Tolerance Estimation of a Load Cell I want to Calibrate

casiano · Jul 18, 2012

I have a Load Cell I want to calibrate it. The manufacturer is giving me some specifications:

Non-linearity: 0.25% of R.O.
Hysteresis: 0.25% of R.O.
Nonrepeatability: 0.1% of R.O.
Zero Balance: 1.0% of R.O.

and others!

Which of those should I use? Do I consider all of them? if yes, is necessary to do the arithmetic sum or square root sum?

thanks!

Stijloor · Jul 19, 2012

A Quick Bump!

Can someone help?

Thank you very much!!

JAltmann · Jul 19, 2012

Yes you need to look at/calibrate (or verify) all of those features. They are important and tell you slightly different items about the load cells preformance and accuracy.

Non-linearity: How close to the nominal values is the close along various measurement points.
Hysteresis: How straight is the measurement curve
Nonrepeatability: How well does the gage re-read the same values over and over.
Zero Balance: How well does the gage return to its zero point after being loaded.

casiano · Jul 19, 2012

JAltmann said:
Yes you need to look at/calibrate (or verify) all of those features. They are important and tell you slightly different items about the load cells preformance and accuracy.

Non-linearity: How close to the nominal values is the close along various measurement points.
Hysteresis: How straight is the measurement curve
Nonrepeatability: How well does the gage re-read the same values over and over.
Zero Balance: How well does the gage return to its zero point after being loaded.

I understand that, but my question is do I need to use all of them?

Regards

casiano · Jul 25, 2012

for example!

Non-linearity: 0.25% of R.O.
Hysteresis: 0.25% of R.O.
Nonrepeatability: 0.1% of R.O.
Zero Balance: 1.0% of R.O.

the arithmetic sum is 0.25% + 0.25% + 0.1% + 1.0% = 1.6%

can I use this like total accuracy? or I should look at individual accuracy? or i need to do the square sum, where the total is 1.07%?

thanks!

toolsengineering · Jun 2, 2015

hi there..

i also have problem during calibration..
i still doubt how to solve the problem..

i have calibrated master loadcell 400t and calibrated loadcell 200t..
how to compare reading between master and calibrated loadcell?
how to get the best scale factor?

please help!!!!!!.... thanks

dwperron · Jun 2, 2015

It depends on what you are looking for.

It is possible (but extremely unlikely) that each of these parameters could be at or near its extreme tolerance limit in the same direction, but none of them out of tolerance. In that case the simple additive result produces a "conservative" specification which would allow the load cell to be "within tolerance". This is good for not having to address out-of-tolerance compliance issues, but not so good for producing trustworthy test data because you are allowing the use of a load cell that is marginal at best. In reality it is far more likely that one or more of these terms would be out of tolerance in order to approach the additive total error, even if the result would be "within tolerance" using the additive total.

The best choice is to evaluate each error term individually. That is how the manufacturer specifies them. It's the right way to evaluate the load cell.

wesatwork · Jun 2, 2015

2 Cents from a technician (not a Metrologist)

The weight of the specification is dependent on your conditions. Each specification should be reviewed independently.

Example:
Creep is a common specification. Creep is the change in the signal when all variables (load, environment) are constant except time. If your conditions are static creep is important. If you conditions are dynamic at 10Hz then creep has less relevance.

If your conditions are understood, repeatable, you have a good base line and are competent in parsing the measurement results then you can start to combine uncorrelated specifications. A common method is Root Sum Squared (RSS).

Example:
Hysteresis 0.25%
Non-Linearity 0.25%
Non-Repeatability 0.1%

RSS 0.367%

Just a side note- Zero Balance is the cells signal measurement in a no load/stress condition. Zero Balance provides evidence to the health of the cell. Zero Balance doesn’t have much effect on performance until the measurement becomes too large of a percent of the rated output (RO). Zero Balance is not typically a contributing factor to consider unless it is high, then it may be time to scrap the cell anyway.

dwperron · Jun 2, 2015

wesatwork said:
2 Cents from a technician (not a Metrologist)

The weight of the specification is dependent on your conditions. Each specification should be reviewed independently.

Example:
Creep is a common specification. Creep is the change in the signal when all variables (load, environment) are constant except time. If your conditions are static creep is important. If you conditions are dynamic at 10Hz then creep has less relevance.

If your conditions are understood, repeatable, you have a good base line and are competent in parsing the measurement results then you can start to combine uncorrelated specifications. A common method is Root Sum Squared (RSS).

Example:
Hysteresis 0.25%
Non-Linearity 0.25%
Non-Repeatability 0.1%

RSS 0.367%

Just a side note- Zero Balance is the cells signal measurement in a no load/stress condition. Zero Balance provides evidence to the health of the cell. Zero Balance doesn’t have much effect on performance until the measurement becomes too large of a percent of the rated output (RO). Zero Balance is not typically a contributing factor to consider unless it is high, then it may be time to scrap the cell anyway.

Well put, and you also have to consider what is important to the end user.

Suppose that your application is highly dependent on the linearity of the load cell. If you have a load cell which has been calibrated by combining hysteresis, linearity, and repeatability terms in RSS form it might give you problems. Let's take the example of a load cell with excellent hysteresis and repeatability, both approaching 0%, but non-linearity of 0.35%. That would meet the conditions of the RSS combined tolerance, but not the end user's requirement for 0.25% non-linearity.

If you don't have specific guidance on what is "important" to the end user you have to assume that all parameters are equally important and should be evaluated individually.

toolsengineering · Jun 5, 2015

Thanks to both of you..
But i still have some doubt...
The zero balance i already check and the result same with manufacturer's data sheet.
The question is how to play with the indicator scale to make sure the comparison reading in tolerances 1%?

Tolerance Estimation of a Load Cell I want to Calibrate

casiano

Stijloor

JAltmann

casiano

casiano

toolsengineering

dwperron

wesatwork

Learning what I can.

dwperron

toolsengineering

Similar threads