Article - Issue 19, May/June 2004

From securing stealth to ensuring health – making ultrasound ultra-safe

David Bell

Download the article (99 KB)

Making ultrasound treatment ultra-safe

Ultrasound is used regularly throughout the UK to treat sporting and other soft tissue injuries. However, many of the 10,000 ultrasound physiotherapy units currently in use fail to deliver within 30 per cent of the indicated power and some do not work at all. This was highlighted in a recent New Scientist article based on a survey carried out by Dr Stephen Pye, a medical physicist with the Lothian Region of the National Health Service. Although the survey was carried out ten years ago, recent contacts with many physicists throughout the NHS and the private sector indicate that these results are not only still valid in the UK, but that it is a worldwide problem.

The only way to ensure that patients receive the most effective treatment for their soft tissue injuries is to provide evidence that the equipment delivers the correct level of ultrasound power. Currently, the only way to guarantee the output of the physiotherapy devices is to use costly, cumbersome and complex measurement equipment (radiation force balances) retailing at around £1500–2500. This equipment is only suitable for use by trained medical physicists in the laboratory and, although it offers accurate measurement of power output (between ±7 per cent and ±10 per cent), it takes significant time to set up and use – typically between 15 minutes and 1 hour. Additionally, the ultrasound physiotherapy equipment needs to be removed from the clinic to a laboratory.

How it all started

Precision Acoustics Ltd (PAL) is an SME and one of the leading manufacturers of ultrasound measurement equipment. In order to address the measurement problem, we carried out some initial development work on a new power measurement device and asked the UK national standards laboratory, the National Physical Laboratory (NPL), for comments. At the same time, NPL was investigating a new method of measuring power, which it described as a solid-state meter (with no moving parts). We recognised the synergy in our work and NPL requested help to develop a power meter based on their investigations. At this point we became involved in developing a solution to measure physiotherapy ultrasound equipment.

Our collaboration resulted in the development of a solid-state ultrasonic power meter. This power meter ensures that physiotherapists worldwide can deliver the safest and most effective ultrasound treatment to their patients. NPL owns the patent around the sensor technology, which uses the pyroelectric effect of PolyVinylidene diFluoride (PVdF) – an established sensor material for acoustics. Whereas the R&D work was a joint project for PAL and NPL, PAL is now responsible for the manufacture and marketing of this low cost, user-friendly power meter, with NPL taking a royalty on each unit sold.

How does the power meter work?

In order to produce a robust, cheap and easy-to-use device, an alternative to the conventional method of measuring radiation force was necessary. The team produced a prototype device in under nine months using the known pyroelectric properties of PVdF and a specially designed polyurethane rubber material originally developed for use as a submarine anti-sonar coating. It is considerably cheaper to produce than the conventional radiation force devices and has the added advantage of being portable and extremely simple to use.

The sensor comprises a thin metallised layer of PVdF (less than 0.1 mm in thickness) bonded to the polyurethane rubber. There is an additional insulating layer bonded to the top surface of the PVdF, which provides electrical insulation and protects the PVdF from damage. The transducer is coupled to the sensor through a small water reservoir. For reasons that will become apparent, the sensor is tilted approximately 20° horizontally. The construction of the sensor is shown in Figure 2. The transducer signal passes through the PVdF layer and is then rapidly absorbed into the polyurethane rubber. This produces a temperature rise at the lower surface of the PVdF, which causes a pyroelectric signal across the PVdF. The magnitude of this signal is proportional to the rate of temperature rise. A measurement is taken by turning the transducer on for a few seconds. This produces the characteristic shape illustrated in Figure 3. The peak of the time waveform corresponds to the maximum rate of temperature increase, which occurs about 100 ms after turning on the transducer. About 100 ms after turning off the transducer, the trough corresponds to the maximum rate of temperature decrease.

