Using a test bias from a Rockwell Tester (HRB, HRC) to pass specifications.

[Dan Watson]

I hope I can explain this situation. I am a Quality Control Supervisor. My company is both ISO 9001 and IATF 16949 registered. We do make fasteners for automotive and for distributors that may ship to automotive or aerospace or industrial customers.

In the past, when the receiving inspectors found parts from an outside heat treater the Part Manager would approve to ship, without any formal process. So, since we had this potential system finding, I ordered the receiving testing cancelled and to use the supplier certificate as the approval document.

We had an industrial customer audit my company. He found that the automotive control plans still had incoming hardness testing noted. The QA had not removed.

The QA manager is fine with removing all testing requirements. Other interested parties in management want the testing to resume. They want QC inspectors to take hardness test values and average the numbers so the probability of the shipment passing.

I just read an old NIST standard that spoke to using the bias value of the test disks to apply to your generated values. We have limited resources, either in testing publications or the test disks for the hardness tester. I could use the bias value from the calibration certificate but I am not sure how to apply this value to our test values from the receiving samples.

Any help or direction that anyone can provide is appreciated. I have the same issue with our X-ray plating thickness unit, but one problem at a time.

Thank you.

 

[Dan Watson]

I hope I can explain this situation. I am a Quality Control Supervisor. My company is both ISO 9001 and IATF 16949 registered. We do make fasteners for automotive and for distributors that may ship to automotive or aerospace or industrial customers.

In the past, when the receiving inspectors found parts from an outside heat treater the Part Manager would approve to ship, without any formal process. So, since we had this potential system finding, I ordered the receiving testing cancelled and to use the supplier certificate as the approval document.

We had an industrial customer audit my company. He found that the automotive control plans still had incoming hardness testing noted. The QA had not removed.

The QA manager is fine with removing all testing requirements. Other interested parties in management want the testing to resume. They want QC inspectors to take hardness test values and average the numbers so the probability of the shipment passing.

I just read an old NIST standard that spoke to using the bias value of the test disks to apply to your generated values. We have limited resources, either in testing publications or the test disks for the hardness tester. I could use the bias value from the calibration certificate but I am not sure how to apply this value to our test values from the receiving samples.

Any help or direction that anyone can provide is appreciated. I have the same issue with our X-ray plating thickness unit, but one problem at a time.

Thank you.

Any help or direction that anyone can provide is appreciated. I'm not sure how to handle this without causing a major finding in an audit.

 

[Miner]

I just read an old NIST standard that spoke to using the bias value of the test disks to apply to your generated values. We have limited resources, either in testing publications or the test disks for the hardness tester. I could use the bias value from the calibration certificate but I am not sure how to apply this value to our test values from the receiving samples.

Do you mean that you cannot purchase test disks, or that you cannot take the time to use them?

Compensating for the bias is pretty straightforward. If your bias on the calibration certificate is negative, it means the tester is reading low. In this case, you would add the bias value to the test value. If your bias is positive, the tester is reading high, and you would subtract the bias from the test value. For greater detail, see page 49 of NIST Special Publication 960-5 Rockwell Hardness Measurement of Metallic Materials

It has been 20 years since I was in a business that required hardness testing, but it is good practice to check against a test block daily. Also, if you remove or replace the indenter, the bias usually changes. You would have to check the ASTM E18-22: Rockwell Hardness Standard Test Methods standard to be certain, but it may recommend that you take multiple measurements, discard the extremes and average the remaining values. However, I may be confusing this with the tensile strength standard. Also, the Rockwell test results can be influenced by many factors such as cleanliness, proximity of indents to the edge or each other, indenter wear, and many others. It takes a well-trained technician to obtain accurate and repeatable results.

 

[Dan Watson]

Do you mean that you cannot purchase test disks, or that you cannot take the time to use them?

Compensating for the bias is pretty straightforward. If your bias on the calibration certificate is negative, it means the tester is reading low. In this case, you would add the bias value to the test value. If your bias is positive, the tester is reading high, and you would subtract the bias from the test value. For greater detail, see page 49 of NIST Special Publication 960-5 Rockwell Hardness Measurement of Metallic Materials

It has been 20 years since I was in a business that required hardness testing, but it is good practice to check against a test block daily. Also, if you remove or replace the indenter, the bias usually changes. You would have to check the ASTM E18-22: Rockwell Hardness Standard Test Methods standard to be certain, but it may recommend that you take multiple measurements, discard the extremes and average the remaining values. However, I may be confusing this with the tensile strength standard. Also, the Rockwell test results can be influenced by many factors such as cleanliness, proximity of indents to the edge or each other, indenter wear, and many others. It takes a well-trained technician to obtain accurate and repeatable results.

Thank you Miner. I agree with you about checking the tester everyday with the standard. How Due to management decisions on spending, it is not an option. I know we have an older ASTM standard so I will check this out. Thanks for the guidance.

 

[jmech]

Thank you Miner. I agree with you about checking the tester everyday with the standard. How Due to management decisions on spending, it is not an option. I know we have an older ASTM standard so I will check this out. Thanks for the guidance.

If you are not performing daily verification, then you are not meeting ASTM E18. Per ASTM E18-22:
7.1 A daily verification of the testing machine shall be performed in accordance with A1.5 prior to making hardness tests.

 

[Dan Watson]

Just ot update. I wsa told to only take one sample as a pass/fail. 100,000 sample of automotive parts and we take ONE sample. We don't do daily verification as was suggested here, and the unit only gets calibrated once per year. I tried to compromise by saying we'll take at least five samples per 100,000 and average the five pieces together. Using the measurement error (which is 1 point HRC) I built a "confidence interval" around the upper and lower specification limits. No, take one sample as pass/fail. I guess I fail at trying to run an ethical department.