- THE MAGAZINE
- WEB EXCLUSIVES
The long awaited moment has arrived. Your blocks are back from the lab complete with a calibration report outlining the measured size of each piece in the set. Now what?
Before you put the set into service, quality standards require that you review the report to ensure everything is up to whatever standard you are working to, or tolerances you have decided on as being acceptable for your work. The gage record for the set should indicate this review has taken place and they’re good to go.
But what if you’re unsure about some of the numbers or notice a difference between some of the reported sizes on this report and one from the same lab a year ago? What do you do about it—if anything? The first order of business is to get a coffee, relax and think about it. Too often this situation comes up and folks hit the panic button due to these differences. The usual outcome is a call to the lab for explanation, but hopefully I can save you—and the lab—the call.
Assuming the blocks in question have hardly been used in between calibrations, conventional thinking would suggest that the size should be the same as last year or a tad smaller due to wear. But the reported values indicate these particular blocks appear to have grown! Most folks would say that’s good—blocks that never wear out. But not everyone would consider this a plus factor.
There are a number of reasons that can cause this to appear that have nothing to do with the abilities of the calibration laboratory. The first is that the blocks may have in fact grown. I know, we’re all familiar with the claims that the blocks are stabilized to the point where they are rock solid like the Strait of Gibraltar or the Egyptian pyramids. But not all of them are. At one point the U.S. standard for gage blocks listed an acceptable amount of such drift that could be expected. But reputable makers have refined their processes significantly and it rarely occurs from the major block manufacturers. However, today’s gage blocks may come from anywhere in the world and their makers may not be as diligent. Also, sudden shocks from temperature or physical action could undo some of the better processes.
It could be that the values from last year’s report were incorrect or for that matter, they might have been accurate while this year’s numbers are out to lunch. If you’re using a lab that is careful in what it does, this is unlikely. Measurement uncertainty is the more likely factor for differences in reported values. As a rough rule, if the difference between the values is within the stated uncertainty for the calibration lab you have used, the only way to improve on this would be to use a lab with a lower uncertainty. Of course, you could go to the top of the metrology pyramid and have NIST calibrate your blocks. They have uncertainty regarding their measurements like anyone else, but it’s so small that it can be ignored for most applications.
If you did not instruct otherwise, the lab will return the bad blocks with the good ones so it’s up to you to decide which ones will be replaced—if any. Many replacements come with a factory calibration report that can be used for replacements rather than send them out to your lab for calibration. Many customers prefer this arrangement because they remain in control of the situation. It’s time for another decisive moment. This time you have to determine whether those blocks that are being replaced have compromised some measurements to such an extent that a product recall is required.
It just keeps getting messier doesn’t it? If you’re like most readers of my column you have already figured a way to keep from getting into this box. You’ll recall determining, when the set was new, what the maximum deviation you can live with was and then adding a sludge factor to that deviation which became the point that triggers a replacement block. Wear and uncertainty are kept out of your critical zone.
The biggest advantage of the calibration report is that even though the blocks will wear over time, you have actual values for them so their new tolerance doesn’t matter anymore.