The maximum rate of temperature change, and hence the pyroelectric voltage, is therefore proportional to the absorbed power. Consequently, the power meter will give an accurate estimate of the total power produced by the physiotherapy unit provided that it is proportional to the absorbed power. Measurements made on the power meter over a range of known powers from a calibrated physiotherapy unit produce a very linear response that validates this assumption.

Early experiments

All sensors created use an absorber (polyurethane rubber) produced by Applied Polymer Technology Ltd (APT) which NPL was involved in developing. The absorber is available in two forms: solid pre-cast sheets of 10 mm thickness; and a 2-part liquid that can be cast to any desired shape.

Preliminary experiments carried out at NPL by Bajram Zeqiri and Adam Shaw with David Bell from PAL established basic sensor characteristics. This included the calibration of a physiotherapy unit used in subsequent tests at PAL. Working with NPL, we developed several generations of sensors using a variety of PVdF thicknesses with both solid and liquid forms of the acoustic absorber. Versions using the solid form of the absorber required the PVdF to be glued in place. This caused unwanted reflections from the watersensor interface, possibly due to trapped air. In later versions we switched to the liquid form of the absorber, which was cast onto the back of the PVdF. This eliminated the glue bond line and reduced the unwanted acoustic reflections; however, due to insufficient mixing of the material, a bubble developed between the PVdF and the absorber. Improvements in our mixing technique and the use of a primer between the PVdF and absorber enabled us to produce a sensor using the liquid form of the absorber.

During the development phase we experimented with different PVdF thicknesses, all of which had gold electrodes deposited in-house directly onto the PVdF surface. However, the production devices use pre-silvered film, which is a more cost-effective solution with no degradation in performance. Since the sensor is a thermal device, there is some expansion and contraction during the measurement process. The front insulating layer is made of 9 ìm unpoled PVdF (i.e. it doesn’t produce a pyroelectric signal) in order to avoid differential expansion between the front insulating and PVdF layers. The thickness of this layer was chosen to minimise surface reflections.

In order to reduce the dependence on transducer orientation, which appeared to be a consequence of ultrasonic standing waves between transducer and sensor, it was necessary to tilt the sensor 20° horizontally. This meant that under normal measurement conditions, in which the user attempts to hold the transducer vertically, the variation in power is less than 2 per cent. This increases to about 5 per cent when the transducer is angled at about 20° vertically. Also, the depth of the water well was increased on the production devices compared to the early prototypes in order to reduce the dependence on transducer-sensor distance.

Problematic electronics

At the beginning of the project we believed that the pyroelectric signal from the sensor would be very low and so very early on we built an amplifier to boost the signal to a useful level. There were various sensor designs during the development phase with widely differing characteristics, which caused some problems for the electronics. As the sensor design was improved, the sensitivity increased and the final amplifier design needed only a small amount of amplifier gain. Despite these difficulties, the finished electronics are brilliantly simple, involving an amplifier, data acquisition module and integrated display – all of which take up very little space.

Although physiotherapy machines work in the frequency range 1–3 MHz, the new sensor is a temperature rate of change device, which is a much slower process at approximately 30 Hz. Since the frequency range of the sensor is close to mains frequency we had a problem with mains noise; but careful orientation of the PVdF film and earthing of the sensor’s water bath has reduced this. Under normal circumstances, this earthing arrangement is sufficient. However, if the meter needs to be used in an electrically noisy environment, an external earth connection can be made via a connector provided on the rear panel of the unit.

Another design feature we used is a programmable data acquisition module. This allows corrections to be made for non-linearity and frequency dependence, permits a meaningful display in watts and provides a route for improvement to the device without having to make hardware changes. An optional serial interface can be included to permit data output to a PC, if required.

Another technological challenge was to ensure that power consumption was low enough for the unit to be truly portable. With measurements taking approximately 5 seconds and an auto power off facility, the use of a 9 V battery gives 45 hours of routine measurement time, keeps the costs down and avoids mains interference.

