Do Errors in Stacked Gage Blocks Add or Cancel?

Steve H. Graham · Nov 21, 2016

I have a question for a real metrology expert who knows something about gage blocks.

People are telling me to get gage blocks to calibrate my micrometers. Gage blocks come in different grades. The cheap ones have a tolerance of +/- 0.0005". The good ones are much more accurate. I wouldn't mind blowing $150 for +/- 0.0002", but I don't machine for a living, so I don't want to spend much more than that.

To calibrate a big micrometer, I might have to stack a bunch of blocks. If I do that, are the errors going to add up, or are they likely to cancel out so the overall error is small?

It would depend on whether manufacturers are likely to err consistently on the low or high side. I have no idea whether that's true or not, but I thought someone here might know.

If I have to stack five or six blocks to calibrate a micrometer, and I'm using cheap blocks, I could end up a thousandth or more off if the errors add up. My micrometer would be only slightly more accurate than calipers.

Someone suggested getting standards, but I notice they tend to claim tolerances of +/- 0.0002". They sound like they're not a whole lot better than relatively cheap blocks. To measure tenths reliably. I would have to send the standards in to be measured, at considerable expense. Am I wrong about that?

DMF_TomB · Nov 21, 2016

Steve H. Graham said:
I have a question for a real metrology expert who knows something about gage blocks.

People are telling me to get gage blocks to calibrate my micrometers. Gage blocks come in different grades. The cheap ones have a tolerance of +/- 0.0005". The good ones are much more accurate. I wouldn't mind blowing $150 for +/- 0.0002", but I don't machine for a living, so I don't want to spend much more than that.

To calibrate a big micrometer, I might have to stack a bunch of blocks. If I do that, are the errors going to add up, or are they likely to cancel out so the overall error is small?

It would depend on whether manufacturers are likely to err consistently on the low or high side. I have no idea whether that's true or not, but I thought someone here might know.

If I have to stack five or six blocks to calibrate a micrometer, and I'm using cheap blocks, I could end up a thousandth or more off if the errors add up. My micrometer would be only slightly more accurate than calipers.

Someone suggested getting standards, but I notice they tend to claim tolerances of +/- 0.0002". They sound like they're not a whole lot better than relatively cheap blocks. To measure tenths reliably. I would have to send the standards in to be measured, at considerable expense. Am I wrong about that?

.
piece of paper on box says error amount of each block like like +20 or -30 which is in millionths. you have to add nominal value AND error value to get exact value
1.00002
0.99997
.
no different than when i set indicating gage to ring gage if ring 2.0000 and +50 thats 2.00005"
.
you got errors on day 1 when first bought. just a matter of error amounts
.
obviously you can measure a bore thats different size than ring gage you just need to know what indicating gage should read when in ring gage like +.00005"

Johnnybar · Nov 21, 2016

How big is this "big" mic?

Gordon B. Clarke · Nov 21, 2016

Steve H. Graham said:
People are telling me to get gage blocks to calibrate my micrometers. Gage blocks come in different grades. The cheap ones have a tolerance of +/- 0.0005". The good ones are much more accurate. I wouldn't mind blowing $150 for +/- 0.0002", but I don't machine for a living, so I don't want to spend much more than that.

Who are "people" and how accurately do you need to calibrate? If you don't machine for a living what do you use your micrometers for?

Don't mean to be impolite but the information you give doesn't make giving helpful advice easy. How many gauge blocks do you expect to buy for $150?

± tolerances on gauges should mean that statistically they should even out errors. Again, in theory, the more you stack the more they should even out.

scraper · Nov 21, 2016

ive never seen even the cheapest set of gauge blocks with those loose tolerances. even your cheapest set of shop spacer blocks are usually within a tenth or two. you ought to be able to get from .1000" to 9.0000" in .0001" increments using no more than 4 blocks if you have a standard 81 pc. gauge block set. other than that, do what DMF-TomB says, and pay attention to how you handle both the blocks and the mics. your mics should have come with standards unless you bought used mics and the standards were lost, in which case you could call mititoyo or starrett and order some.

Steve H. Graham · Nov 21, 2016

Gordon B. Clarke said:
± tolerances on gauges should mean that statistically they should even out errors. Again, in theory, the more you stack the more they should even out.

This is all I'm really after. I don't want to bog people down with trivia about what I machined today. It's just a general inquiry.

So it sounds like stacking a few blocks should give me an error pretty close to what I would expect from a single block.

The real Leigh · Nov 21, 2016

Gordon B. Clarke said:
± tolerances on gauges should mean that statistically they should even out errors.
Again, in theory, the more you stack the more they should even out.

