New to 17025 and need a little help in proficiency testing

[Christopher B. King]

I am new to 17025 and need a little help in proficiency testing. I understand that you need to do interlaboratory comparisons. The confusion comes in reporting of the results since the 2 labs may have different methods of testing. For example we calibrate DC torque wrenches and our method is to test the wrench in 4 positions through the torque range of the tool where as the lab we sent the tool to only test max torque of the wrench. This leads to a different value for degrees of freedom. We also compare the torque reading of the wrench on each sampling to an independent transducer to produce multiple xi, xi-xbar, and Square of xi-xbar readings which are then averaged. The lab we used only compares the torque value of the wrench to the independent transducer until an acceptable Cal Factor is obtained then basis the Uncertainty calculations on the readout of the wrench alone. This causes us not to pass proficiency since they can not achieve the numbers we do. My question is there not a provision for labs having different methods to still obtain acceptable results? We are being required to relax our standards to meet this labs requirements which I find difficult to swallow. This seems to be a flaw since if we adjust to pass with this lab and they go out of business and we have to send the artifact to another lab with an alternate method we will be changing our method again.
Thanks,
Confused

 

[Ken K]

Welcome to the Cove Christopher.

Sounds like you need to get everyone on the same page. I would specify to the lab your are using how the torque wrench should be tested. Send them a copy of your calibration / test method and let them know this is how you prefer to have it tested.
If they can't / won't use your method, seek another lab who will.

IMO, for a valid comparison, you all need to use the same method.

 

[Ryan Wilde]

Well, first of all, from what I've read here, this is not a valid proficiency/interlaboratory test. Who was the pivot lab? What was the nationally recognized lab measurements? Who is correct? There is no control in your test, and therefore, the test gives you no useful data, except that the other lab has some mighty strange practices (the most common failure for a torque wrench is linearity).

First off, get a copy of ISO Guide 43, which is the official method for conducting proficiency tests.

Generally, a proficiency test or interlaboratory comparison will have one lab that acts as the "pivot lab". What this means is that the pivot lab develops the test, specifying the points to be measured, but not the method (Remember, the proficiency test is a means of proving that a method is effective and within stated uncertainties).

The pivot lab will run an initial test on the artifact to get baseline data. The pivot lab then generally sends the artifact to either the National Lab (or a well-known lab with very small uncertainties) for actual values. The pivot lab will then determine their bias from actual. The artifact will be tested before it is sent to a participant in the proficiency test to ensure that the artifact has not shifted. The participating lab then tests the artifact at the exact same test points, reporting what they measure at those points (the actual values as reported by the national lab are not shared with the participating lab) and regardless of measurements, the artifact is never to be adjusted. The artifact and the data are sent back to the pivot lab, which then verifies that the artifact has not shifted or been adjusted or damaged.

From this, the pivot lab and all participating labs know whether or not their uncertainties are somewhat correct, regardless of method. The number of points taken is not a method (although a single point on a torque wrench is very suspect, and I've never met an assessor that would ever accept a single point cal on a torque wrench as a valid procedure), the comparison versus a calibrated transducer is a method, the hanging of weights at a known distance from center of arc is another, etc.

Hope this helps a bit,

Ryan

 

[Ken K]

Does anyone know where I might obtain a copy of IAW Policy 002.
It pertains to proficiency testing.

Thanks.

 

[Torque Guy]

What do you mean by "DC torque wrenches"??

 

[Christopher B. King]

DC Torque Wrenches is a phrase we use for torque management systems used in critical torque applications. These nutrunner, tubenut, ect wrenches are DC electric motor driven units that are controlled by a central controller. All aspects of tightening can be controlled. Speed, shift down speed, torque, angle, ect.. they also report accept and reject indicators based on the parameters supplied.

The following link is one example.
(broken link removed)

 

[Torque Guy]

Why do you calibrate the "wrenches" at 4 points? I believe you will run into problems with other labs because every dc nutrunner manufacturer will tell you the cal value they give you is based on how the tool is used.

For example a Gardner Denver model 48BLC149AM3 nutrunner running a 110 degree joint from a threshold of 5 Nm to a target of 21 Nm will calibrate differently than the same tool running a 280 degree joint or a different threshold or a different speed for that matter.

I believe that a "DC Wrench" should be calibrated at the customers nominal target value on a similar joint at the same threshold, downshift and rundown speed. The very best situation being at the customers facility on their production part. If they change the process the cal is no longer valid.

When you calibrate at 4 points you are demanding that the tool be capable of doing a job it isn't doing at your customers facility. I have found that only a fraction of facilities use a "DC Wrench" throughout it's range. Most set the tool and leave it. Therefore you would only need to calibrate it at one point.

 

[Christopher B. King]

I would definately agree that the best possible calibration is in the field using the customers specs and joints. Most of what we get is the customer sending the tool in to us for repair and calibration. Since we do not know most of the time what he is fastening we must use the best possible method. I also have sever customers that will keep the tool in the crib until it is needed on the line. Depending where on the line it is put in production will determine the setup of the tool.