Texture Analyzer – Accuracy, Resolution and Precision Explained

Systems work on the principle of supplying a large range of easily interchangeable load cells, which allow the users of texture analyzers to choose specific capacities relevant to the texture measurement forces they wish to measure.

Texture Analysis accuracy

Texture analysis instruments have the ability to measure load or distance to the absolute true and correct value. Generally defined as a percentage of the full scale range of the measuring device, where the manufacturer states the ability of the instrument to report registered load or distance correctly in relation to a calibrated standard.

Some texture analyzers intelligent load cells operate to 0.1% accuracy of full-scale capacity. If you are using a 10N (1.0kgf) capacity we can expect a margin of error (or inaccuracy), no greater than, +/-1gf. Therefore if the value reported by the instrument for a force is 300gf we can expect our true reading to be between 299gf and 301gf

Resolution is the smallest unit of measure or the smallest change that can be displayed or recorded by an instrument referred to as granularity

This characteristic can be expressed as a percentage or in actual units and is highly influential on the level of accuracy that the instrument can measure at. For Example, A load cell can only operate to 0.1% accuracy of full scale, say +/-1gf therefore the potential error must always be much greater than the 0.15gf resolution that the unit will measure individual changes.

Some texture analyzers have a resolution of 1 part in every 6,500. To calculate this in real terms we divide the capacity of the load cell by 6,500. For Example 1000gf capacity load cell 1,000/6,500 = 0.153gf measurement increments for load, This effectively means that a new load will increment every 0.153gf e.g. For position measurements. A data resolution of 0.0025 mm in its travel span. In other words, there is an event signal from the positional sensor for every 0.0025 mm of movement. Data is recorded at this level of resolution but only displayed at a resolution level of 0.01 mm of travel.

Precision is how reproducible or close identical measurements will be reported as a percentage of full scale.

This is extremely difficult to demonstrate practically as all samples will be slightly different. If calibrated weights are used we can anticipate that our repeatability will be directly a function of accuracy

DATA CAPTURE RATE:

Intelligent filter data from 16,000 down to 2,000 points per second. This enables texture analyzers to operate at a much higher data rate than competitor systems. This means that the load displayed is truer providing the user with a much greater number of “peaks” and “troughs” from trace, although this can be smoothed later within the software.

A high data capture rate is offered so that food technologists are able to collect all available information relating to the textural properties of their samples. This is of particular relevance with brittle foods where the micro fractures caused by the test probe entering or deforming the sample are of importance.

Simple three-point bend test on a cookie.

The brittle structure of the sample exhibits micro-fractures from the minute the test probe touches its surface. These fractures are clearly visible at the 2000pps data capture rate and allows the food technologist to obtain a full and detailed interpretation of the biscuits structure. Had the data been filtered to 500pps these initial fractures would have gone undetected and the food technologist would have potentially made an incorrect assessment of the biscuits fracture characteristics.

AMPLIFICATION OF LOADS:

Systems available to day can collect unfiltered force readings at 16,000 points per second using a 14 bit processor. Not following the principles of amplifying or digitally enhancing the raw data through averaging and calculation of new data points when measuring in the lower region of load cell capacity is a mistake. This is why it is recommend that the use of load cells which match the preferred range of measurement are used. It is often better to have more than one load cell to cover a variety of products.

It is important that you choose the correct load cell. Each load cell has a level of accuracy relating to its full scale measuring capability.

If you measure a very soft product such as cream cheese you will generate very low forces of around 2-3N. The +/- accuracy value means that if I use a 2500N load cell the potential instrumental error will be greater than the force one is wishing to measure.

SUMMARY

1. 10 load cell capacity options from 200gf to 255kgf capacity
2. Choose load cell capacity to match measured range ideally between 10 and 90% of the load cell’s capacity
3. Filter data post test to smooth peaks post test in software
4. Collect 2000 points per second of load, position and speed
5. Accuracy is measured as % of Full Scale. If Accuracy is not stated resolution is irrelevant as measurement will not be true
6. Load cell is calibrated over full range of capacity
7. No amplification of signal to improve accuracy and resolution below 10% of load cell capacity