Correct.

And to make matters worse, gage block faces are seldom exactly parallel.

A good cal cert will show the block length in the middle and at all four corners.

But this is all of strictly academic interest.
Any decent gage block should be much more accurate than your mics.

- Leigh

Steve H. Graham · Nov 21, 2016

The real Leigh said:
Any decent gage block should be much more accurate than your mics.

- Leigh

This is the thing that surprised me. Micrometers purport to measure to the tenth, but lower-end blocks have tolerances bigger than that. You did say "decent gage block," though, so maybe you're talking about better blocks.

The real Leigh · Nov 21, 2016

Hi Steve,

Good blocks have errors measured in millionths of an inch per inch.

The precision standard used in metrology is that the gage used should be 10 times more accurate than the instrument being calibrated.

So to calibrate a micrometer spec'd accurate to a tenth (0.000100"), you need a block with an error spec of one hundredth of a thousandth (0.000010" = 10 millionths).

From the Mitutoyo catalog US-1002 page E-4 (from ASME B89.1.9-2002) the limit of error on a 1" gage block in four different grades is: 3, 6, 12, or 24 microinches.

- Leigh

Steve H. Graham · Nov 21, 2016

Thank you, Leigh. So basically it sounds like you have to spend some money, and not $150 for cheap blocks, if you really want your mikes to be within a tenth or two.

The real Leigh · Nov 21, 2016

Steve H. Graham said:
Thank you, Leigh. So basically it sounds like you have to spend some money, and not $150 for cheap blocks, if you really want your mikes to be within a tenth or two.

Exactly.

You never save money by spending less.

You always save money by spending more.

Those are universal truths (within limits like where the item was made).

- Leigh

6061Mike · Nov 21, 2016

Steve H. Graham said:
Thank you, Leigh. So basically it sounds like you have to spend some money, and not $150 for cheap blocks, if you really want your mikes to be within a tenth or two.

Well... that and measurement technique, temperature... etc. The list goes on and on for people who routinely measure to this spec.

I will chime in again about the comment you aren't a machinist, but wondering how accurate your 'big' (which you haven't said yet) mics can/will be?? :crazy:

Don't get me wrong, I'm not trying to be the turd in the punchbowl, but the aforementioned items don't add up.

Honestly, when I get above the 2-3/3-4" range, which I can't get a proper 'feel' for, and I need to be within tenths, I always verify with something else, indicator, etc. Really, how big is your 'big' mic, that will give us better info to go on??

And yes, you have to spend money, and I mean MONEY, to be really accurate. BUT, your $150 gage block set will probably do, like others said, even those are measured into millionths, not tenths...

Steve H. Graham · Nov 21, 2016

It's a complicated story. I was thinking about getting some inside micrometers. Then I wondered about calibrating them. Then I started looking at large micrometers, like 6" and up. Not for a particular job, but just to have on hand, the way you have calipers and screwdrivers on hand. I got into reading about them, and I thought I needed big standards to go with them, and then I read about the dubious accuracy of standards. That got me thinking about blocks, and then I started wondering if the errors canceled, because I knew about error propagation and random walk.

I don't want to give everybody a headache or a big homework assignment. It was just a general question. I found some blocks which supposedly have a tolerance of +0.000004/-0.000002, and the price wasn't bad, so maybe I'll invest and be satisfied.

This has been very enlightening, so thanks.

The real Leigh · Nov 21, 2016

For a dedicated task such as this, you might consider buying individual blocks.

McMaster-Carr and others sell Mitutoyo blocks in whatever tolerance and lengths you want.

- Leigh

Gordon B. Clarke · Nov 21, 2016

The real Leigh said:
For a dedicated task such as this, you might consider buying individual blocks.

McMaster-Carr and others sell Mitutoyo blocks in whatever tolerance and lengths you want.

- Leigh

I read that twice. Will Mitutoyo in the USA really make ".... blocks in whatever tolerance and lengths you want"?

That doesn't sound like the Mitutoyo I know.

The real Leigh · Nov 21, 2016

Gordon B. Clarke said:
I read that twice. Will Mitutoyo in the USA really make ".... blocks in whatever tolerance and lengths you want"?

Hi Gordon,

I did not say "make".

I said "sell".

You choose which blocks you want from among their standard products.

My point was that the OP could probably assemble a focused set with tighter tolerance at lower cost than buying a complete set, of which most would go unused.

- Leigh

Gordon B. Clarke · Nov 21, 2016

The real Leigh said:
Hi Gordon,

I did not say "make".

I said "sell".

You choose which blocks you want from among their standard products.

