Micrometer accuracy in use

[Patio]

I started machining 4 years ago this fall, so I am no authority on anything. I have followed Harold's and many others advise, to work to the greatest degree of accuracy that I am able to, all the time. The reason is, that when I actually need to, I will have the skill to do so.
When I started I had the same question as most, What is close enough? The answer was always the same, "It depends on what you are doing!". What I have learned is, a saw can cut to about .020" and mill or lathe will cut to about .0010" and a grinder can cut to about .0002" any better than that is best done by hand, with a scraper or by lapping. That is a general guide. With practice you can split the smallest of those values to get even closer. I do not have a grinder, I do have a mill and a lathe and neither of them are new, in fact the lathe is older than me. When I need something to less than a .001 fit, I try to hit the mark, but error to the side of over sized. The way it was explained to me to get there is to turn the part within about .010", the last .010" is taken in three passes. The first pass would be .003" then measure, take another .003" and measure. By now you will have a good idea of what is needed for the final pass to hit the mark, it may be .002" or .004". The reason for this, is so the tool is loaded evenly for each pass, which will, when measured (with a micrometer)give you a good idea of what your dials (or DRO) is really telling you, in regards to what is actually happening. If I am I can get within .00075" to .0005" I will polish the last bit off with emery paper. The other thing I have learned is the finish is important when turning for close fits. If the finish is not smooth, there could be a .0015" to .001 of difference between the peaks and valleys of the finish. When the two parts are put together the peaks will wear off. So if the finish is poor and the parts fit well to begin with, in time the part may become loose as the peaks wear off. This is the same reason one does not want to use knurling to help make parts fit, as they may not last, depending on use.

Back to the "It depends" statement. Most jobs have a tolerance factor given by the engineer, and that is how close, is close enough. Lots of surfaces on projects are not critical and +- .010" won't make any more difference than +- .500" Either way, I pick a number and try to hit that within +-.001" every time, just for practice. So every time I am at the machines, it is practice, for when I need to get it right! Even when I am roughing, I measure things as I go, to see if what I think is going to be taken off, is really what I am taking off.

I hope this helps someone newer that me, with the, what is close enough question.
Like I said, I am not an expert, this is just what has worked for me so far.

[Jaxian]

One thing I will just throw in as an addition to the observations on technique. As you have already mentioned one of the things you can notice is the values changing as your hand warms up the micrometer. One thing you can't change is the temperature of your hands. The closer you get the measuring instrument and the thing you are measuring together temp wise the less ability they have to influence each other.

What I am trying to get across is if your shop is 60 degrees and you pull out a part that just got turned and was throwing tan/blue chips and is well over 100 degrees and grab a micrometer that is somewhere in between depending on how long you have been holding it or where it was stored you have a whole lot of moving targets all of which are going to skew any high accuracy reading. It's all about getting everything as consistent as possible. Let the part cool down, let the micrometer warm up and if at all possible get your shop as close to 68 degrees which is the standard temp most measuring tools are bench-marked at. If you can be consistent you can get repeatable high accuracy results. Every time you alter when you measure in relation to one of the moving variables you will have issues.

I had an incident where I was making a small reasonably thin aluminum part. I was a bit stumped as to how to do the next piece and was just kind of brainstorming. I was sitting in my garage where it was 64 degrees (about as cold as winter gets here indoors), I had the drawings in front of me and the part sitting there to look at for reference. I measured it all up and wrote it down. After about a half hour I had what I thought was good next part to fit the existing piece so started to measure again to double check as I dimensioned the new drawing. I was amazed that just sitting on the table a foot or two from me while I breathed on it for a half hour had made all my measurements change by well over .001. I measured a few times and they just got worse as I handled the part.

Now to try to stay consistent I went down and bought a good quality IR temp gun. I check the temp of the cutter, I check the temp of the vise or chuck, I check the temp of the part as it's being cut. It only takes a second to point and shoot and if I am doing something like roughing aluminum the temps skyrocket and I allow for that. Also I read the micrometer or caliper. It's readings move a lot depending on whether its been on the bench in the light or sun or near me, or whether it was sitting in a drawer full of other pieces of cold metal.

I try to make my process so the temps are the same at any one point. Tends to make my measurements and fits much better. Plus those IR guns are so fast and easy to use.

[warmstrong1955]

By memory..... (scary)

Thermal expansion of 6061 aluminum would be equal to 13.0 x 10^-6 x temp increase x L

So...... a 2" pin, with an increase in 50 deg F, would be 13.0 x .000001 x 50 x 2 = .0013"

The coefficient for steel is 7.3....or ....why do pipelines have zigs & zags......

I have a chart of several coefficients somewhere....

Bill

[Downwindtracker2]

I'm not a machinist, rather I'm a retired millwright. If the parts didn't fit, I told someone. And he told the machine shop owner to get his butt down here and get that part back tomorrow morning or sooner. So I used micrometers and telescoping gauges more often for differences than overall sizes.

When I retired, I wanted a fully equipped woodworking shop. I bought older machines and started rebuilding and modifying. That's when I realized I needed a lathe and mill. Sweet,two more to sort out. I had bought a 0-4" Chinese set on sale, as they were cheap, but never used them. I found a 0-6" Polish VIS set at a pawn shop for less money than adding 4-5 and 5-6. Since the lathe spindle nose is metric, from KBC, a 0-75mm set, oddly enough, also VIS.

Now I get to micrometers. The Chinese set had ratchets which I had to calibrate to my twisting, it's quick and easy. Check out THELAZYMACHINIST on YouTube. If the clowns at the mill could get to 1/10ths using ratchets, anyone can. The VIS had friction thimbles and none of the nine worked ! An afternoon later they are really nice, I like them more than the Mits or the Starrett, but I certainly would suggest the ratchets over thimbles.