The box

The box was designed to be flat, wide and reasonably heavy to provide stability. This permits the operator to hold the transducer in the measurement position whilst providing adequate support for their hand. Having support for the hand holding the transducer is important as it reduces vibration. Since PVdF is a piezoelectric as well as a pyroelectric material, it is important that the box doesn’t move during the measurement time.

The sensor is recessed into a water-filled well. Although ultrasonic gel would be preferable, the use of a water reservoir allows heat to be transferred away once the measurement is done and allows thermal equilibrium to be re-established so the measurements can be repeated quickly. Another special feature of the box design is that the sensor has been housed in such a way that it can easily be removed and replaced, which gives an easy method of replacement if the sensor fails. Since this is a new device, it is impossible to predict the longevity of the sensor at this stage. However, more importantly it creates a product that physiotherapy equipment manufacturers can incorporate into their own ultrasound machines, holding out the hope that future physiotherapy systems may be self-calibrating.

A successful outcome

Most of our goals and design challenges have been achieved, with the exception of frequency independence, as ultrasound is more effectively absorbed at higher frequencies. However, by selecting the operating frequency of the physiotherapy machine a correction figure can be applied within the software to ensure accuracy. The sensor has been angled and made deeper, making the device largely independent of position and orientation. Over the physiotherapy power range 0.5–12 W, the device is very linear. After the device has been calibrated, the measured power is within 20 per cent of the true value from an independently calibrated physiotherapy unit.

The power meter is easy to use and can provide a measurement in a matter of seconds. By simply placing the treatment head (the transducer) – the part of the physiotherapy equipment applied to the patient’s body – in the water-filled well in the top of the power meter, the practitioner can check instantly that the equipment is delivering the right amount of power.

The PAL team believes that practitioners and thousands of patients in physiotherapy departments worldwide will benefit from this latest technology and that it will ensure a step forward in the efficacy of delivered ultrasound treatment.

‘Practitioners in physiotherapy departments worldwide will benefit from this latest technology’

The challenge

Ultrasound physiotherapy units currently deliver high frequency sound to body tissue by using a treatment head or transducer that is placed on the skin through a thin layer of ultrasonic gel. This water-based gel allows the ultrasonic signal to penetrate to the required tissue region without being reflected on the skin surface. The challenge was to design a meter that would measure the power delivered by the treatment head. As Pye’s research shows, the meters displaying power or intensity on the physiotherapy equipment are frequently inaccurate. Indeed, what often appears on the meter can bear little resemblance to what is coming out of the transducer. This means that patients can receive the maximum power, or indeed receive no power at all, regardless of what the meter indicates. This highlights that physiotherapists are not complying with their own institutes’ guidelines or IEC standards, which state that they should know the power produced by their equipment to within ±20 per cent of the true value.

The goal was, therefore, to provide a power meter that would meet and exceed these standards and be:

operable by the physiotherapists in the clinic

simple to use

portable (small, light and battery-powered)


quick to give a result so that the power level can be checked immediately before the treatment head is used on a patient

In addition, there were a number of design challenges requiring that the measured signal should be:

proportional to power in the range 0.5–12 W

independent of the operator

independent of ultrasonic frequency

independent of transducer position

independent of transducer orientation


David Bell

Precision Acoustics Ltd

David Bell was born in Epsom, Surrey, England in 1962. He received a BSc (Hons) in Physics with Geophysics from Bath University in 1984 and an MSc in Remote Sensing from London University in 1985. From 1986–1988 he worked as a software engineer at Racal-Decca Advanced Development Ltd. From 1988–2001 he worked as an Ultrasound Medical Physicist at the Royal Marsden Hospital, Surrey with responsibility for quality assurance, clinical support and research. He joined Precision Acoustics Ltd in 2001 as a physicist with particular responsibility for acoustic calibration and measurement. He is a member of the Institute of Physics.

[Top of the page]