Temperature effects on the measurement

[aliptpatel]

Hello everyone,

So far the companies I have worked for, always have a temperature and humidity controlled quality lab like most of you guys are working. But in my current company, I am working with a ROMER ABSOLUTE ARM 7525 6-AXIS ARM AND WE USE TOUCH TRIGGER PROBES (PHYSICAL CONTACT PROBES). I was so shocked when I see the arm on the shop floor in a dirty environment where people work with drills, sand papers and abrasives to do rework or to make holes and other regular stuff with hand held devices. They all use a vacuum to suck in any dirt particles but still I can see a visible layer or dirt around my workplace in a week after cleaning. At the time of posting this thread, the temperature on the shop floor would be 37C or 38C thanks to our ovens that bake the composite parts. I used a temperature gun pointing on my Romer arm and reading 36.1C and humid work environment. (that's insane). I have never used temperature compensation but as the temperature keeps varying to a lot depending on the duration the ovens stay open and running at 50 Celsius. I was told that the Romer absolute arm can compensate results to 68F, I want to know from you guys that how accurate our results can be? I definitely know this is not ideal condition to even measure the parts.

Just for information, Romer arm holds accuracy of 0.020mm or 0.0008". Also to add in, so far, the building don't seem to have any HVAC system on the shop floor except the CNC room and LAMINATION room. Only Lamination room has "controlled environment" because it "needs". I have tried to explain the need of a controlled Quality Lab not only for CMM but for inspection also (because many parts just go to inspection area to do manual inspections and not come to CMM) but management might not thinking about it for our "SOON TO BE NEW" facility as we are moving to bigger place. So I want to explain what kind of terrible accuracy we can hold at 36C or even more.

Your help is much appreciated guys. Thanks.

[Ncogneto]

https://www.hexagonmi.com/products/portable-measuring-arms/absolute-arm-7-axis

" Built on a foundation of high-tech carbon-fibre tube construction, the Absolute Arm maintains strength and thermal stability under any environmental conditions, offering reliable accuracy whatever the environment."

you should be running the Length and SPAT on the floor before running a part good reads here;

https://www.hexagonmi.com/solutions/technical-resources/metrology-101/temperature-compensation-for-portable-metrology

now why you are using a touch trigger on an arm is the biggest question, unless what you are checking is delicate. I used to work on some stuff called "spam" breathe on it and it would flex on you, even at over an inch thick. we moved to a noncontact scanning arm for that stuff...

[aliptpatel]

Thanks for the information. I usually measure carbon fiber parts and aluminium tools which are solid and won't flex as you mentioned. However management is thinking to get laser scanners for measurement of Styrofoam parts.
On the link you have provided says a skilled operator should consider all thermal properties and stuff to design measurement routine. while we keep the portable arm on the shop floor, temperature always changes in a wide range due to ambient temperature, oven heat and lack of HVAC system in the building. Its obvious that the part measured in summer and in winter will make different results due to temperature change and no one should question repeatability of the ARM like one of our customer does. Either management is ignoring the "temperature change" or simply doesn't understand that it can have effect on measurement.

Mentioning length and SPAT check, you mean to check them on the machine floor before checking the part?

[louisd]

So there's two factors to temperature variation:
1 the changes it makes to the machine and its scales, which romer arms are supposedly very stable and apparently aren't affected by TCOE (Thermal Coefficient Of Expansion).
and
2: the changes that temperature makes to the part you are measuring, which PCDMIS has temperature compensation (available as applies to equipment), in which you need to determine the TCOE expansion value (see https://www.engineeringtoolbox.com/l...ents-d_95.html or ask the Google for more info). You need turn on temperature compensation within each routine's code, and define the material type/TCOE value, in order for PCDMIS to compensate for part TCOE deltas. Once you implement PCDMIS temp compensation, you should mitigate almost all of your variation you are experiencing.

To explain thermal coefficient of expansion, learn more about it. it's pure math.

[aliptpatel]

Thanks Louisd,

I have tried that and notice a list of material that I can select. I work with resin and carbon fiber which are not listed on the drop down list. How can I use temp comp for them?

[louisd]

in edit window, type in tempcomp, press tab (or go through pop up window, selecting any material you want temporarially). then change CTE value from whatever the material you chose or it defaulted to, to the CTE value of your material, the material type will automatically change to "Custom Material", and retain the CTE value you input. Keep in mind it's in metric units, micron * meter^-1 * °C^-1, regardless of what unit format you select for the routine.

[oddam]

I will put my two cents in this one. I agree with you that a more stable environment is desirable, but not always needed. As long as the parts and arm have the time to adjust to the temp, you should be ok. Now, if you have very tight tolerances, say less than .025mm, this might be an issue. I suggest you take a few different parts, (take into consideration material, thickness and other items), measure them and with your measured readings, record both inside and outside weather conditions, (temp, humidity, location checked in the plant). Keep the parts and measure them over a period of time, maybe even a year for different season weather conditions. Review your findings and see if there is a concern, if you think there is, you can now show data that backs your concern. If there is no real change in readings, then the working conditions your in just suck. It would be interesting to see your results this time next year. Has anyone else ever done a study like this when conditions change? Especially with an arm.

