Calibration Problem - Range of Use, Calibration Acceptance Tolerance

MASTERQ

Registered
Hi everyone,

Firstly, I would like to thank all the contributors to this forum for your posts which had been a great source of ideas and solutions to the problems that I encounter at work.:thanks:

Recently, we had a customer audit and they had pointed a problem with our internal calibration method.

The following had been stated in our calibration master list (please refer attachment):-

Equipment name: Tanita Digital Scale KD200
Range of use: 2g ~ 600g
Process characteristic variation: 5g
Calibration acceptance tolerance: +/-1.67g

The auditor pointed out that for a weighing scale with smallest scale division of 1g, the stated calibration acceptance tolerance was too large and she asked us why did we set such a tolerance.

I was unable to provide an answer to the auditor but I told her that I would review all the tolerance and see if it is suitable for that equipment but I know nothing about calibration.

The guy who was in-charge of the calibration had left my company a few months before and I have no one to turn to for answers.

I would really appreciate if anyone could clarify what the below terms means and how they relate to each other:-
a) range of use
b) process characteristic variation
c) calibration acceptance tolerance

Thank you.
 

Attachments

  • EQU CAL_ MASTER LIST.pdf
    12.2 KB · Views: 862

BradM

Leader
Admin
Hi everyone,

Firstly, I would like to thank all the contributors to this forum for your posts which had been a great source of ideas and solutions to the problems that I encounter at work.:thanks:

Recently, we had a customer audit and they had pointed a problem with our internal calibration method.

The following had been stated in our calibration master list (please refer attachment):-

Equipment name: Tanita Digital Scale KD200
Range of use: 2g ~ 600g
Process characteristic variation: 5g
Calibration acceptance tolerance: +/-1.67g

The auditor pointed out that for a weighing scale with smallest scale division of 1g, the stated calibration acceptance tolerance was too large and she asked us why did we set such a tolerance.

I was unable to provide an answer to the auditor but I told her that I would review all the tolerance and see if it is suitable for that equipment but I know nothing about calibration.

The guy who was in-charge of the calibration had left my company a few months before and I have no one to turn to for answers.

I would really appreciate if anyone could clarify what the below terms means and how they relate to each other:-
a) range of use
b) process characteristic variation
c) calibration acceptance tolerance

Thank you.

BTW.... thanks for coming by the Cove.:agree1: It would be our pleasure helping you with your questions.

I would say that chart you have is pretty impressive; better information than most places I have seen.


The auditor pointed out that for a weighing scale with smallest scale division of 1g, the stated calibration acceptance tolerance was too large and she asked us why did we set such a tolerance.
I might tend to agree with your auditor of reviewing your tolerance. For one, if you can only read to one gram, how could you possibly estimate 1.67?:tg: You would minimally have to make your tolerance +/-2. Hopefully that makes sense... what I'm suggesting is that if I understand the information given, the balance can never display a number like 1.67; only 1 or 2. Correct me if that is wrong.

If your process tolerance is 5 grams, then I assume that is the tolerance for the balance. What I understand that to say is that as long as the error with the balance is less than 5 grams, it will not affect the process. Is that correct?

If it was mine, I sure would feel more comfortable with a 1 gram tolerance. If the balance cannot meet that, I think you might consider a balance with a little more resolution.

I know you mentioned this is your first go in calibration, so please, feel free to ask any questions at any level. We are here to help.:agree1:
 

MASTERQ

Registered
BTW.... thanks for coming by the Cove.:agree1: It would be our pleasure helping you with your questions.

I would say that chart you have is pretty impressive; better information than most places I have seen.


I might tend to agree with your auditor of reviewing your tolerance. For one, if you can only read to one gram, how could you possibly estimate 1.67?:tg: You would minimally have to make your tolerance +/-2. Hopefully that makes sense... what I'm suggesting is that if I understand the information given, the balance can never display a number like 1.67; only 1 or 2. Correct me if that is wrong.

Thanks for the reply Brad. You are right about the balance as it cannot display a number like 1.67. It can only display numbers like 1, 2, 3 and so on.

I would like to know how to determine the process tolerance. We are using the balance to weigh rubber compound blanks for our compression moulding process.

The required weight for the rubber compound blanks is specific for each part which we are moulding. The required weight is stated on the work instruction of each part and it is stated as a range for e.g. 305g ~ 310g, 30g ~ 32g or 17g ~ 18g.

As you can see, there are 5g, 2g and 1g range. Sometimes, the range is larger than the example given. In this case, what is process variation?

Then, how do we calculate the acceptable calibration tolerance from this process variation? (is there a specific formula etc?)

Currently, our acceptable calibration tolerance is calculated as 1/3 of our process characteristic variation i.e. 5g process variation then tolerance will be 5/3g = 1.67g.
 

BradM

Leader
Admin
Thanks for the reply Brad. You are right about the balance as it cannot display a number like 1.67. It can only display numbers like 1, 2, 3 and so on.

I would like to know how to determine the process tolerance. We are using the balance to weigh rubber compound blanks for our compression moulding process.

The required weight for the rubber compound blanks is specific for each part which we are moulding. The required weight is stated on the work instruction of each part and it is stated as a range for e.g. 305g ~ 310g, 30g ~ 32g or 17g ~ 18g.

As you can see, there are 5g, 2g and 1g range. Sometimes, the range is larger than the example given. In this case, what is process variation?

Then, how do we calculate the acceptable calibration tolerance from this process variation? (is there a specific formula etc?)

Currently, our acceptable calibration tolerance is calculated as 1/3 of our process characteristic variation i.e. 5g process variation then tolerance will be 5/3g = 1.67g.

