Thermal expansion of measurement tools 5

[TerryR1]

I have a 75in diameter pilot that needs to fit with a minimal interference. The two parts are made in different places and likely at significantly different ambient temperatures.

I thought I was being smart in calling out different dimensions depending on metal temperature (ie 75.000@80F, 74.989 @60F).

But now I'm thinking that isn't necessary, as the tools used to measure this pilot (ie pi tape, calipers, machine readout) will also thermally expand/contract. So as long as the measurement tool has close to the same temperature as the part it is measuring, I should not need to worry about temperature effects.

Can any of you much better engineers confirm this? Thanks much!

 

[dgallup]

Only if the measuring instrument and the object being measured have similar coefficients of thermal expansion. If it is super precision, it is probably going to have to be measured in a temperature controlled environment (usually air conditioned to something slightly below normal room temp.). Be sure to allow plenty of time for the parts to reach equilibrium.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[JJPellin]

I would not assume that the thermal expansion properties of the measuring instruments are comparable to the part being measured. Most of our precision micrometers are made of Invar (iron/nickel alloy) because it has a very low coefficient of thermal expansion. But, I doubt anyone has an OD micrometer made of Invar that is capable if measuring a 75 inch diameter. I would make a measurement like that using a Faro arm or laser mapping tool. Either of those tools would tend to be accurate over a fairly large temperature range. I would specify the two dimensions under defined standard conditions and require the two manufacturers to meet my dimensional specifications. Depending on the measurement method they used, they would have to certify that it would meet my dimensional requirements at my defined standard temperature.

Johnny Pellin

 

[racookpe1978]

Actually, both tool and part WILL change due to temperature, but, will it MATTER whether one, both, or the other has changed temperature?

Obviously, only you can decide that: It depends on the accuracy you need in the final product.

Two cases should show what I'm talking about. In the first, we were installing a shrink-on, interference fit valve liner with about 10 thousands overlap. If the part were machined so fast it got "hotter" and expanded while in a 70 degree shop, that ID needed to be corrected for a 120 or 130 degree actual temperature so the actual cooled dowm measurement would still have a 10 thousandth interference fit. If the interference fit went down to 8 or 6 thousandths, I'd have lost a substantial part of the locking force. If the part were machined at 80 or 85 degrees metal temperature, and the valve and liner were installed together at 70 or 80 degrees, I couldn't justify any thermal correction in either. The overlap would be maintained by the machined 10 thousandth overlap.

Second case. The parts were outside in the rack at 5 and 10 degrees daytime, and -5 to -10 at midnight. The "shop" was barely hotter at 30 degrees, but the micrometer was at that 30 degrees. While machining the 36 valve actuator throats, the metal got up to 60 or 80 degrees. So, we had to be very careful about temperature of the part, measured temperature of the part (not all of it was warming up), temperature of the micrometer as it varied, and the different thermal expansion coefficients.

But, this machining was for a replacement o-ring groove, so it was more critical (but smaller!) than a longer or less critical application.

Distance matters. If its a small part, it can't expand very much. Temperature difference matters.

 

[tbuelna]

TerryR1-

Most serious machine shops work to QA standards like AS9100 or ISO9001 which control dimensional inspection procedures, including the environmental conditions of the inspection area and calibration of the inspection equipment. Inspection equipment is typically calibrated for the temperature conditions the inspection area is maintained at. It is also typical for high-precision machine tools to be located in a controlled environment for the same reason.

The machinist will always prefer to inspect a feature like your close-tolerance 75" diameter pilot while the part is still fixtured on the machine. If the part is removed from the machine to take a measurement and the feature measures out of tolerance, then the part must be precisely re-fixtured on the machine for rework, which can be very difficult to do.

As a suggestion, if your main concern is the relative fit between the mating pilot surfaces rather than the actual dimensional precision of each part, then you could have one vendor machine their part to match the other vendors part. This would eliminate the cost of precision machining one part, and it would allow for all of the dimensional tolerance previously used for both parts to be applied to the part being match machined.

Hope that helps.
Terry

 

[ajack1]

To my understanding unless otherwise specified should be measured at 20 degrees C 68 degrees F. This is what all standards should be kept at and all calibration done at.

If you actually start specifying what temperature things should be measured at you will add a large cost increase and also throw in variables like you have described.

 

[TerryR1]

Wow thanks much guys.

My 2 shops are not climate controlled, and moving the parts around is difficult and expensive due to size. Parts will be inspected while on the machine, at temperatures foreseen to be between 40F and 100F.

I'm going to remove my temperature table. Replace with a note for the inspector specifying fit intent (0.025+/-0.024" interference), and instructing to account for thermal expansion of part and measurement tool (20 deg is 0.011" on this pilot). Along with another "If in doubt, ask" note.

 

[racookpe1978]

May I ask about your "(0.025+/-0.024" interference)" dimension?

You want an interference fit of course, but 25 thousands overlaps is larger than what I usually see for our products. Your machined part, your needs are correct for your appilicatino thougfh.

But should you not dimension the overlap with the tolerances on each explicitly stated? (part 1 = 75.000 +

http://wwww/-xxxx,

part 2 = 75.000 +yyyy/-zzzz).

Here, I could see somebody giving you a .001 sliding fit (0.024-0.025) where you really wanted a maximum of 0.024 - 0.025 overlap.

 

[byrdj]

I some times get involved with in place field machining where the mating componenets can be at much differetnt temperture. contact temperture and time are taken along with the dimensionals. then the responsible engineer makes the final call. But the engineer bases his dicesion basicaly off of an temp/expansion table.

 

[TerryR1]

Yes the dims are called out on respective parts (75.007/74.993, 75.032/75.018)

I came up with this by allowing each shop +/-0.007, and accounted for 10deg of error in measurement for temperature, another +/-0.0055.

Parts will fit by heating one to 150F. Worst case cooled stress works out to 10ksi, which is fine for my application.