CMM CALIBRATION CERTIFICATE

[[Me...]]

IF there is failure in calibration certificate for as found, How to read it and justify the failure.

[[De...]]

A lot of this is going to be dependent on your companies quality policy and what it specifies needs to be done in the event of a failure during calibration.

In general you will need to determine how much the machine failed by and then if the machine was used for the acceptance of finished parts then a determination needs to be made as to whether any of the parts that were accepted could have been found as good due to the error in the calibration.

As for how to read the charts, the linear and spatial charts on the calibration cert use a formula for MPE0 which is the allowable error in length measurement based on the machine specification. This will look like A+L/K. The spec will give numbers for the values for A, L and K. By evaluating the formula based on different lengths you will be able to determine the amount of error allowed at any given length. That is why the charts are small on the left side and the allowable error gets larger as the chart moves to the right, this is due to the fact that the machine has more allowable error as measurements get larger.

Based on this you can look where on the charts that your machine failed and then make a determination of how much it failed by.

Unfortunately based on the spec this is about the best you will be able to do, there is no hard and fast way to turn the amount of error in a calibration into an exact number of how this amount of error will effect an actual part. The thing is, in the real world its rare to be taking simple point to point distances parallel to one axis. In most cases parts are checked by collecting points on different faces and then answers are calculated by creating a 3d coordinate system and determining the difference between different features within this coordinate system.

The error in the machine is going to skew this coordinate system to some degree. How much? Good question. The safest answer is probably to just assume the error is at a minimum the amount the machine was out of tolerance, an argument could probably be made to actually increase that amount if the error is in more than one axis.

[[Me...]]

Thanks feed back.
Dilip