Can 400mm Swiss (SIP) scale grow by 16 microns in 50 years?

[greif1]

I have a 400 mm scale that came from a SIP MU-214 measuring machine (vintage 1960). It is steel about 20x20mm square with a slightly recessed area on one side which has very fine engravings every mm.

The original calibration data says the max deviation for this scale was 0.7um, when calibrated by SIP in 1960.

I have sent this off to be calibrated by an ANAB accredited cal lab and they have given me a preliminary report of a reasonably gradual stretching of dimensions, ending at 400.019mm

Original SIP cal was at 20 C, the new cal was at 69F.

Is it possible a chunk of steel would grow by 19 microns in 400mm?
Seems like an awful lot.

It looks like the cal temp. of 69F will cause it to be 2.6um big, so it supposedly is 16.4um bigger at 400 mm than the original cal.

 

[Monarchist]

Is it possible a chunk of steel would grow by 19 microns in 400mm?

There is published info on the phenomenon in general.

Metals can be patient critters. Some grain structure changes move at lower rates than even multiple human generations do. 56 years isn't much time on that sort of metric, so, yes, seems possible.

My compensation is to use ONLY "OLD" metrology gear, but that works only because I've no one to please but meself. "Customers" are some other Pilgrim's rice-bowl.



Bill

 

[Gordon B. Clarke]

You seem to believe all calibration measurements are absolute certainties. Try sending your "chunk of steel" to 3 different labs and I'll be amazed if you get the same result from each.

What "measurement uncertainty" is given in your calibration reports?

 

[Monarchist]

You seem to believe all calibration measurements are absolute certainties. Try sending your "chunk of steel" to 3 different labs and I'll be amazed if you get the same result from each.

You are congenitally "amazed", Sir. Or play it well on TV.

Kindly allow the gentleman to retain enough hard-earned cash to still be able to EAT after hiring expensive calibrations done.

They are not exactly team sports one can charge admission fees to watch.

Bill

 

[greif1]

The lab states the 2016 uncertainty as 2.4um
The original SIP measuring machine the scale came from said guaranteed accuracy of 2um in 400mm.
I just looked at the temp. difference of 20C and 69F- it means the modern lab measured 2.58um big due to the temp.

 

[greif1]

Any suggestions on other labs I could send this off to? I only found this one by accident.
It seems that labs that can measure micron precision optically over 400mm , are hard to find.

 

[Monarchist]

Any suggestions on other labs I could send this off to? I only found this one by accident.
It seems that labs that can measure micron precision optically over 400mm , are hard to find.

It would help if you shared where you plan to go with this. If you have contract specs to meet that need that sort of accuracy and 'any' sort of traceability, I can't see starting the march with a 56-year-old scale, no matter how calibrated.

If the sort of accuracy you are fretting over is NEEDED, you'll have to start with a temperature - possibly also particulate - controlled facility. Not many SIP-Genevoise goods were operated in carparks.

With the sort of costs THAT entails, buying newer, better, more automated and integrated metrology equipment should look less costly.

If it is but curiosity? Or bragging rights? Putting an older machine tool back into hobby or prototype service?

Please share.

Others here may have the same equipment, parallel goals, plenty of relevant experience.

Bill

 

[The real Leigh]

Try sending your "chunk of steel" to 3 different labs and I'll be amazed if you get the same result from each.

Hi Gordon,

This is precisely the reason that some large companies maintain their own primary standards labs, as I stated in another recent thread.

See... You agree with me even when you don't mean to.

- Leigh

 

[Monarchist]

See... You agree with me even when you don't mean to.

Only for openers, Leigh. With Gordon, it's a trap!



Bill

 

[greif1]

It would help if you shared where you plan to go with this. If you have contract specs to meet that need that sort of accuracy and 'any' sort of traceability, I can't see starting the march with a 56-year-old scale, no matter how calibrated.
Bill

I don't plan on using the old SIP scope , but was going to use the scale as calibration check for other devices.
(Metal is nice because it does not break easily, and SIP quality is unsurpassed)
It appears that metal is not as stable as I had thought! Long term I will switch over to glass or fused silica scale for a reference tool.

 

[Monarchist]

I don't plan on using the old SIP scope , but was going to use the scale as calibration check for other devices.
(Metal is nice because it does not break easily, and SIP quality is unsurpassed)
It appears that metal is not as stable as I had thought! Long term I will switch over to glass or fused silica scale for a reference tool.

Do your research. May not be an advantage.

I've looked at acquiring an old favorite. A Cadillac Gage Pla-Check. Far handier local ref for transferring settings, surface plate up.

Mind... the stacked gage blocks that make it up are subject to the same issue of age movement. But so are the ones in my gage-block SETS. Old Dearborn Gage one, anyway. Newer import set prolly isn't accurate enough to yet care.

And there you are. Why should I give a taxatwoshits?

Still way closer than anything I have a real-world, not dick-swinging-imaginary NEED for.

Nanotech guys and gals have to eat, too. Leave it to them to put post-it notes on individual electrons.

Bill

 

[Gordon B. Clarke]

Hi Gordon,

This is precisely the reason that some large companies maintain their own primary standards labs, as I stated in another recent thread.

See... You agree with me even when you don't mean to.

- Leigh

It wasn't that I disagreed with you as much as you didn't make me feel more relaxed when flying

I can only hope that those that "go by their own rules" are both trustworthy and reliable

That I don't like to fly is NOT going to stop me from travelling. Logic and common sense tells me flying is one of the safest transport methods. Maybe it's the fact that I know I have no influence in anything that happens once we're "up there".

