Browse over 10,000 Electronics Projects

# A filter for a noise generator

BROWSE

Some computations involving analog computers require a random noise source – think, for example, of a system consisting of several masses which are coupled to each other using springs and dampers (like a car suspension system). Such a system may be excited by a random noise source, thus simulating a bumby road.
Other uses involve the study of filter simulation or stochastic processes. There are many applications requiring a good random noise source. Since real time simulations of mechanical systems require quite low frequency random signals as excitation values, it is rather difficult to build a random signal generator for this purpose.

Professional devices will amplify the Nyquist noise of a resistor, mix it with a fixed frequency (normally about 300 kHz) to create an IF down to 0 Hz and amplifiy and filter this. When you have a look at the drawings of a professional random noise like the Wandel und Goltermann RG-1, it becomes clear that building a good random noise source is a major project which will not be completed in a weekend.

Fortunately, I have a Wandel und Goltermann RG-1 noise generator which seemed ideally suited as a random source for my analog computers. The problem to be overcome is the frequency range of the device. Normally it delivers white noise between 0 Hz and 100 kHz (there are some filters like a speech filter, but these filters are not satisfactory for my plans), but I need something like noise between 0Hz and

1 Hz

5 Hz

10 Hz

50 Hz

100 Hz

500 Hz

1000 Hz

5000 Hz

So it was clear that I had to build a high order low pass filter to be switched between the analog computer and the random noise source. I decided to employ a switched capacitor filter to facilitate the overall design – after having a look at modern integrated circuits I decided to use the MAXIM MAX293 – an elliptic low pass filter of eighth order which requires a TTL clock signal of 100 times the cutoff frequency.

Visit Here for more.

BROWSE