What is the Resolution of my Load Cell?
What is the lowest Force my Load Cell can detect accurately?
What is the difference between Sensitivity and Resolution?
MTS load cell resolution is a measured value.
*ASTM requirements are generally based on ½ the noise or one digit whichever is greater.
*Resolution shall not exceed ½ the tolerance specification.
*Resolution is required in ACS calibrations
*Resolution is assessed at the lowest non-zero calibration point.
*Data cannot be recorded below 200 x resolution (ASTM E4)
As an example, if I have a 50kN load cell calibrated at 2% of full scale (1kN) and measured 0.008kN noise (Figures 1 & 2), the resolution would be ½X0.008 kN or 0.004 kN.
The minimal acceptable data point based on the measured resolution would be 0.004X200 or 0.800 kN.
Note: Because ASTM E4 assesses resolution at the lowest non-zero calibration point, and MTS uses 2% of full scale force capacity rating of the load cell, the lowest acceptable data point for an MTS performed calibration would be 0.005 kN (lowest calibration point) X 200 or 1 kN.
This would not prohibit the user from recalibrating to a lower calibration point if the noise resolution was less than 0.01% of the full-scale force capacity rating of the load cell. In this example, if MTS had used 0.005 kN as the lowest calibration point but the noise resolution was 0.008 kN, the load cell could be recalibrated to 0.008 kN and the lowest recordable data point could then be 0.004 kN.
Conservatively, the ‘noise’ errors from the transducer and electronics combined should not exceed 0.00040 mV/V. As a percentage of full scale, this is small: 0.0004 / 2 = 0.02%. But, if the transducer is used in an application with a different full scale, the errors can be large. The noise at 10% of full-scale load is 0.2% error. The noise at 2% of full-scale load is 1% error. For this reason, MTS does not recommend the general use of transducers below 2% of their full-scale capacity, because the errors due to noise become excessively large
Generally, the worst-case acceptable data point should not exceed 2% of the full scale. In this case .02*50,000kN=1.000kN. A value higher than 2% of full scale would suggest a problem.
Figure 1 – Screen Shot of Noise at a Static Load of 2% full scale
Figure 2 – Meter to measure the peak to peak noise
Precision is a nebulous term. It is related to accuracy and really involves all the components that go in to the accuracy of the load cell along with its calibration. It is better to compare the calibrated output vs. nominal load standard output to determine error between the two. There are other factors that go into it; filtering, noise, dynamics, temperature change, linearity, hysteresis etc. but there is no one number for precision.
CALIBRATION ACCURACY AND CLASSIFICATION
When load cells are calibrated, they are measured against a standard. The accuracy is how far the measurement is from the standard load cell’s reading.
When MTS does calibrations, the error can be no more than 1% of the reading. This means if you have a 250N load cell the acceptable error could be as much as + or – 2.5 N (.56 lbs) at full scale but at 2% of full scale (5 N) it would be + or - .05 N (.01 lbs).
This is why resolution is measured during calibrations to ensure measurements fall within 1% accuracy even at the lowest calibration point. If the measured resolution was .009 for example, the error at 2% of the full scale would not be more than the 1% of the reading or .05N since any noise would be out too many decimal places to have an effect.
When we do calibrations, we typically do them to 1% accuracy. If noise prevents calibrating to this accuracy due to resolution being worse than 1% accuracy at the lowest calibration point it has to be reclassified or the lowest calibration data point must be raised. Our procedure is to raise the lowest calibration point thus reducing the usable range of the load cell.
MTS standard load cells have a sensitivity of 2 mV/V.
Our software/hardware use a +/10 V min/max range.
To achieve full scale voltage output, the ideal gain is 500, therefore, at full scale…
0.002V\V * 500 *10V = 10V
Note: Some smaller MTS load cells have a sensitivity of 1 mV/V.