scraping that is mainly to do with machine reconditioning and scraping them back to factory or better specs.
Our home shop lathes or milling machines are never going to cut repeatably to these low numbers, so for the average guy working in the back shed there is no need to worry about that sort of precision.

Dave

Originally Posted by steamingbill

Can I start off by being fairly agricultural and innaccurate ?
Bill

Yes, that is how most metalworkers would have started off. It then entirely depends what you do. Different levels of accuracy are required between fitting a hoof iron to a horse, or to make an egine, or to make a camera head to take pictures from inside your blood vessels. And tat is not to say that the fitting of the hoof iron requires less skill than the other tasks. But it is different skills, and different knowhow. Maybe for starters it helps to know, that a machinist apprentice would spend his first year of training working to an accuracy of +/- 0.1mm. That is what I would recommend you initially aim for, at least for items small enough to be carried by hand. General metalworking reading is all you need, there is some good stuff available free online from the US armed forces. You will know when the +/- 0.1mm is not anymore accurate enough, and then there is still time to read about metrology and machine tool testing etc. I guess its like a beginning woodworker would not set out to learn about intricate intarsia or Holzapfel turning before he can make a basic stool with basic tools. Chris

Thanks for the answers

Will order the books from local library - they may have to get them from Melbourne somewhere.

Reason that I started the thread was that I have been reading up on the Gingery stuff and have joined their forum

Was wondering about the feasibility of building a metal lathe out of slabs of Aluminium similar to the Shopnotes lathe - photo attached - see issue 73 for how to make - Am 3/4 way through making the wooden version.

The guts of this shopnotes one is basically slabs of plywood bolted together with a hole that lines up through 4 of the pieces

Will have a read of the books


Regards

Bill

Thanks for the answers

Will order the books from local library - they may have to get them from Melbourne somewhere.

Reason that I started the thread was that I have been reading up on the Gingery stuff and have joined their forum

Was wondering about the feasibility of building a metal lathe out of slabs of Aluminium similar to the Shopnotes lathe - photo attached - see issue 73 for how to make - Am 3/4 way through making the wooden version.

The guts of this shopnotes one is basically slabs of plywood bolted together with a hole that lines up through 4 of the pieces

Will have a read of the books


Regards

Bill

shopnotes lathe.jpg

Sorry I'm not really answering the recommended reading part but here goes anyway.

If you are making things to work together then understanding tolerance or permissible error as prefer to think of it is important, for example if something is to big or too small it will not work. This notion of permissible error is in any kind of work that involves the manipulation of physical things and you know it intuitively, but it is a concept more clearly defined in science and engineering

This I'm sure you already know but I will say it anyway. Nothing is straight, square flat, round, etc. When taking a measurement you are always using something as a reference, it could be a tape measure,your thumb or a laser interferometer you are making a relative estimate of magnitude and because it can only ever be an estimate there is an difference between its true magnitude and the estimate. So the best you can say is that the true magnitude is somewhere between magnitude A and magnitude B. The difference between magnitude A and magnitude B is your measuring error. If you are working with standard measurement like mm or pounds then the reference is also a estimate and has an error, its error arises from its comparison to a defined standard, typically these errors are very small compared to the total range of measurement of the reference used so are not often quoted with a measurement. The total error is the sum of these errors.

One of the reasons why a surface plate is of particular interest is that is one of those reference standards. The surface plate it is the workshops or labs primary standard for flatness. Now the better the standard the smaller the relative estimation of error can be. One of the interesting qualities about flatness is that you can build off it for other geometries like squareness, straightness, roundness, parallelness etc.

The important thing to note here is that it is all relative and additive. An error in a standard adds to an error in estimation. The typical rule for the estimation of magnitudes is that you should use a "working" standard that has 10 times more resolution than the magnitude of the permissible error. For a concrete example: If you are making something that an error in estimation of 1mm will make it not work/fit then you should use a method of estimating that will give you 0.1mm resolution. That being said as low as 2 times the resolution is workable in this case that would be 0.5mm.

-Josh


.

WOW Josh....For a "metal man" you are very articulate
I have been a Carpenter for 40 plus years. Accuracy is always a prime objective. With timber there will always be variances due to the nature of the material. Man made sheet goods are a lot better with consistency than ever before, but they are made from natural commodities and so they are not 100% consistent. In practice it is not a problem because the measurements we work to have a fair tolerance for =/- (industry accepted if you like.

I once did a coarse in Workshop Practice where we were taught how to use a metal lathe and a brief go at a milling machine. Did you know there is a whole civilisation that lives below a mm? I had to reevaluate reading measurements. Learning how to achieve an "oil clearance" or an "interference fit" has to be learned and appreciated to know where they are appropriate.

There was a story going around about a class of Fitters and Machinists apprentices. The teacher gave them an exercise to read a micrometer. The kids answers were generally good except for one kid. The teacher in sheer frustration said to him "Well how many thous are in an inch?" the reply...."I dunno, but there must be a lot of them!"

A dear old Tradesman I worked with said that if you can read a drawing and can mark out and measure you can make something out of any medium if you have the right cutting tools. He was right. I have been able to work with timber, sheet materials, concrete, steel in fabrication and machining, glass...its just a matter of having a go.