Question About Inside Micrometers

[GlennW]

The second is a name brand from Poland and was certified by a cal lab in the very late 1900's. Both are in top notch condition. It is good to actually know the block size.

They can actually change, hence the need for calibration on regular intervals. Some as often as a few months depending on their use. Annual calibration is fairly common though.

Study up on it, as it's a metallurgical thing that causes steel blocks to change.

I've had a few blocks cancelled over the years.

[Harold_V]

What Glenn said. Because steel and other alloys are quite active, they are always on the move, in spite of having been heat treated and cold cycled. Blocks change, and can grow or shrink. One of mine had changed about .000050" from the original calibration to the next calibration. It's not just a wear concern, in other words.

On that subject, a good set of blocks will come with a set (or two) of "wear blocks" which are used on each end of a stack. That limits the exposure of the precision surfaces of the balance of the blocks with less than desirable materials. I make it a habit to use mine at all times.

Harold
edited to correct error in size from .00050" to .000050"

[ctwo]

I was in bliss. Perhaps these would be more stable: http://www.msi-viking.com/Starrett-RC-8 ... _2920.html

and this seemed to be a good reference: https://www.nist.gov/sites/default/file ... ono180.pdf

[NP317]

I was in bliss. Perhaps these would be more stable: http://www.msi-viking.com/Starrett-RC-8 ... _2920.html
[snip]

Yikes! >$8K!!
But I get it.
~RN

[Carm]

The PM dudes claim errors cancel instead of adding.

It's funny; people say to use blocks to calibrate mikes, but when they evaluate their blocks, they say they're "dead on"...according to their mikes.

Shars has a reasonable set of AS-0 blocks. I think that might not be a bad thing to have. Says "Tolerance +0.000004/-0.000002."

If that's even close to correct, they will be better than I will ever need.

Yes, that's funny, but not illogical if the tacit rule of thumb for metrology is kept in mind.
The error in the verification device has ~ ten X less error than the measuring tool.
Micrometers as per OP can only resolve so close. One thousandth of an inch is reasonable and repeatable, and expected across shop usage by operators who possess different "feel" for the tool.
Splitting that is tricky and becomes more so with more than one machinist/mic. I can assure you that many machinists sweat those numbers.

Using the tolerance given for the Shars blocks, you would never be able to read the error with a mic. So the reverse verification actually makes sense and would were the error 10x greater.

[SteveHGraham]

If you buy blocks to check your mikes, you can't use your mikes to check the blocks. Unless your mikes have been checked. And if you could do that, you wouldn't have bought the blocks.

You wouldn't measure your height with a ruler and then say, "Let's see if the ruler is right. I'm six feet tall (according to the ruler), so we'll check it against me."

Because steel and other alloys are quite active, they are always on the move, in spite of having been heat treated and cold cycled. Blocks change, and can grow or shrink. One of mine had changed about .00050" from the original calibration to the next calibration.

This appears to bring us back to, "You can't measure to the tenth without paying someone to check your equipment," so we're back where we started.

The clear conclusion seems to be that you have to spend a lot of money or forget about measuring to the tenth. So if you're cheap, the tenths marks on micrometers have to be ignored, and you have to be happy with the nearest half-thousandth.

It sounds like the old caliper argument. People claim their trusted Starretts are accurate to within a thousandth, but they're really just looking at repeatability, or they're squeezing the jaws just right to get the results they want to see. They don't really know how accurate their calipers are.

I guess now I should focus on figuring out how to split a thousandth reliably. As for sizing stuff for press fits, it sort of sounds like emery cloth is the only hope for a cheap machinist.

[liveaboard]

it sort of sounds like emery cloth is the only hope for a cheap machinist.

That's me; but I just bought a couple of rolls of emery and I don't think it was cheap at all!

Just so you know how cheap a cheap machinist can be...

[Magicniner]

If you buy blocks to check your mikes, you can't use your mikes to check the blocks.

If the blocks are in test and the micrometer checks out you can, to the spec of the micrometer, but it wouldn't get you anywhere as your blocks are in test to a higher spec than your micrometer can measure and so don't need checking.

It's all about Standards, a man has got to have Standards

- Nick

[Carm]

If you buy blocks to check your mikes, you can't use your mikes to check the blocks. Unless your mikes have been checked. And if you could do that, you wouldn't have bought the blocks.
(snip)

Paraphrasing Nick, and attempting clarification of my statement:

Say you have a gage block of known provenance, 1" nominal, .00002+ (a cheapie)
You wish to verify your mic, it reads 1.0002. The mic (or usage technique) needs further looking into since the test indicates discrepancy.
You wish to verify a gage block for nominal size, you use your calibrated mic. It reads 1.0000.
The gage block is good for the test even though it may be +/- past the fourth digit. The mic cannot qualify the block beyond nominal.
Akin to using a mason's level to determine twist in a lathe bed.

No need to ignore the tenth marks on a mic, they tell you something relative and can be interpolated. In your own shop, very reliable.
There's nothing wrong with abrasive cloth, you can't get grinding tolerance with cutting tools in terms of Ra/RMS.

[ctwo]

It's all about Standards, a man has got to have Standards
- Nick

Yeah, they are so important, that everyone must have their own.

I'm going back to bliss. My gage blocks, for all intents and purposes, are perfect. I would never think of using any of my mics to check my gage blocks. That would be like using a 2x4 to check my calipers. I am going to trust the certs that are with my blocks too, and use the certified values in my measurements, and I'll sleep well at night. The nist paper I linked states 52100 Gauge Block Steel stability as 0.01E-6 per year. Even the old British yard standard only drifted a few millionths of an inch over half a century. The following tested a sample of 52100 steel gage blocks as average stability in -0.1 X 10^-6 in./in./yr and they measured 7 commercial blocks with an average change in length for the seven blocks as -1.1 X 10^-6 in./in./yr.

We should have a contest like Tom Lipton. Someone grind up a standard, we'll ship it around to each other and each measure it to the best resolution we can, and then compare notes. In the end, I bet we have no idea of its actual size.

[Harold_V]

I've commented on measuring to tenths before, and will, once again, relate my personal experience.
When I worked in precision grinding, I could (routinely) measure to .000050" with my Starrett micrometer (1"---all our work was quite small---control devices for launching the missile). We had a Sheffield Shadow Graph at the machine, with which we could verify readings. We also used fine wheels (100 grit), so surface finishes were excellent.

It would be unreasonable for anyone to expect to measure that closely when turning metals. Surface finish plays a huge role in size, as do the bearings of the spindle. It isn't uncommon for lathes to not turn true cylinders, although the degree of error may be small. Still, that eliminates the possibility that one can measure extremely closely, assuming finish was acceptable.

That said, it isn't unreasonable for one to hope to measure to a tenth. It would be more unreasonable to dismiss the concept and fall victim to the notion that it isn't possible, therefore it can be ignored. Less than a half thou, for small work, spells the difference between a successful fit and a failed fit. Work as closely as you can, and do so until that becomes routine for you. You'll enjoy success when requirements are stringent, and you won't be intimidated when called upon to perform at that level. That was the lesson I learned as a trainee.

Harold

[SteveHGraham]

Okay, but...how do you get your home shop mikes CLOSE to a tenth without paying for metrology every year? Let's cut to the chase. Most of us won't be doing A+ work, so what do we use to get to B+?