Blog

In the pharma industry there has always been great concern on the impact of out of tolerance calibration weights (OOT’s).  “As Found” OOT weight calibration results have always set off alarms and initiated investigations in the pharma industry, not just for calibration weights but for all instruments.  In regards to specifically calibration weights used for internal balance checks, the impact of the OOT “As Found” results on the weight calibration certificate in most cases is surprisingly negligible.

There first thing that needs to be assessed is how far out of tolerance was the calibration weight and how does that value compare to the readability of the balance it was being used on.  In most cases the tolerance for the calibration weights are very tight, Class 1, Class 0, Ultra Class, etc.  So in many cases the OOT condition or amount that the calibration weight is out of tolerance would not be detected or measured on a particular balance.  For example, if we have a 100mg calibration weight that has a Class 1 tolerance of 0.010mg and was found to have an OOT value of -0.014mg from the weight calibration certificate.  Then we know that the calibration weight was -0.004mg past the allowable Class 1 tolerance.  If we were using this calibration weight on a Mettler Toledo XS205DU analytical balance, with a readability of 0.01mg (or 0.00001g) at the 100mg range, then this -0.004mg value can not be measured or displayed.  So in this case, there would be no impact of the OOT calibration weight.  The calibration weight would most likely be downgraded to Class 2 or would need to be replaced to with a new Class 1 calibration weight.

The second issue that would need to be addressed in assessing the impact of OOT calibration weights is what is the internal balance check tolerance being applied.  A 0.1% of the applied calibration weight is a very popular tolerance in the pharma industry.  Lets use the same data from the previous 100mg calibration weight example of the -0.014mg value from the weight calibration certificate but this time we will say the calibration weight is a Class 0 calibration weight with a tolerance of 0.005mg and the balance the weight is being used with is a Mettler Toledo XP6 micro balance with a readability of 0.001mg (or 0.000001g).  The 100mg calibration weight was -0.009mg out of the Class 0 tolerance.  This -0.009mg value can be measured and displayed on the micro balance, so this value now needs to be compared to the applied internal balance check tolerance (0.1% of the calibration weight), which in this case would be 0.100mg.  The 100mg calibration weight’s OOT value of -0.009mg is well within the 0.100mg of the internal daily balance check, so the impact of the OOT calibration weight would be negligible.

The above two examples are that of metrological or scientific in nature, where the data is used to help determine the impact.  It is also important to remember that the accuracy of the balance is not affected by the OOT calibration weights used in the respective internal balance check.  The calibrations weights are used as independent checks and have no effect on the accuracy of the balance.  So, just because calibration weights were found to be OOT from the weight calibration certificate results, this does not mean that the accuracy of any previous calibrations or measurements performed on the balance should be deemed inaccurate or questionable.

The one scenario in which the impact of the OOT calibration weights would be of concern is when the calibration weights have been used to “calibrate” the balance.  Calibrate the balance would be defined as going into the balance’s menu program, as per the operating instructions, and adjusting the accuracy of the balance by following the calibration steps for that balance with the specified calibration weights. This is usually performed by an experienced balance calibration service technician.

There are no published standards are requirements in regards to weight calibration frequencies for your “precision” calibration weights. “Precision” would be defined as any calibration weights having the following, or equivalent, Accuracy Classes; ASTM Class 0, ASTM Class 1, ASTM Class 2, ASTM Class 3, and ASTM Class 4.

Calibration weight manufacturers will recommend one year to six months, depending on usage of the calibration weights and criticality of the weighing process. In servicing the pharmaceutical industry for over 30 years, most of the industry seems to have followed the calibration weight manufacturer recommendations of one year and six months for their weight calibration frequencies. About seventy five percent of the weight calibrations we perform for the pharmaceutical industry are at a one year frequency, with the rest being at six months.

With that being said, unless your weighing process is part of a cGMP critical operation, or getting very heavy usage, or the calibration weights are being used for calibrating laboratory balances, especially analytical and micro balances, then a one year calibration frequency for the calibration weights should be sufficient. If the calibration weights are being used in a cGMP Pharma critical operation, or getting very heavy usage, or are being used for calibrating laboratory balances, then a six-month calibration frequency is highly recommened for the calibration weights.

