From: hill@rowland.org (Winfield Hill) Subject: Re: Good Pink Noise Generator Date: 8 Nov 1997 21:27:37 GMT Organization: The Rowland Institute for Science Kevin Nakano, said... > Has anyone come across a good design for a pink noise > generator. I need something in the audio range (20-20k) > +/- 1 dB or so. Something similar to the old MM5437 chip. The 5437 was a white-noise generator. It's true that this chip was often used as the source for a pink-noise filter, but the RC filter networks involved will work with any white-noise source. Here are a few usenet references I've saved. Thomas G Booth, on 6 Jun, suggested Howard Tremaine's _Audio Cyclopedia_ page 363, Figure 7-114B, and a white/pink noise generator kit project, based on a reverse biased BJT B-E junction, was described in an article by J. Petzold in the 3rd quarter 1976 issue of The Audio Amateur magazine. Don Klipstein, on 14 Jun, suggested that noise out of an FM tuner is approximately pink noise from approx. 600 to 7500 Hz, And Roy McCammon, on 5 Oct, suggested learning about 1/f (pink) noise at http://linkage.rockefeller.edu/wli/1fnoise/ Richard Chadderton, posted a .GIF schematic sent to him by Ian Du Rieu, on 11 Jun, saying it was from "Electronics Today International", April 1976, p.23. +V +V +V | | | 10k | 5.6k ,---+---+---+---, | C | | | | | | +--- B ,----+-- 56k --+- 10uF R1 R2 R3 R4 | | E | | | | | | | | | | | OUT 25uF E C 39k C1 C2 C3 C4 +---- | -----+-- 10uF -- | B --- B | | | | | | | | GND C E '-----+---+---+---+ 1Meg | 2.2k Q1 | Q2 | | C | GND '---+-- B GND E | 1 2 3 GND Rx 390k 100k 18k Cx 5.6nF 2.7nF 820pF The original schematic used BC548 transistors, however almost any good small-signal NPN types should work, such as 2N3904, 2N5089 etc. While the V+ supply can be 12 to 30V, it should be well filtered. Transistor Q1 is in reverse E-B breakdown at about 20 uA with its collector open, acting as a weak white-noise source, and amplified by Q2, and an R-C feedback filter network provides the -3dB/octave rolloff. I made a few changes, adding an emitter-follower stage for low output impedance. According to a simple SPICE analysis, the feedback network above has a very rough -3 to -4dB/octave gain response from 10Hz to 10kHz. It has unity gain at 3.4kHz, and a broad +1.5dB hump near this frequency, to make up for a bad dip near 1kHz. The defective mid-range pink-noise performance can be fixed by changing the R-C values and adding another R-C. To choose R-C values, it's useful to note if R and C are scaled by sqrt(10), the R-C product will span 3 decades with 4 R-C stages. According to SPICE, these values work well: 1 2 3 4 Rn 470k 150k 47k 15k Cn 4.7nF 1.5nF 470pF 150pF Finally, one of the best pink-noise circuits is in AoE, see fig 7.61, page 452 for a -3dB/octave pink noise filter circuit which uses 5 resistors and 4 capacitors. This circuit is pink to within 0.4dB from 9Hz to 16kHz (sorry, the book over-states its accuracy). To improve this circuit's accuracy to 0.3dB and extend it to over 100kHz, change the 2.49k to 3.32k, the 2.9nF to 3.0nF and add 1.5nF across the 3.3Meg resistor. Now a -45 degree phase shift is maintained for nearly 4 decades. -- Winfield Hill hill@rowland.org _/_/_/ _/_/_/_/ The Rowland Institute for Science _/ _/ _/_/ _/ Cambridge, MA USA 02142-1297 _/_/_/_/ _/ _/ _/_/_/ _/ _/ _/ _/ _/ http://www.artofelectronics.com/ _/ _/ _/_/ _/_/_/_/