Our experience is based upon designing and testing very many circuits, mainly based upon personal computers running Windows controlling software.

Specialisations

• Analogue circuit design 1 Hz to 20 MHz.
• Analogue signal processing.
• Digital circuits to 150 MHz.
• Digital signal processor circuits.
• PC plug-in expansion boards.
• Programmable logic circuits using VHDL.
• Software development to support hardware design.

Cambridge Ultrasonics for hardware design

Cambridge Ultrasonics has many years experience in developing electronic instrumentation circuits. We also have extensive software development experience.

• We use a range of software to assist in the design phase, including FEMLAB, MATLAB, SPICE, VHDL and PCB design software.
• Our experience is based in developing novel instruments particularly for inspection and quality assurance.
• We have worked mainly on stand-alone or PC-based systems.
• Our staff are trained to a high level in this field (graduates and PhDs with several years experience after graduating typically) and offer an intelligent and creative approach to problem solving.
• Our experience result in quick turn-around times for our clients.

Application - detecting ultrasonic waves in air

A client had an existing ultrasonic system for use in air with a miniature amplifier that functioned from time tot time as an oscillator. Not only that the system was not functioning as expected. Our task was to stop the oscillations, reduce sensitivity to environmental noise and perform a technical audit of the entire system - all to be done in about 7 days! We re-designed the amplifier providing a new PCB design on-time, re-using as much of the original design as possible and maintaining compatibility with the rest of the circuit and performed the technical audit to the point of providing full report. Delays in the client supplying us with information meant that the report took 10 days and it took the client a further 4 weeks to get a PCB made and populated, we then had to debug faulty parts and repair the amplifier before providing a demonstration to the client of a working amplifier with a range of 20 m and much improved environmental noise rejection. The client put the amplifier straight into its system with no further oscillations and improved perfromance of the system.

Application - acousto-optic visualisation equipment

The present design, based upon improvements since 1976, uses a regenerative oscillator that can be triggered from an external source. It includes a digital measurement system. Important controls are strobe delay (100 ns to 300 us), flash duration (30 ns to 2 us), transducer excitation (500 KHz to 5 MHz), transducer enable period (100 ns to 10 us), external trigger delay. It uses MOSFET output transducers to drive the ultrasonic transducer from a 500 v intrinsically safe supply.

As well as accepting a trigger the system can provide a trigger for synchronising an oscilloscope/signal capture system monitoring the output signal from an ultrasonic receiver. This creates a powerful mode of investigation in which it is possible to identify ultrasonic pulses causing signals at the receiver. We are now able to make changes to the system, such as: changing the transducer(s) geometry, modifying the transducer(s), adding absorbing materials or changing the signal processing and see the effects on the received signal. It is the ultrasonic equivalent of debugging software.

The system has been used to help develop and investigate a very wide range of ultrasonic systems, including: medical scanners, concrete inspection, systems for inspecting oil-wells, inspecting nuclear fuel rods and inspecting railway lines.

Application - quality inspection using ultrasonic resonance spectroscopy

Cambridge Software was the lead partner in a collaborative European research project in which our technical tasks were to develop a new lock-in amplifier circuit on a PC plug-in board and to write controlling Windows software.

A digital synthesiser was used to give a drive signal with stable control of frequency over the range 1 Hz to 10 MHz with a frequency step of only 0.01 Hz. Input signal conditioning included a multi-stage op-amp amplifier, with differential inputs, high-pass filters, low-pass filters and notch-filters - all selected under software control from the PC. By sampling the received signal in-phase with the driving signal, phase-locked to the synthesiser, we were able to perform the lock-in amplifier function (mixing and low-pass filtering) in software using an on-board DSP. The low-pass (Chebysev II) filtering was simulated first using Mathcad and the filter characteristics were linked to the driving signal frequency, a strategy that made the DSP software much faster and smaller.

As well as our own extensive testing the system has been evaluated by several partners on the collaborative project and was found to work well. It has application to testing the quality of a wide range of industrial components, to find faults and defects. We are now marketing systems in collaboration with a US corporation.

Application - finding defects in concrete

We were approached by the Institut fur Massivbau, Darmstadt, and asked to design and build a novel inspection system to probe concrete. The Institut fur Massivbau was sponsored by the German Institute of Standards and the German Concrete Association.

The hardware was based upon two commercially available PC plug-in boards, one synthesise chirp signals for transmitting, and the other for capturing received signals. A custom control plug-in board and external power amplifier and multiplexer units were designed and built by us to interface to an array of up to 20 ultrasonic transducers (also our design). A daisy-chain method was used to allow one plug-in card to control up to 4 multiplexers; in this way up to 80 ultrasonic transducers could be controlled by one board.

The system built has worked well in many inspection projects and is acknowledged to set new standards for inspecting concrete.