Odda

[Mike Ruff]

It's a bit less straight forward when you're working with composites, because depending on how it's laid, they can end up expanding one way more than another. The same goes for other mechanical properties of composites as well (tensile strength, etc.). Keep in mind that thermal expansion is a percentage growth based on length AND temperature change. So the further your temperature is from 68°, the more the part will grow (or shrink), and the longer your dimension, the more you will see it grow (or shrink). Check matweb.com for the coefficient of thermal expansion, if it's not there, you might have to measure a part in your shop, then remeasure the same part with the same method with the same tools in a climate controlled environment, and calculate it out yourself.

[aliptpatel]

Wow, its an interesting study. Most of our measurement tolerances are .005" to .010" I would certainly try to find out a simple and solid part, do some measurement on it and keep doing that thru the year to find out differences.I agree with Mike it grows or shrinks depending on the oven cycles and temperature at CMM. We have measured one part on our shop floor and sent the same part for third party measurement where they have used bridge cmm and the differences were big for some of the features. I did measure part with portable arm 3 time and results were in .002" range for each feature. However, the only issue with portable arm is you can't hit the same point everytime even if you align the part in same way. and most composite parts are not a straight blocks but various curves. So even in same temperature the hits will be on different point everytime resulting difference between each measurement taken it will be smaller or bigger depending on layup irregularities.

[fatso666]

We used to have a CMM that was not temperature compensated, and our shop environment is not environmentally controlled. We created an aluminum tool (aluminum has the largest CTE of the materials that we use, AND it is the most commonly used material in our industry) that is just a couple of tooling balls in a long, straight, thick piece of aluminum. We send it out once a year to be measured at an environmentally controlled facility, so we know the dimension between the TBs at 68-degrees. Whenever it is too warm (or cold), we measure the tool and compare to the 68-degree nominal. This has given us a set of 'factors' to correct for temperature (in aluminum). Perhaps you could create a couple of tools out of CF - one with the 'grain' going length-wise, and another with the 'grain' going width-wise. Like someone upthread said, this isn't all about the material being measured -- the expansion and contraction of the various materials in the CMM will also affect your measurements. This real-world measurement should give you real expansion/contraction factors with the material of the tool AND the various materials in your measuring machine. We now have a new CMM with temp comp sensors built into it, so our CTE aluminum tool just collects a lot of dust (LITERALLY, because our CMM facility is out on the shop floor, subject to all the dust/dirt/crud coming from all the machinery) now.

[aliptpatel]

I agree with you fatso666. The longest part we make are order based so I can't keep them with me for calculating worst case scenario but I will talk to some engg or our CTO about getting a part thru the summer to calculate worst case scenario. Our CTO knows about requirement of controlled environment of CMM lab but sadly, he don't have that power to implement that. I will talk to them and will find out. My manager agreed to do cmm calibration at hottest day adn coolest day of the year but as romer arm is made for shop based operation it wouldn't represent the actual issue of temperature changes of the part itself.

Also I want to know that the temperature sensors that are available externally, do they compensate for real-time temperature or just at the beginning of the measurement? If it is real-time, then if I measure one feature at 80F and measure another one at 84F (assuming I go for a meeting or break for an hour) would that compensate both of them differently?

[fatso666]

Also I want to know that the temperature sensors that are available externally, do they compensate for real-time temperature or just at the beginning of the measurement? If it is real-time, then if I measure one feature at 80F and measure another one at 84F (assuming I go for a meeting or break for an hour) would that compensate both of them differently?

I do not know what is available externally. The sensors we have are integral to the CMM --- we bought the CMM with the sensors installed at the factory. Also, I have often wondered about the timing of the Temp Comp. I suppose it could make a difference in a long running program where it is cool in the morning but heats up as the day wears on. I just don't know if the compensation is based on the temp at a particular time (probably at the start of the program), or if it is based on some real-time continuous reading.

[aliptpatel]

I figured out that we use so many different kind of resin mix, fiber glass and number of different types of carbon fiber fabrics. some of the resins are invented by our company that are used at only our facility. Considering that mix would not be same everytime and each part will be mixture of many different things, temperature compensation for part would not be a real thing and will never be accurate. So for that, lab is the only option unless they allow me to work in engineering office, give me a small space as temperature over there is maintained to be 22-23C. Its not ideal but its far better than shop floor.

[Quality ish]

Not to be "that guy", but you may find a better audience (with more applicable experience) in the PC-Dmis Portable board.

This board is home to folks that are very knowledgeable about CMMs, but Romers are a slightly different breed of CMM.

[aliptpatel]

Thanks for your info. I will also post this thread in Portable board.