 

[Heavey Metal]

I have a 400 mm scale that came from a SIP MU-214 measuring machine (vintage 1960). It is steel about 20x20mm square with a slightly recessed area on one side which has very fine engravings every mm.

The original calibration data says the max deviation for this scale was 0.7um, when calibrated by SIP in 1960.

I have sent this off to be calibrated by an ANAB accredited cal lab and they have given me a preliminary report of a reasonably gradual stretching of dimensions, ending at 400.019mm

Original SIP cal was at 20 C, the new cal was at 69F.

Is it possible a chunk of steel would grow by 19 microns in 400mm?
Seems like an awful lot.

It looks like the cal temp. of 69F will cause it to be 2.6um big, so it supposedly is 16.4um bigger at 400 mm than the original cal.

Invar not steel

 

[greif1]

Invar not steel

Nope- the original SIP manual says they used steel in the scale (and state the coeff. of expansion)to match what was expected, would most often be measured on the measuring machine. The machine has a sliding table with the scale embeded, one microscope for looking at the part and another microscope for reading the scale. A very fancy optical filar was built into the scale read microscope to direct read in-be-tween the 1mm divisions to the nearest 0.5um.

 

[The real Leigh]

I can only hope that those that "go by their own rules" are both trustworthy and reliable

Hi Gordon,

The point being that no measurement is exactly correct.

The sole exception is the Metre stick maintained in Paris (IIRC).
It's exactly 1 metre long by law, not by measurement.

All length measurements derived from that metre bar have errors... ALL of them.

The errors may be in parts per million or even parts per billion, but they're NEVER zero.

When I worked at the metrology company, we sent all of our standards directly to the National Bureau of Standards to be calibrated and certified.

All of the cal certs from NBS came back with error terms identified and quantified.

The point of large companies having their own primary standards is that the same error terms apply to all measuring instruments in the company.

There are no absolutes in this world.

- Leigh

 

[Gordon B. Clarke]

Hi Gordon,

There are no absolutes in this world.

- Leigh

Leigh,

You're absolutely right.

Gordon

 

[Screwmachine]

Nope- the original SIP manual says they used steel in the scale (and state the coeff. of expansion)to match what was expected, would most often be measured on the measuring machine. The machine has a sliding table with the scale embeded, one microscope for looking at the part and another microscope for reading the scale. A very fancy optical filar was built into the scale read microscope to direct read in-be-tween the 1mm divisions to the nearest 0.5um.

That's weird, all the Sip scales I've seen were Invar, including the same model as your machine.
Regardless, if it grew and you know the error then it's fine. Sip provided a chart showing the deviation at each mark for those machines back in the day, just make an updated chart.

 

[SteveBausch]

Reminds me of the guys who want to dyno their vintage auto and wonder why it doesn't match the original specs.....

When it goes up for sale, does it matter?

 

[hanermo]

L .. Disagree, but politely.
Your competence is not in Q at all.
And I was in Paris looking at 3 versions of the meter stick 2 months ago, at the "museo de arts metiers".
Have Fotos. Lots of them.
Fascinating stuff.

The meter is, iirc, now defined by a certain nr of divisions or wavelengths of light (x counts/second).
And the nr is absolutely unvarying, and reproducible.

Afaik.

My opinion, no way did the scale grow.
But I cannot state why, nor offer any supporting evidence.

IF the error bar is 2.x microns, and the temp error is 2.5 microns, that leaves about 14 microns.
I suspect SIP had more error than stated ?
And could be the lab had error, but unlikely.

Most probably error, imho, is human cya, and something went wrong in the manufacture of the scale.
Could be its steel and was supposed to be invar, or something like that.
Or it was calibrated as one or the other, by mistake, back in the day.

Much bigger companies have made much bigger errors.
Atomic bombs come to mind, at least 2 major snafus with 1+ guy dead (one dropped the stuff playing with it (dead soon after from rads), iirc the other was the guy dropping a wrench in the missile, or something like that, maybe caught fire).

Or using imperial/metric wrong, and the mars multi-billion-dollar thingies shute did not open - bang goes the weasel.
You would expect nasa to take more care, and they do, but ...
stuff happens.

My experience is that stuff happens, and size or cost does NOT equate to quality.
Telecoms billing, multi trillion $ industry, has about 30% errors.
(And yeah, the analogue holds, but only partly, I know).

Hi Gordon,

The point being that no measurement is exactly correct.

The sole exception is the Metre stick maintained in Paris (IIRC).
It's exactly 1 metre long by law, not by measurement.

All length measurements derived from that metre bar have errors... ALL of them.

The errors may be in parts per million or even parts per billion, but they're NEVER zero.

When I worked at the metrology company, we sent all of our standards directly to the National Bureau of Standards to be calibrated and certified.

All of the cal certs from NBS came back with error terms identified and quantified.

The point of large companies having their own primary standards is that the same error terms apply to all measuring instruments in the company.

There are no absolutes in this world.

- Leigh

 

[greif1]

I have a hard time believing that either SIP or an accredited lab can be wrong by 16 microns, so I will have to get some other measurements. I found a paper by NIST that talked about growth of gauge blocks and various other alloys over time- which is the only reason I think such a distance is possible for the metal to have grown. Some of the examples were twice as bad as this!