# Principles of Calibration

### What the curriculum thinks you need to know:

PC_BK_89 Principles of calibration of monitoring equipment A,C,E 1

This gets asked in VIVAs mostly as its something that is nice to describe with the use of diagrams and graphs (I had 5 minutes in my Primary VIVA on calibration!).

This is all simple, and stuff that we do every day. But if you can use the terms and sound confident you can breeze through the content in about 2-3 minutes in a VIVA.

### What you need to know (The theory):

The College likes these definitions…..

#### Accuracy

The ability of a measured value to equal the actual value

So, more accurate devices will give a result closer to the actual value, whilst inaccurate devices will give a result far from the measured value.

#### Precision

The ability of a device to reproduce a value (given the same input)

So a device can be precise even if its measured value is just downright wrong. As long as it gives the same (or similar) value every time.

Principles of accuracy and precision

Calibration

The process of ensuring a measured value equals the actual value.

So if an arterial line transducer is saying 110mmHg when we apply 100mmHg against it, calibration is the act of changing the displayed value to 100mmHg.

To understand how we calibrate, we need to understand the types of error from devices:

#### An Ideal Device

An ideal measurement device would have its measured value equal the actual value across its range (e.g. X = Y)

An ‘Ideal’ Device

There are no ideal devices!

#### Offset Error and One Point Calibration

This is used where a device has a linear and equal response to a step change in the measured value. E.g. if a pressure increases by 20mmHg, the measured value will increase by 20mmHg. (e.g. X +/- n = Y)

The best example of this is an Arterial or CVP transducer. The response to a step change in pressure is linear and equal but the device needs to be ‘told’ where zero is. So we open the device to atmospheric pressure and set this pressure as zero.

A One Point Calibration, Correcting an Offset Error.

#### Slope/Gain Error and Two Point Calibration

This is used where a device has a linear but proportional response to a step change in measured value. E.g. if a pressure increases by 20mmHg, the change in measured value will be variable. But, if the increase is 40mmHg, then the change in measured value will be double that of 20mmHg.

This type of error needs calibrating at two points to get the slope to equal the ideal. A nice example of this is a BM/Blood glucose meter (if you’ve ever calibrated one!). There are two solutions of known glucose concentration, which give a high and low point for calibration. These two points are inside the measured range (e.g. the expected range of blood glucose!).

A Two Point Calibration, Correcting a Slope/Gain Error

Note that slope/gain errors can co-exist with offset errors, but a two point calibration will sort out both. The Diagram above shows only a slope error for clarity.

#### Non Linearity Error and Multi-point Calibration

Some devices have a non linear response from the actual to measured value. Obviously this is going to need a more complex correction to get the measured values in line with the actual values. This is where a multi-point calibration comes in. The number of points that are required depends on the shape of the response curve, the more complex and odd the response, the more points are needed.

A good example of this is a thermistor which shows a non-linear response (see here for more on this)

A Multipoint Calibration, Correcting a Non Linear Response.

### Random Exam factoids (i.e. the things the college like asking):

1. Make sure you know an example of a type of device which needs each type of calibration so you can explain it in the exam.

© Sam Beckett and Physics4FRCA, 2018. Unauthorized use and/or duplication of this material without express and written permission from this site’s author and/or owner is strictly prohibited.