It's kind of looking to me that your balance is not adequate to meet your need. For the 1 gram range, all of your error is encompassed within your resolution. Make sense? Even for the 5 gram range, your widest process tolerance is 2 grams.

I just had a thought (and probably not a very good one:lol:). Many balances have adjustable resolution. Can you adjust the resolution to .1 gram?

BTW... how is the history on this balance? If the balance has not been right on every time, I think you should look at trading it in for another one. If it has been "on" each time, then see if you can gain some more resolution.

Let me know if I poorly explained anything.:)
 

MASTERQ

Registered
It's kind of looking to me that your balance is not adequate to meet your need. For the 1 gram range, all of your error is encompassed within your resolution. Make sense? Even for the 5 gram range, your widest process tolerance is 2 grams.

I just had a thought (and probably not a very good one:lol:). Many balances have adjustable resolution. Can you adjust the resolution to .1 gram?

BTW... how is the history on this balance? If the balance has not been right on every time, I think you should look at trading it in for another one. If it has been "on" each time, then see if you can gain some more resolution.

Let me know if I poorly explained anything.:)

I can understand your explanation:).Our balances do not have adjustable resolution, the resolution its fixed.

As for the balance's reading, it is accurate and I never had any problems with it.

Buying a balance with smaller resolution it not a big issue but for weighing process is not a very critical for the moulding process. The rubber compound blanks are weighed before it is placed on the mould to avoid overuse of the material (cost saving).

So, is it necessary for smaller resolution?:confused:
 

jerry_Malaysia

Quite Involved in Discussions
I can understand your explanation:).Our balances do not have adjustable resolution, the resolution its fixed.

As for the balance's reading, it is accurate and I never had any problems with it.

Buying a balance with smaller resolution it not a big issue but for weighing process is not a very critical for the moulding process. The rubber compound blanks are weighed before it is placed on the mould to avoid overuse of the material (cost saving).

So, is it necessary for smaller resolution?:confused:

Hi Masterq

Many people always thinking about "what is required by the quality system" or "what is required by the auditors", instead one should be thinking about "what is required by the process / equipment / resources / etc in order to fulfill customer requirements".

Your case, for instance, assuming that the range of specification applied are 305g ~ 310g, 30g ~ 32g or 17g ~ 18g;

by logic, you will need a balance that meets the following requirements

1) Measuring range : at least from 0g to 310g
2) Accuracy : at least 0.1g - applying 1:10 rule.
3) Resolution : at least 0.1g. This is because when the balance displayed a result of 17g, you wouldn't know whether it is 16.5g or it is 17.4g. I believe if the resolution of the equipment is at 1g, it will display a result of 17g even if the measured value if 17.4g

If that is the case, your acceptable tolerance of 1.67g or 2g is not acceptable at all.

Something that i do not understand is how you get the process characteristic variation as 5g?

You mentioned that weighing process is not a very critical for the moulding process, the rubber compound blanks are weighed before it is placed on the mould to avoid overuse of the material (cost saving). My previous superior told me the same thing as well, but i wasn't too agree about it. If it is not critical, then we shouldn't be controlling it. Furthermore, I would say it is critical if it is related to cost saving, because that involves money.
 

MASTERQ

Registered
Hi Masterq

Many people always thinking about "what is required by the quality system" or "what is required by the auditors", instead one should be thinking about "what is required by the process / equipment / resources / etc in order to fulfill customer requirements".

Your case, for instance, assuming that the range of specification applied are 305g ~ 310g, 30g ~ 32g or 17g ~ 18g;

by logic, you will need a balance that meets the following requirements

1) Measuring range : at least from 0g to 310g
2) Accuracy : at least 0.1g - applying 1:10 rule.
3) Resolution : at least 0.1g. This is because when the balance displayed a result of 17g, you wouldn't know whether it is 16.5g or it is 17.4g. I believe if the resolution of the equipment is at 1g, it will display a result of 17g even if the measured value if 17.4g

If that is the case, your acceptable tolerance of 1.67g or 2g is not acceptable at all.

Something that i do not understand is how you get the process characteristic variation as 5g?

You mentioned that weighing process is not a very critical for the moulding process, the rubber compound blanks are weighed before it is placed on the mould to avoid overuse of the material (cost saving). My previous superior told me the same thing as well, but i wasn't too agree about it. If it is not critical, then we shouldn't be controlling it. Furthermore, I would say it is critical if it is related to cost saving, because that involves money.

Thank for your comment Jerry.
I am also puzzled as to how to obtain the process characteristic variation as 5g. This figure was set by my superior who has already left the company a few months back.

May I know how you determine your process characteristic variation for your company?

I would also appreciate if you could also explain to me 1:10 and 1:4 rule with examples.

Thanks.:thanks:
 

Stijloor

Leader
Super Moderator
Thank for your comment Jerry.
I am also puzzled as to how to obtain the process characteristic variation as 5g. This figure was set by my superior who has already left the company a few months back.

May I know how you determine your process characteristic variation for your company?

I would also appreciate if you could also explain to me 1:10 and 1:4 rule with examples.

Thanks.:thanks:

Look at this thread.

Stijloor.
 

Hershal

Metrologist-Auditor
Trusted Information Resource
MasterQ, that seems more like an uncertainty value than a tolerance to me. If I am correct, given the scale parameters you identified, then it is likely very reasonable.
 

MASTERQ

Registered
MasterQ, that seems more like an uncertainty value than a tolerance to me. If I am correct, given the scale parameters you identified, then it is likely very reasonable.

I am confused:confused:
Now if I place a 50g standard weight on the balance shows 49g or 51g, is the reading acceptable.

How do I determine if the reading is acceptable or not?
 
Top Bottom