Load cell MSA help

[cpkguy]

Hello all,

I am having trouble convincing my coworkers that we need a known standard to perform a MSA study on a load cell. They want me to pull data from multiple machines and throw it in Minitab. I am trying to tell them all that is doing is giving us the process capability, not the measurement system capability.

Am I wrong here? If not, does anyone know if there is a requirement to use a known standard that I can show them?

Thank you for your time, any help you can provide will be greatly appreciated.

 

[Johnnymo62]

I wouldn't use any uncalibrated device for a MSA.

 

[Bev D]

First calibration is far different from MSA. You cannot and should not perform an MSA on a ‘a known standard’. But as @Johnnymo62 said you do need to perform a MSA using a calibrated load cell. So it needs to be calibrated.
Next simply perform a true MSA on actual results. BUT you must use repeated measurements of the SAME thing. At least 2 measurements or 3 if you are using the straightforward AIAG method.

You cannot never just take some data and throw it into MINITAB and hope MINITAB knows what you want…

You can check out my presentation on effective measurement system analysis in the resources section. It will cover the fundamentals of a replicable study - which is what you have with the load cell. I have performed - or managed the performance of - 10s of load cell studies. It works. Come back if you have questions regarding the study design or analysis of the results.

 

[Ron Rompen]

First, I agree with Bev. Calibration is not MSA.....the two are completely distinct things.

To perform you MSA study correctly, you first need to ensure that your gauge is calibrated and functioning correctly.

Once that is established, I would suggest a Type 1 gauge study on the load cell, using 3 different parts (not reference standards, REAL parts). If you have a known weight for them thats good, if not don't worry about it. Although you can use the 'real' weight as the reference standard in Minitab, it isn't an absolute requirement - use the first 10 (or 20) measurements to determine what the reference weight is, and go from there.

 

[Miner]

Let's take a step back and get the terminology straight. First, MSA is an umbrella term that includes several different types of studies.
MSA includes the following types of studies:

1. Bias (Accuracy) - Difference between the measured mean and the actual mean (typically covered by calibration)
2. Linearity - Bias over the entire measurement range of the instrument (typically covered by a GOOD calibration program)
3. Stability - Bias over time and changes in the environment (may be a special study or covered by an EXCELLENT calibration program)
4. R&R(Precision)
   1. Repeatability - variation seen within a single appraiser measuring the same feature multiple times with the same instrument
   2. Reproducibility - variation seen between multiple appraisers measuring the same feature multiple times with the same instrument

Then we have the different types of studies:

1. Type 1 - Evaluates the accuracy and precision of a measurement device by assessing the bias and repeatability of measurements taken by one operator on a single part with a known reference value
2. Type 2 - Also known as Gage R&R, assesses the repeatability and reproducibility of a measurement system by involving multiple operators measuring multiple parts
3. Type 3 - Similar to the Type 1 study but using multiple parts. Used for automated testing equipment to assess repeatability.

Now lets consider what is a measurement device. Is a stand-alone load cell a measurement device? Or is the measurement device an instrument that contains a load cell? What is the instrument containing the load cell measuring? Is it a tensile or a compressive measurement? Is it a yield strength or an ultimate tensile strength?

You can no more do a single study on a load cell than you can do a single study on a micrometer. The study must include the entire measurement system AND the feature being measured. Tension and compression must be studied separately just like you would study the inner and the outer diameter of calipers separately.

 

[Bev D]

Thanks Miner. I just went and assumed the load cell was in a measurement setup for a push or pull force measurement.

I agree the OP needs to tell us more.