Thread: good quality micrometers +/- 0.0001" really and truly ?

Hi Bill,

Originally Posted by steamingbill

Although some sort of estimating "between the lines" is possible.
Bill

I used to do this with analog electrical test equipment. Different field but the "principle" was the same. Improve the accuracy of the reading by "guesstimating" between the marks.

Then one day I read an article about human behaviour relating to exactly this process. I don't remember much about the article just recall that it made me seriously question the process and my motivation.

If I remember the argument it ran something like this...

If a scale has a marking at "1" and a second marking at "2" with no markings in between, at what point does "1" become "2"? If you are really honest with yourself then "1" becomes "2" at the 2nd mark and not ever before the 2nd mark.

Keeping in mind thermal stability, calibration, parallax reading errors, your state of physical and mental well being at the time (to name a few variables) how well are you really able to pick "1.5"? (There being no "1.5" mark to read)

Still up for guesstimating?? What if your instrument has a log scale? (I now remember that the instruments I was guesstimating had log scales!) What if you have a digital instrument that gives a continuously flickering reading of "1" then "2" then back to "1"? is that really "1.5"? do you understand the sampling and the machines internal bit rate as it relates to accuracy?

Lets go the other way....

Once in a wild flight of fancy, I bought a Mitutoyo 0-25mm mic which would read to 0.001mm from ebay. Got it at a pretty good price. Nice bit of kit....

My eyes aren't quite what they used to be, sadly. I can't easily read the vernier scale for 0.001mm. Damn. Need a magnifying ring lite or similar....

The wild flight of fancy really got out of hand when I considered how I would need to checks it's accuracy. A nice set of second hand Starrett Croblox in 0.5 specification were the obvious answer. Well of course how did I know they were within specification?? I got the blocks calibrated of course! Now once I get the workshop fully insulated, air conditioned and thermally stable all will be well, won't it??

We will not discuss the cost of this episode

That episode was somewhat reality driven. One question we should all ask ourselves is "what level of accuracy do I want to be able work to?" As RayG points out you need to be able to measure to one decimal place more than that. I arbitrarily decided on 0.01mm as my level of accuracy (hey, I was new to this stuff at the time!) without fully understanding exactly how high my goal was and the consequences that would flow from it. Another understanding that I have had since then is that while you may have a lofty goal for accuracy, for a lot of work a lower level of accuracy is acceptable.

One last point. After struggling with the vernier scale on the mic, I decided that digital scales with f'n big characters was the way to go and promptly bought a 450mm set of digital verniers with thumping big readout. I love it.

Cheers

The Beryl Bloke

Originally Posted by Theberylbloke

I used to do this with analog electrical test equipment. Different field but the "principle" was the same. Improve the accuracy of the reading by "guesstimating" between the marks.
...
One last point. After struggling with the vernier scale on the mic, I decided that digital scales with f'n big characters was the way to go and promptly bought a 450mm set of digital verniers with thumping big readout. I love it.

One thing that normally slips people's minds is that there is an uncertainty on measurement too. For analogue equipment this is usually taken as plus or minus 1/2 the smallest division. So, while you can estimate that the reading is 1.5, it is really 1.5 +/- 0.5. This assumes the reading device is accurate too. A typical digital caliper has an uncertainty of 0.05mm or such like, so when it says something is 1.5, it is really 1.5 +/- 0.05. Analogue devices can have an uncertainty quoted of say "0.5% FSD +.01" (FSD = full scale deflection).

In metrology terms measuring devices are spoken of as having 3 characteristics - accuracy (how close they are to the 'real' number), repeatability (how good they are at getting to a number consistently) and precision (how refined is the scale/ readout). These are normally all lumped together as uncertainty. We've had this discussion before I think - whether DTI's make absolute or relative measurements.

To get absolute accuracy then you do need to go with the calibrated gauge blocks etc, but really in a home workshop a lot of the measurement is relative - I may want to turn a piece of bar down to 18.95mm, but that is so it is in clearance with a 19.05mm hole - which I've probably measured with the same calipers. Close fits on bought in parts are really the only thing to worry about (bearings for example) but they can be measured too and because of the tolerances they are made to, probably considered a reference standard anyway (that is, if you measure a bearing with a nominal 25mm OD and get 24.9mm, the smart money is usually on your measuring equipment reading incorrectly)

Michael

Originally Posted by Michael G

In metrology terms measuring devices are spoken of as having 3 characteristics -
accuracy (how close they are to the 'real' number),
repeatability (how good they are at getting to a number consistently) and
precision (how refined is the scale/ readout).

Michael

Thanks Michael - have been wondering about these things - I see various things on the internet quoted as having "resolution".

In your 3 point model above is precision the same as resolution ? One simple definition that I read and remember is "the smallest discrete thing that your eyes can see is called resolution and for human eyeballs that is approx 0.1mm or approx the thickness of a piece of paper". Although I think I can see defects much smaller than that on a turned cylinder.

As an aside, can you feel things that you cannot see ie a .0001" step in a smooth cylinder wall ? Will do some tests in shed.

I suspect that in some of the ads I read there is some confusion regarding the definition and appropriate use of the various terms.

A quick google shows discussion on various forums regarding what these 4 words mean and how they are quoted or misquoted in advertisements.

Is there a well recognised machinists glossary somewhere that people use - Maybe an appendix in a measuring standard ? Anybody know ?

I guess at the end of the day, for me in my own shed, its all fairly academic as somebody pointed out above - if things fit together then you've done it right, and you use the measuring tools to help you to get there.

