hello

we have a set of internationally traceble cast iron test masses that we use for validating / calibrating the 45 different scales and balances we have.

we are measuring passenger and cargo weights for aircraft load control, so only consider weights rounded to the nearest kg. typically the smallest scales are 0-120 kg and go through to 0-300 kg electronic scales and beam balances etc. Most units can do down to 0.1 kg though many are analogue with a 0.2 kg minimium division.

what is a suitable tolerance to use to accept or reject the scale / balance? Is plus / minus 3% ok or should this be tighter? presently we are proposing minus 3% to plus 7% as a limit...... though is this unrealistic?

can a 'blanket' tolerence be applied for a calibration method? what variation / tolerence is typically acceptable?

Any help appreciated.

the NIST website (www.nist.gov) has a handbook establishing official tolerances for many types of scales based on range, resolution, application and other factors that I use when ever I have a question as to tolerance.

It is NIST Handbook 44, "SPECIFICATIONS, TOLERANCES, AND OTHER TECHNICAL REQUIREMENTS FOR WEIGHING AND MEASURING DEVICES".

It is a long, detailed handbook. So expect to spend some time finding your application. This, I think, is one of the best resources for such questions.

I notice you are in Australia. I don't know the name off the top of my head of your national measurement organization. But it is possible the Australian version of NIST may have an equivalent handbook (although I am betting there is little difference between them). If you must comply with Australian requirements, I suggest you check the Australian organization first.

As usual, Jerry's the man. Good to see you here, Jerry.

A couple of extra things.

1. What class weights are you using? Depending on the accuracy required, you may want to assure you have sufficient standards.

2. Handbook 44 is long and winding.. but it is the standard for your application and should suite your needs.

The only thing that you merit concern with Handbook 44 is that it allows for limited test points above a certain weight (Table 4). If you use your scale above the limited test points and you are not verifying that accuracy, you could have problems later on.

I know this is really basic stuff... but just to be sure. Use safety with the heavier weights and don't mess your back up. Also, use as much care as possible in placing the weights on the scale to avoid shock.

As far as the acceptable tolerance, that's up to you. I would think the industry would have a requirement (FAA). If Handbook 44 is too tight, you may need to check with your industry requirements and go with that. As far as setting a standard tolerance, that would be nice to do; but with that many different types of scales I would think it would be difficult for all of them to have the same capability.

hello


what is a suitable tolerance to use to accept or reject the scale / balance? Is plus / minus 3% ok or should this be tighter? presently we are proposing minus 3% to plus 7% as a limit...... though is this unrealistic?

can a 'blanket' tolerence be applied for a calibration method? what variation / tolerence is typically acceptable?

Any help appreciated.


The only way that these tolerances would be unrealistic is that they would be far too wide. According to your description, I would say that these scales fall into NIST's "Unmarked" or "Unclassified" scales category. For the capacities that you stated, the tolerances would be based on the test load applied.... More test weight applied, the higher the tolerance.

Example: 300 kg X 0.1 kg scale
Tolerances:
@ 50 kg = 0.1 kg
@ 200 kg = 0.2 kg
@ 300 kg = 0.3 kg

This example is based on NIST Handbook 44, Tables T.1.1 and 6

You can see that the nationally suggested tolerances are FAR smaller than your suggested 3 - 7%.

However, if your process can allow you to have as much as a 7% error on weight and still be ok, then more power to you. You will have a very low scale maintenance bill. You could probably get by with calibrating your scales once every ten years and still meet this tolerance criteria.


You need to determine what your maximum allowable error for reported weight could be. Then take that number and divide it by ten. The result should be the tolerance that you apply to your scales during verification with calibrated and traceable mass standards.

:thanx:

thankyou all for your replies.

a work in progress......