The other scenario in which the weight calibration frequency would best be at six months is if an Out of Tolerance “as found” condition would be an issue for any previous measurements or daily checks that particular calibration weight may been used for. This is a common issue in the pharma industry that usually triggers investigations and a look at available historical data.

A weight calibration frequency of once every two years would be acceptable if the calibration weights are not getting used on a regular basis or the weighing application is not critical in the least. But no matter how non-critical the weighing process or how little the calibration weights are being used, a once every two year weight calibration frequency would not be recommended for calibration weights that are Class 0, Class 1, or Class 2.

For a variety of reasons, (as found out of tolerance conditions, weight handling, traceability and confidence in your weighing process) anything beyond a 24 month weight calibration frequency for precision calibration weights would not be recommended.

So in summary, a quick and dirty breakdown of calibration weight frequencies would be as follows:

6 Months
cGMP critical weighing operations; the calibration weights are being used for calibration of laboratory balances; avoidance for Out of Tolerance conditions and investigations; the calibration weights are getting very heavy usage.

One Year
The calibration weights are used on a moderate basis; there is a normal level of critacality to the weighing process; the pharma industry has adopted a one year weight calibration frequency as their industry standard.

Two Year
The calibration weights are not being used that often; the calibration weights are not part of critical weighing process.

Our Weight Traceable Certificate is a NIST Traceable report documenting weight calibration results.

The Weight Traceable Certificate documents the testing environmental data (Temperature, Pressure, Relative Humidity and the Air Density) taken at the time of the weight calibration. The air density is used to apply and adjust for the required air buyoancy corrections for the weight calibration results. In addition to the environmental conditions the weight certificate also contains the descriptive and identification info (S/N, ID No., Range, Accuracy Class) for the respective calibration weights under test.

Both the as found and as left values, for all the calibration weights that were calibrated, are documented on the weight certificate. The weight calibration values documented on the certificate are conventional mass values. The conventional mass values could also be called “mass in air”. These weight calibration values are relavent when using the values in measurements with balances. The conventional mass values should not be referenced when using the calibration weights in a vacuum chamber or when “true mass” values are required.

Also documented on the weight certificates are uncertainty values. These values are the uncertainy of measurement values for both the as found and as left weight calibration values. For example, if a 100g Class 1 weight has an as found value of 0.050mg and an uncertainty value of 0.034mg, then the weight certificate is stating that the value of that weight is 0.050mg +/- 0.034mg. The uncertainy values are calculated at a 95% confidence interval using a coverage factor of k=2.

Lastly, the stated tolerance and the as found disposition columns are contained in the weight traceable certificate. The stated allowable tolerance for the calibration weights is based on the fixed accuracy class values, which are referenced from national or international standards. The most recognized domestic calibration weight standard would be ASTM E 617-97 which uses a numerical accuracy system. The OIML R 111-1 calibration weight classification standard would be most recognized internationally. The OIML standard has an alpha-numeric accuracy breakdown.

The as found disposition columun is stricly a “pass/fail” disposition for that calibration weight for the as found value only. If the resulting as found value of the calibration weight is outside of the allowable tolerance for that respective weight class then the as found disposition would state “Out of Tolerance”. If the resulting calibration weight value is within or at the allowable tolerance for the respective weight class then the as found disposition would state “In Tolerance”. Again, this would be for as found calibration weight values only.

It is also of note that a calibration weight will not be certified to a respective accuracy class if the resulting weight calibration value when added to the uncertainty of measurement value brings the calibration weight out of the allowable tolerance range. For example, if the resulting weight calibration value for a 500mg ASTM Class 1 calibration weight is -0.0089mg and the uncertainty of measurement of that weight calibration is 0.0016mg, then the weight calibration value added to the uncertainty of measurement would bring the potential calibration weight value to -0.0105mg which is outside of the allowable tolerance range of 0.010mg. The calibration weight would then have to be removed from service or downgraded to ASTM Class 2.