Its when I interact with the outside world that my measurements have to agree with everybody else's.

Bill

Originally Posted by steamingbill

As an aside, can you feel things that you cannot see ie a .0001" step in a smooth cylinder wall ? Will do some tests in shed.

You sure can. On the surface grinder chuck you should always run your hand over it to make sure it is clean-you will feel specs of grit way smaller than you will see. Same if you have a "wave" in a piece of bar-you may not be able to see the change in size but you will feel it if you run your thumb and a finger along either side of it.

Resolution is the smallest step you can measure with whatever tool (in this case anyway) A tape measure or ruler might have a resolution of 1mm. Your mics a resolution of .01mm, TBB's mity mic has a resolution of .001mm, Rays interferometer has a resolution of .00001" (or is it .0000001"?).

The other thing to note about TBB's digital caliper-the resolution of these digital units is ofter far greater than there accuracy and repeatability. You find yourself counting on that .005mm or .0002" (thats what my mit absolute resolution is) because its there in big bold letters. The bottom line is though you cannot measure to that level accurately (not such a problem for us) but more importantly you cannot do it repeatedly either.

FWIW i think .01mm is all you can expect from a mic with a ratchet thimble, you need a friction thimble to get repeat readings under that.

TBB- I have a .001mm 0-25 mity mic too- i would class my eyesight as very good but i still have trouble reading the vernier scale.
Ew

When do the Metroligists take their hand off it.

Originally Posted by pipeclay

When do the Metroligists take their hand off it.

They wouldnt touch it in the first place.

You beat me to it Stu, if you touch it it may get warm and start to expand.....

Originally Posted by Ueee

You beat me to it Stu, if you touch it it may get warm and start to get bigger.....

Be nice... .... I can feel things smaller than I can see...



Seriously, when you wipe your hand across a surface plate, you can feel the tiniest bit of dust.... also when spreading out blue, Marco and Phil, recommended using your hand, that way if there is any grit or dust you'll feel it.

Ray

Originally Posted by Ueee

You beat me to it Stu, if you touch it it may get warm and start to get bigger.....

Well now wouldnt that all depend?........ it gets bigger but reads smaller.

Originally Posted by Ueee

Rays interferometer has a resolution of .00001" (or is it .0000001"?).

.00001" would be my master height gauge. Rays interferometer would add at least two 0's I think.

On the repeatablity side of things, remember the manufacturers spec are for "the worst one they let out the door" so most would be much better(Of course you cant be sure which one you have lol).
When I was testing DTI's I was reading "between the lines" as some times it's right in the middle of "not 0.00005" and "not 0.00006" so I called it 0.000055" as I recall all the readings* (I repeated each test 3 times I think) agreed within 0.00005"

*one good thing about having a crap memory is I couldn't remember what the reading was from the last test so wasnt tempted to massage the figures. Though I'm likely recalling it all wrong anyway.... it will be in the DTI thread somewhere

Stuart

Originally Posted by Stustoys

On the repeatablity side of things, remember the manufacturers spec are for "the worst one they let out the door" so most would be much better(Of course you cant be sure which one you have lol).


Stuart

I think you may be an optomist.

Dean

Originally Posted by Stustoys

.00001" would be my master height gauge. Rays interferometer would add at least two 0's I think.

The interferometer resolution is 0.01 micron, or 0.3 millionths. 0.0000003"

Ray

Originally Posted by steamingbill

Is there a well recognised machinists glossary somewhere that people use - Maybe an appendix in a measuring standard ? Anybody know ?

I guess at the end of the day, for me in my own shed, its all fairly academic as somebody pointed out above - if things fit together then you've done it right, and you use the measuring tools to help you to get there.

Its when I interact with the outside world that my measurements have to agree with everybody else's.

Bill

Hi Bill,

The bible on this would be an Australian Standard where all technical words well pretty well defined. (Dare I say, "to an accuracy of....") Having said that, I do have a copy of the standard applicable to gauge blocks. It talks a lot about uncertiantity and other large words that I found tended to make my head hurt.

Cheers

The Beryl Bloke

Originally Posted by Theberylbloke

...................One question we should all ask ourselves is "what level of accuracy do I want to be able work to?" As RayG points out you need to be able to measure to one decimal place more than that..............

Ah, but that is still the wrong question. It still is the question that made you waste money on metrology items that are of no use to you.

The real question you have to ask yourself is, whether or not you want or need to work to absolute dimensions.

- If you are into machining to earn money, chances are you have no choice other than work to absolute dimensions. You work from exactly specified drawings, and cannot afford your work to be rejected.
- If this is a hobby for you, chances are you do not want to work to absolute dimensions. You can fit things together to micron precision, without ever knowing the absolute dimensions.

There are exceptions of course. If you are into it to earn money, but what you do resembles more blacksmithing work, you may well get away without absolute dimensions. If this is your hobby, and metrology happens to be part of this hobby, you may want to work to absolute dimensions even if it is not necessary for what you do.


By the way, that elusive 0.001mm micrometer is just about the cheapest bit. It lets you measure shafts. But how do you measure bores, anything capable of truly 0.001mm resolution is going to cost a fortune, many times more than the micrometer. What use is that micrometer, if you can only make a shaft for a ball bearing to tolerance, but not the bore?

This is what Chris is pointing at Bill (and anyone else who doesn't know) I have never seen a cheap one of any size (that is not made in China) Tesa 40 50 0 002 Internal Bore Micrometer Dreipunkt Innenmikrometer | eBay

Ew