My point was that the OP could probably assemble a focused set with tighter tolerance at lower cost than buying a complete set, of which most would go unused.

- Leigh

That's not how I understood what you wrote but now I can go to bed knowing you know the same Mitutoyo as I do :cheers:

hanermo · Nov 21, 2016

0.000004 == 0.0001 mm.
0.1 um or 0.1 microns.
No way- As a specified plus accuracy.

I do believe good gage blocks are or were lapped to something similar in size/flatness/parallelism, but not as in *specified accuracy*.
The specified accuracy is the killer.
Smooth - yes. Accurate - no.

Actual size better than 1 micron error .. sure. Yes.
Better than 0.1 microns ... well.. heck no.

Guys lap telescope mirrors by hand in basements to this level, or even better, yes. But..
I know of no process than can make cheap mass produced components to near 0.1 - 0.2 micron flatness / size.
Does anyone ?

I believe, think, lapping machine vendors quote about 0.5 microns..
The quoted level implies about a 5-10x better process, to end up within +0.1-0.4 um.

Any link to the provider please ?

Steve H. Graham said:
I don't want to give everybody a headache or a big homework assignment. It was just a general question. I found some blocks which supposedly have a tolerance of +0.000004/-0.000002, and the price wasn't bad, so maybe I'll invest and be satisfied.

This has been very enlightening, so thanks.

dcsipo · Nov 21, 2016

Steve H. Graham said:
I have a question for a real metrology expert who knows something about gage blocks.

People are telling me to get gage blocks to calibrate my micrometers. Gage blocks come in different grades. The cheap ones have a tolerance of +/- 0.0005". The good ones are much more accurate. I wouldn't mind blowing $150 for +/- 0.0002", but I don't machine for a living, so I don't want to spend much more than that.

To calibrate a big micrometer, I might have to stack a bunch of blocks. If I do that, are the errors going to add up, or are they likely to cancel out so the overall error is small?

It would depend on whether manufacturers are likely to err consistently on the low or high side. I have no idea whether that's true or not, but I thought someone here might know.

If I have to stack five or six blocks to calibrate a micrometer, and I'm using cheap blocks, I could end up a thousandth or more off if the errors add up. My micrometer would be only slightly more accurate than calipers.

Someone suggested getting standards, but I notice they tend to claim tolerances of +/- 0.0002". They sound like they're not a whole lot better than relatively cheap blocks. To measure tenths reliably. I would have to send the standards in to be measured, at considerable expense. Am I wrong about that?

hey Steve,

You are off by a factor of 10....grade B or grade 3 gage blocks are accurate to 0.00005" grade A or grade 2 is about 0.00002" and AA or lab grade is 0.00001" so those are 50 20 or 10 millionth of an inch, that means that stacking blocks with 3 junctions you are within 0.00015" of an inch with shop grade blocks In an 81 piece set almost all sizes up to 5" can be made with 4 blocks or fewer. if i needed to have over 5" lengths i would buy extra long blocks for each length i need or just an extra set of 1", 2", 3" and 4" blocks. you can get a cheap set of long blocks from Shars SHARS 8 Pc GRADE B 5-2" LONG LARGE GAGE BLOCK CALIBRATION W/ USA NIST CERT NEW | eBay

dee
;-D

Derek Smalls · Nov 21, 2016

hanermo said:
0.000004 == 0.0001 mm.
0.1 um or 0.1 microns.
No way- As a specified plus accuracy.

I do believe good gage blocks are or were lapped to something similar in size/flatness/parallelism, but not as in *specified accuracy*.
The specified accuracy is the killer.
Smooth - yes. Accurate - no.

Actual size better than 1 micron error .. sure. Yes.
Better than 0.1 microns ... well.. heck no.

Guys lap telescope mirrors by hand in basements to this level, or even better, yes. But..
I know of no process than can make cheap mass produced components to near 0.1 - 0.2 micron flatness / size.
Does anyone ?

I believe, think, lapping machine vendors quote about 0.5 microns..
The quoted level implies about a 5-10x better process, to end up within +0.1-0.4 um.

Any link to the provider please ?

I'm not sure what you are getting at here, that better than .1 micron blocks don't exist, OR, that they don't come cheap?

http://starrett-webber.com/GB40.html

Not shown here, but grade AAA blocks are also available.

Do Errors in Stacked Gage Blocks Add or Cancel?

Aluminum

Diamond

Plastic

Banned

Aluminum

Aluminum

Diamond

Aluminum

Diamond

Aluminum

Diamond

Hot Rolled

Aluminum

Diamond

Banned

Diamond

Banned

Titanium

Diamond

Stainless

Similar threads