In a direction finding system that is to operate in real time, such as the SESAM system, computational efficiency is a key issue. However the best superresolution algorithms such as MUSIC and WSF do require a computationally heavy multidimensional search to find the DOA:s, unless the array manifold can be given a Vandermonde structure. In the latter case efficient closed form algorithms such as the Root methods exist. The problem of transforming the manifold of a sparse circular array into Vandermonde structure, and finding closed form 2D algorithms that work over at least two octaves of frequency, are considered in this report. The latter requirement makes it necessary to run the array with element separation much in excess of the Nyquist limit. Unfortunately this makes the only known 2D algorithm for circular arrays, UCA-ESPRIT, fail. A remedy to this problem is given comprising a transformation of the output vector from the array onto the corresponding output vector of an imaginary smaller array that do comply with the Nyquist limit. The transformation hinges on the special property of circular arrays to retain full rank manifolds despite an element separation much in excess of the Nyquist limit. Using this transformation in a 16 element application the bandwidth of UCA-ESPRIT is shown to increase from one to two octaves. This result is supported by simulations. The report is concluded with some recommendations for the SESAM system, including a cost effective scheme that uses 16 antenna elements but only 8 receiver channels.