Metal Scale (SIP-Societe Genevoise) change in size over time?

[greif]

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.

 

[Ninja]

Not an answer, but some perspective...

depending on the type of steel...let's say TCE is ~6.5x10^-6 mm/mm/F
Let's say that 20C = 68F...the one degree F difference across 400mm is ~3um all by itself.

How accurate is the temp each time? another 3-5um worth?
How much error reported in the measurements in 1960 and today?

And lastly...19um on a scale with engravings every mm across 400mm is not so bad...especially when your markings are only every mm.
It would be interesting to see how much error there is between marking pairs...

 

[greif]

The uncertainty reported with the prelim 2016 report at 400mm was 2.5um.
The original SIP report did not state uncertainty. The chart was in the form of a correction to be made with each mark (add or subtract), in tenths of microns.

The SIP manual states the thermal expansion of the scale and for the difference between 20C and 69F ((0.56C), the scale will be 2.58um longer at 69F.

The guaranteed measurement accuracy of the MU-214B that the scale came from was 2um in 400mm (from their lit.)

Still - 19 um is quite a bit!

 

[Ninja]

Quite a bit....compared to zero.

does 19um matter for a scale where 1000um is a single tick?

 

[greif]

The number of graduations is not particularly important- it is to be used to check the accuracy of various measuring machines...... so the 19um is rather important (at least the determination if it really is off by 19um!)

 

[Ninja]

Silly question coming...

1. If this is a critical part for machine setup, and
2. has to be accurate to within much better than 19um on 40cm, and
3. has been used since 1960....

Why is it being questioned now?

Point of the question:
If you've been using it since 1960, and 19um is way too much...then you've been making bad parts for 50+ years?

Or if you haven't been making bad parts for 50+ years...19um is really that important?

Easiest solution....buy or make a new one.

 

[greif]

"If you've been using it since 1960, and 19um is way too much...then you've been making bad parts for 50+ years?

Or if you haven't been making bad parts for 50+ years...19um is really that important?"

Nope, pretty much been on the shelf and not used since we acquired it. Being such a high quality scale (Swiss!) I would have like to have used it as a handy (unbreakable) check tool.
My concerns are;
-is metal really that unstable?
-Did this accredited lab really do a proper job?

 

[Ninja]

Ah, understood.

I'm not sure there are really answers to your question...but food for thought...

metal is not unbreakable...especially in the um ranges.
Drop it, leave it in the sun...um is not unreasonable to expect.

for um over 400mm...everything is unstable to some degree, even granite.

When faced with a choice between questioning a current certified lab and questioning "Swiss made in 1960"...I would likely question the 1960 side of things. (All the same, what was the reported measurement error on your recent test?)

HTH

 

[greif]

Summary;
SIP cert from 1960 says off ~.4um at 400mm at 20C.
new 2016 cert says big by 0.019mm at 69F

Taking the temp. difference into account (SIP lit states the coef. of exp.), the new cert says it is 16.5um bigger than 400mm.