This paper presents the analysis and design of a chaotic noise generator using statistical mechanic tools. The noise generator is a CMOS analog circuit operating in current mode, which generates chaotic signals using four Bernoulli chaotic maps with topological dependency on each other. They are randomly or deterministically selected to be applied by iterating an initial condition. These variants of the Bernoulli map are different in their slope and attenuation process. The initial condition can be considered as the secret key in the generation of chaotic noise. The advantage of this chaotic noise generator is that its’ control parameter can be selected within an interval greater than the one obtained when basic Bernoulli map is used.