Application - coin recognition systems

We were asked by a market-leading manufacturer of coin-recognition systems to provide help in investigating ultrasonic sensors for recognising coins. Counterfeiting is a serious problem for makers of coin-recognition systems. Traditionally, magnetic sensors are used but alternative methods were sought. Initially, we provided theoretical modelling and the findings pointed towards an ultrasonic impact sensor. We investigated some designs, building prototypes of the most promising. After statistically significant testing we proposed a method of recognition based upon certain impact measurements. A linear processing circuit was preferred and we designed a circuit using SPICE. The circuit was built by us and tested and provided to the client for full evaluation.

Application - measuring the speed of particles in an erosion test using correlation

A binary correlator was developed to perform both cross-correlation and auto-correlation with the purpose of finding the distribution of speeds of sand grains in an erosion test. The motion of the grains was detected optically, using a beam-splitter and cylindrical lens on the collimated beam from a laser to create two light curtains -through which the sand grains had to pass. As the grains passed through first one light curtain then the next they scatter light. Scattered light was collected using fibre optic cables and taken to two photomultipliers.

Signals from the photomultipliers were passed through a thresholding circuit to create binary signals then collected in a digital memory with a sampling speed or clock speed of 10 MHz. Correlation was done by a popular microcomputer.

The digital data was first compressed to record start and stop times of the binary pulses. A novel correlation algorithm was developed to perform correlation on the compressed data, which gave a very useful improvement in computing time.

The system was tested first using data created for the purpose. The first channel of data was created using random numbers for separation times of pulses; the second channel is identical apart from a delay of 500 clock pulses. After cross-correlation a strong pulse emerges at 500 clock cycles as required. The results are averaged over 20 sets of data, which has the effect of improving the signal-to-noise ratio of the peak by approximately 12 dB.

When data from an erosion test is used the peak is not as distinct as with synthesised signals. This is to be expected because the sand particles do not all travel with the same speed. Again the data is averaged over twenty record sets. The dominant peak corresponds to an average speed of 31 (+6-10) m/s.

Auto-correlation can be used on single channels of data to estimate the size of particles.

Application - measuring mass changes in a hostile erosive environment

Changes in mass as small as 20 microgrammes have been measured for solid samples with total masses in the range 0 to 5 g. A magneto-optical system was built to show the mass varying with time during an erosion test.

A sample, to be subjected to an erosion test, was mounted onto a hollow tube fixed at the other end. The tube is free to vibrate transversely. Inside the tube at the sample end was a concave cylindrical mirror; at the fixed end a fibre-optic cable was attached that comrpised three sets of fibres. One set delivered light from a source into the tube, to be reflected from the mirror; the other two sets of fibres collected the reflected light. The focal length of the mirror was chosen to focus the light onto the fibre bundle.

The two receiver bundles were taken to photo-diodes and amplifiers before entering a differencing circuit. The output of the differencing circuit then went into an automatic gain control circuit before driving a power amplifier. The output of the amplifier was sent through two commutating diodes to drive two electromagnetic coils mounted close to the sample end of the vibrating tube and aligned to cause it to vibrate transversely. The system is designed to work in a sing-around self-oscillating manner at a frequency determined by the stiffness of the tube and the mass attached to the end of it.

When first powered-up the tube does not vibrate. The time constant of the AGC circuit is of the order of a few seconds but, after a few seconds, the tube begins to vibrate due to random noise stimulus from the coils. The large Q (200) of the system helps to guarantee that it oscillates in a stable manner but stability appears to be about 2 parts in 1,000,000, when frequency is measured for a period of 1 s - this is attributed to the AGC circuit.

The frequency of oscillation of a tube is not independent of amplitude and this presents a problem. It is best to drive the tube at low amplitudes because then changes of frequency are proportional to changes of mass of the sample. Unfortunately, low amplitude results in noisy electrical signals. The purpose of the AGC circuit is to stabilise the amplitude of vibration at an acceptable level of amplitude. Typical frequencies of operation are in the range 400 Hz to 1 KHz.

In use the system has followed the change in mass of an eroding sample. Clearly visible is an incubation period at the start of the test followed by an approximately constant rate of erosion of material from the sample. It is probable that the rate of erosion is not constant over times of the order of tens of seconds but instead proceeds by a series of steps - this has significance for the mechanisms of erosion. Prior to the development of this system it had been impossible to follow the progress of erosion in-situ because conventional micro-balances using gravity do not work well in erosive environments and the erosive stream imparts a force onto the sample which can be greater than the weight of the sample.

The system developed is virtually insensitive to gravity and has application in conditions of zero gravity, space for example. It is understood that NASA uses a similar principle to measure the weight of astronauts in space and the same principle is used to measure the specific gravity of certain fluids.

Publications:

• "A gravity-insensitive technique for measuring mass changes in hostile environments, including erosion" D.R.Andrews, J. Phys. E:Sci. Instrum., 16(1983) 803-6.
• "A versatile laboratory correlator with application to determine particle velocity in an erosion test" D.R Andrews. J. Phys. E: Sci. Instrum., 16(1983) 706-9.