This paper reports on the development of a low-cost, rapidly deployable sensor for surveying live sewers for blockages and structural failures. The anticipated cost is an order of magnitude lower than current techniques. The technology is based on acoustic normal model decomposition, The instrument emits short coded acoustic signals which are reflected from any sewer wall defect. The acoustic signals can be short Gaussian pulses or longer sinusoidal sweeps and pseudo-random noise. The processing algorithms used on the reflected signal can predict the extent and geometry of the pipe deformation, and the locations and approximate size of common blockages. The effect of the water level on the frequency of the fundamental mode has also been investigated. It is shown that the technique can be adapted to work reliably in relatively large 600 mm diameter sewer pipes.