13 replies to this topic

[chrissie 2]

I am developing a procedure for scale calibration. Our final product scales weigh product which is on average 100g-200g. Our scales for weighing ingredients 10-25kg. We have calibrated weights to calibrate the scales. What sort of tolerances would you allow for the above, before we would have to send the scales away to be adjusted and recalibrated?

Any help is much appreciated.

Thanks very much.

Chrissie

[PSchnittger]

Dear Chissie,

for your final product scale you will have to stay in line with the legislation in the country you will sell your products, so it might varey from country to country and from product to product. E.g. in Germany are strikt rules therefor.

For the ingridents it depends from my point of view on two aspekts:

• is it something what can rise some risk, there will be either borders by law or you will have to define them
• in any other case it will depend more or less on your recepture or the specification you definde for the product

I hope it helps.

Peter

[Witch]

What sort of tolerances would you allow for the above, before we would have to send the scales away to be adjusted and recalibrated?

Hey, Chrissie
if it is a hazard problem, you can easily decide what to do. Preferably you should think about what happens with your product, when for example 10% too much or too less is measured. If you will have a problem calculate with 9 % and so on. The moment you would not have a problem is your maximum tolerance!
(Forgive my horrible english, did you get, what I wanted to tell? )

Andrea

[chrissie 2]

Thank you so much for your reply - your english is great. This is alot of help.

[yorkshire]

Dear Chrissie,

We use our scale company's recomendation of a 3 unit tolerance, i.e. if the scale reads to 0.1g then we have a tolerance of +/- 0.3g, 10g is +/-30g etc..

This tolerance seems to be workable.

The important thing to note is what action is taken when the scales are out of tolerance.

It is not just a case of recalibrating the scale if it is found to be outside tolerance.

For example if you are weighing out artificial colours it may be critical that you do not put too much in. If you checked the scale and it was reading under then it is possible that you have been adding too much colour. In this case you would need to be aware of any safety factor you have built in and if the scales are outside this safety tolerance then action would have to be taken e.g. recall, hold, rework etc.

A similar issue would apply to average weight checks but this time packs may be illegal if the scales are reading high.

[Dunaskin]

ChrissieSurely if you have your scales regularly calibrated by an independent body, traceable to a national standard, i.e. a NAMAS accredited laboratory, then you have no worries. The calibration certificate should state the measurement accuracy of the scales.

[Charles Chew]

The important thing to note is what action is taken when the scales are out of tolerance.

I am totally in sync with Yorky

I just what to add that it is likely that you may need to look at traceability of non-conformity opportunity beyond the time and day that you have found your scale to be out of range. Is it possible that it was already out of range 1 week ago?

If so, your traceability of non-conformity product opportunity needs to go back.

[okido]

Hi Chrissie

IMO you cannot go blind on the scale reading and use a 3 unit tolerance.
Digital read outs give the suggestion they are very accurate, but are they.
It is better to check the scales spec's in the manual and go from there.
The basic idea is that variation in the measurement process should not contribute too much to the 'true' value of the item that is measured.
What too much is, is the thing you have to think about.

Have a nice day, Okido

[yorkshire]

IMO you cannot go blind on the scale reading and use a 3 unit tolerance.

The important thing to me is make sure that the scale is the correct spec before you start. For example for accurate moisture testing we have a set of scales which reads to 0.0001g, it has draught shields and is mounted on a marble base; this scale would not be suitable for average weight checks in the factory because that accuracy isn't needed, for this we would use scales with 0.1g divisions. To weigh a 50kg bag of sugar we use a scale with 100g divisions, etc.....

[DAVE84]

Does anyone knows what are tolerance limit for USA? I mean when you are calibrating scale for pre operational check, up to what limit we can accept scale tolerance?

[vinegar]

You may want to get a weights and measurments book it will tell you how many units have to be weighed and it will give all the info you need NIST book of weights and measures.

[cosmo]

Rather than waiting until you have an external calibration report one method I use is monthly internal calibrations, recording any variance between the weight and the scale reading. This quickly shows which balances/ scales have drift and those that remain constant between external calibrations. It is not always the expensive balances that are the most accurate.

[Anish]

I totally agree with DUNASKIN. Please do the calibration of your scale with the accredited Calibration body, who can do calibration traceable to NIST standards. When you enter the number, you will find the certificate of the master balance online. Please ensure that your range (minimum & maximum) is included in the calibration range. Also in the mean time, please do have your own internal schdeule of Monthly Master, Weekly & daily calibration.

Does anyone knows what are tolerance limit for USA? I mean when you are calibrating scale for pre operational check, up to what limit we can accept scale tolerance?

[Foodworker]

Some good comments here.

Yorkshire's point about the correct spec is particularly pertinent.

It seems an obvious statement, but it is more common than you would expect to find a company using a scale which is less accurate than the specified parameter being measured. eg the product specification states +/- 0.001g and a scale which can only read to 0.01g being used.

When defining tolerances for calibration, the manufacturers guidance is not always the best method, as it is what is necessary for the parameter being measured which is important. The rule that I have always worked with, is that the tolerance for calibration should be one decimal point tighter than the parameter being measured.

Most companies have an external certified calibration of their scales once or twice a year and use their own standard weights for a more frequent internal calibration check.

Calibrating these internal weights can be expensive, particularly if you have a large number to do. One way round this is to do your own calibrations immediately after you have have had your scales externally calibrated - you will know the scales are OK and what the deviations are, so you can prove the link back to national standards (don't forget to record everything). In most circumstances this should be fine, but there will be restrictions in some cases such as using the weights as counter balance in some of the older scales used in retail situations. There may also be some national variations