Articles on SharpEye™

New Technology - Transmitters & Pulse Compression

My previous article on NT radar explained the potential benefits of the use of coherent radars, particularly with regard to their potentially enhanced performance in clutter. The possible move towards marine NT radar is enabled by advances in two main areas of technology - semiconductor-based radar transmitters and digital signal processing.

For many years the military have been using coherent radars for a number of applications requiring high performance but these have used extremely expensive transmitter technology. Semiconductor-based radar transmitters have been available for a number of years but these have also been expensive because of the small demand, led mainly by military and space applications. Now, the rapid growth in the mobile communications market, which is forcing transmissions ever upwards in frequency, has fuelled the availability of affordable microwave power devices capable of operating in the 3 GHz radar band. The relative simplicity of using semiconductor devices is that they only need voltage supplies of 20-50 volts, compared to the 10,000 volts that are needed to drive the magnetron in a conventional radar transmitter. In particular, it means that very high reliability designs are possible. The design life of the components is measured in hundreds of thousands of hours; magnetrons typically only achieve 10,000 hours and therefore have to be regularly replaced. Semiconductors can produce the overall (mean) power needed for a marine radar but they cannot easily produce the high peak powers natural to conventional magnetron-based transmitters. This means that NT radar transmitters will normally have to work with very long pulses and employ special techniques to realise the range resolution associated with short pulses.

NT radar transmitters

These transmitters operate as an amplifier, in many ways very similarly to an audio amplifier. A complete radar signal at low level is first generated. This is then amplified by the semiconductor power devices and radiated by the radar antenna. In contrast, a magnetron produces an internally generated burst of microwave energy when a short duration very high voltage is applied to its terminals.

Producing the low-level radar waveform is reasonably straightforward these days, using digitally controlled microcircuit signal generators. These can accurately produce the complex waveforms needed for coherent radar. In the limit, the transmitted waveform of an NT radar can be continuous and not pulsed at all. This is known as a continuous wave (CW) radar. The advantage of such a solution is that the power devices of the transmitter only need to radiate a few watts of power - because the peak power and the mean power are equal - and therefore the transmitter can be small and relatively inexpensive. However, preventing the radar receiver being swamped by the continuously radiating transmitted signal is a difficult (and therefore normally expensive) problem to sort out. There are already coherent radars of this type being used for navigational purposes on naval vessels, although they are expensive. Pulsed systems seem to offer an affordable and more flexible approach. The long pulse of a typical conventional system is about 1 microsecond in length. If the equivalent pulse of a coherent radar was one millisecond the peak power could be 1,000 times less. Therefore the range performance of a 10 kW peak power conventional radar could be effectively replicated by a 100 watt peak power NT radar. In reality, other factors come into play with the benefits of coherency, and so for this example an NT radar having a peak power of 100 watts should exceed the 'in-the-clear' performance of a 25 kW conventional radar. Long pulses can degrade the short range performance of a radar as the pulse may still be transmitting when echoes from short range targets are being received. Therefore there has to be a mechanism to ensure that the IMO short range performance requirements are met.

Pulse compression

The energy reflected from targets when using an NT radar is a relatively long pulse of very low signal level. It is therefore normally necessary to collect and concentrate all the energy from a target reflection to be able to properly detect it and accurately determine its range and its velocity characteristics. The process to concentrate the energy is known as pulse compression. Perhaps the easiest way to understand the general concept is to imagine a received radar signal that consists of low level noise overlaid with a small replica of the transmitted signal wherever there is a target. Conceptually, we can slide a replica of the transmitted signal along the received waveform until a match is obtained; then we can establish the range of the target. This is the essence of the process known as correlation. In practice this is not so easy because noise can 'disfigure' the signal shape that we are looking for. A simplified description of a digital method of performing the correlation process may give a better insight. Time, and therefore range, is divided into small intervals and the voltage level of the received signal is measured at every interval. Each measurement is known as a sample.

Starting with the sample at the lowest range, the shape of the received signal is then compared with that of the transmitted signal. (It is the shape that is important, not the absolute levels). If the transmitted signal was equivalent in length to 1,000 samples this would result in 1,000 comparisons. At every sample point, a score of '1'is given if the received signal matches the transmitted signal; if not a score of '0' is given. The scores are then summed over the time interval occupied by the transmitted pulse. If 50 points are matched a score of 50 (out of 1000) is given.

We then shift the comparison forward by one range increment at a time and continually repeat the process. High scores indicate correlation (there is a match in waveforms) and therefore the presence of a target. Low scores, where there is little correlation, indicate the absence of a target.

The magic about this system is that high correlation occurs exactly at the equivalent range position of the target - the pulse has been compressed. Noise will lessen the total score but the peaks in correlation, indicating target positions, will still be visible. There are a couple of drawbacks. The first of these is that it takes an extremely powerful processor to do this task sufficiently quickly. Fortunately, processors with such a capability are becoming affordable. The second can be more problematic - false correlations can occur. A large target at a certain range can additionally be shown as spurious small targets at false ranges. This effect is called 'range sidelobes' and is somewhat similar to the false targets that can occur at erroneous bearings because of antenna sidelobes. Skilled waveform design minimises the possibility of range sidelobes but under certain conditions false targets may occasionally appear, as they do with antenna sidelobes.

What's better?

In summary, it is seen that NT radars as well as bringing potential benefits in sea and rain clutter performance should also offer extremely reliable systems. Not only do they remove the need for planned maintenance to replace magnetrons but the absence of very high voltages considerably lessens the stress on components. The longer pulses need a high level of processing to extract the radar data by a pulse compression process. This is complex but within the capability of modern digital signal processing technology. The lower peak powers will reduce interference with other shipboard and land based systems. Range sidelobes need to be understood and some user training will be necessary such that they can be recognised. On a well-designed system they should normally not be problematic but may be visible when in the vicinity of extremely large targets.

About Andy Norris

Andy Norris works as a consultant in the international marine navigation sector. He is also Chairman of Technical Committee 80 of the International Electro-technical Commission (IEC), responsible for producing international standards for marine navigation and radio communication equipment in close cooperation with the International Maritime Organization. He holds the visiting position of Special Professor of Navigation Technology at the University of Nottingham, UK.