A recalibration can engender confidence in measuring instruments, writes Loc Duong from Power Parameters.

Certain terms have a bearing on accuracy but they are frequently confused.

We pull them apart to help you in selecting the right measuring instrument for the job.

Limit of reading

This is easily understood when looking a graduated analogue scale of a moving coil meter.

For a digital meter, the limit of reading is the least weight digit. For example: four digits, eg: 9999, provides a limit of reading of 1 in 10,000 (0.01%).

Repeatability

This indicates the variation of measurement results for the same current, voltage, etc, being measured.

Electrical ‘noise’ voltage (equivalent to power) of resistors is proportional to temperature (in degrees Kelvin), resistance in ohms and the frequency band of relevance (ie: the pass band in which we are interested for the measurement of current, voltage, etc).

Linearity

Direct proportionality is provided so that scales become linear.

One way of minimising non-linearity is through the use of range switches, which attenuate the current, voltage, etc.

In auto-ranging instruments, linearity is more difficult to achieve and should be looked at carefully when buying an instrument.

Offset and drift

Drift can produce offset, ie: at 0 input there’s either a negative or positive measurement lifting or lowering the measurement.

It is often related to temperature fluctuation, humidity, or the presence of strong magnetic fields.

Drift can also affect the linearity constant. Think of the instrument response as a simple formula 𝑦=𝑚𝑥+𝑏. The symbol b is the offset, and m the linearity constant. Y represents the instrument reading, and x is the actual value of what we are interested in (voltage, current, etc).

Accuracy

Finally, this term is a combination of the above factors, and specifications should be looked at closely.

There is often some deliberate glossing over of facts. One example is sensitivity, which describes – or should be used for – the smallest value of current, voltage, etc that can be detected.

An example might be a milli-amp scale (0 to 1) with a 10 micro-amp sensitivity.

A global figure for accuracy requires specification of temperature and humidity, and it has to take into account the limit of reading as well as the range factor.

In the final analysis, only a calibration check in our NATA-registered laboratory will establish the facts.

Generally, if repeatability and linearity are satisfactory, accuracy can be established by means of re-calibration.