Dick Cappels' project pages http://www.projects.cappels.org
Return to HOME (more projects)
A Noise Generator per IEC 268-1, IEC 268-5, and IEC 268-7
This is a noise generator that provides Broadband Audio White Noise, Pink Noise, and Clipped Pink Noise for the testing of loudspeakers and headphones in accordance with IEC 268-5, and IEC 268-7, respectively, as well as many other uses.
rmA related article on this site: True RMS-To-DC Adaptor for DVM
Photo 1. Most of the circuitry is mounted on a prototyping board with a ground plane to reduce susceptibility to unwanted noise. The IEC 268-1 Pink Noise Filter is mounted on the daughter board not only because of board space considerations but also to facilitate replacement if a different characteristic is desired at a later date.
This project came about when I needed to perform some reliability tests on some headphones according to IEC 268-7. The test requires operating the headphones at their rated input power for a number of hours using clipped pink noise, and muting the noise periodically during the test. There are plenty of white noise generators around, but this test called for a specific filter and that the signal be clipped such that the ratio of peak voltage to RMS voltage is between 1.8:1 2.2:1. More conventionally, we would say that the crest factor is set to 2.0 ±10%. Not having a budget for that specialized equipment but having a large stock of components, I opted to make the noise generator myself.
The noise generator shown in Photo 1 requires a regulated power supply of +12 and -12 volts. It provides 438 mv P-P clipped pink noise on one output, and white noise and un-clipped pink noise on two other outputs. All of the outputs are connected directly to the outputs of op amps and are not intended to drive long cables or other capacitive loads.
The specification for the clipped pink noise signal can be found in International Standard IEC 268-7 Sound Equipment, Part 7: Headphones and earphones; Section 220.127.116.11, Method of Measurement. The pink filter and its response are shown in International Standard IEC 268-7 Sound Equipment, Part 1: General; Figures 1 and 2.
Figure 1. The circuit of the noise generator.
The circuit, shown in Figure 1, is composed of a Zener diode white noise source and its bias supply and a preamplifier. This gets the noise up to the 10 to 20 mv RMS level. A second gain stage provides another 31 db of gain, making the noise signal amplitude sufficient to drive the Pink filter described in EN 268-1. The filter is followed by another 31 db fixed gain amplifier than an amplifier with adjustable gain, followed by the clipper circuit. The adjustable gain stage is to allow adjustment of ratio of Peak to RMS noise.
A transistor switch in the clipper circuit allows muting by shunting the noise signal to ground when biased on.
The Zener Diode Noise Source
Figure 2. The TL431 bias regulator, the 1N4735A Zener noise source and LM833 preamp.
Several methods of generating noise are discussed on the internet (such as this article on this very website). After some careful consideration, I decided that a Zener diode based noise source would provide pretty clean and stable "white" noise. What convinced me was a paper written P.I. Somlo in 1975 in which the suitability of a 6.2 volt Zener diode as a transfer noise standard was discussed.
In his paper, Somlo noted that Zener diodes with a nearly zero temperature coefficient of voltage were also the most stable in terms of noise output as a function of temperature. The Somlo paper can be read at users.cosylab.com/~msekoranja/tmp/04236738.pdf Somlo mentioned a distinct "hump" in the function of noise vs current.
The method of biasing the Zener diode was influenced by Linear Technology Design Note 70 in which Jim Williams described a white noise generator.
As shown in Figure 2, the TL431 bias supply sets the voltage across the 25.5k metal film resistor. It was noted by Somlo and others that noise as a function of current has a peak followed by a valley, followed by another peak, indicating two distinct modes of operation. The first peak can be seen in Figure 3 below, in which RMS noise is plotted against voltage from the bias voltage source into the 25.5k resistor.
Figure 3. Vertical axis is RMS noise across the Zener diode in microvolts and the horizontal axis is the voltage at the cathode of the TL431 bias voltage regulator. Notice the distinct "hump" that was found at 6.750 volts and that was described by Somlo. The Zener voltage is approximately 6.07 volts, meaning that the hump occurred at 26.7 microamps of bias through the Zener. The voltage shown on this plot was inferred by measuring after the Pink filter.
In general, the 10k pot in the bias supply circuit is set to achieve the maximum amount of noise out of the second gain stage, which is before the filter. Two stages of gain, amounting to approximately 61 db raise the noise signal level to hundreds of millivolts RMS at the input of the filter. The gain of the first three amplifier stages was limited to approximately 31 db at each stage to preserve bandwidth, leaving the shaping of the noise spectrum to the filter circuit. The comparatively huge 470 uf capacitor on the output of the bias regulator is intended to keep as much of the 1/f noise from the TL431 as practical away from the Zener diode.
The EC 268-1 Pink Noise Filter
Figure 4. The Pink Filter described in IEC 268-1. The tangent of loss for all capacitors must be less than 0.005
The filter shown in Figure 4 described in IEC 268-1 and was built on the phenolic daughter board shown in Photo 1. The filter reduces the input noise voltage by approximately 31 db. This reduction includes the difference in RMS noise voltage resulting from the decreased bandwidth as well as insertion loss.
The Clipper Circuit
Figure 4. Signal clipping and muting circuit. The negative clipper is similar to the positive clipper except that the diodes point in directions opposite those in the positive clipper.
The clipper is composed of three parts; the positive clipper, shown in detail above as the LM833 and its accompanying circuitry, a negative clipper, and a muting switch, a 2SC2878 NPN transistor.
The positive clipper relies on the op amp to keep the anode of the 1N916 connected to the 10k resistor from being pulled any more positive than the positive clipping reference voltage. The diode with its cathode connected to the inverting input of the op amp closes the op amp's feedback loop during the periods in which the signal is below the positive clipping reference voltage. Keeping the loop closed prevents the op amp's output from saturating, which would result in poor high frequency performance. The 47k resistor provides feedback when the noise signal swings toward the positive clipping reference voltage.
The clipper has some overshoot because of the non-zero response time of the op amp.
The noise signal can be muted by grounding the mute input. When the mute input is pulled near ground, approximately 2.5 milliamps flows into the base of the 2SC2878 transistor, which results in a collector-emitter resistance becoming less than two ohms. The voltage divider made of the 10k resistor and the 2 ohm transistor provides an attenuation of up to 74 db. The actual amount of muting depends a lot on layout and routing of signals.
Shown in Figure 1 is the clipping voltage reference supply, which is composed of a TL431C 1% shunt regulator and several 50 ppm/°C 1% resistors.
I built the circuit on a ground plane prototyping board with one pad per hole. The use of the ground plane reduces susceptibility to external noise sources, which is an important consideration since the noise generated by the diode is less than 1 mv RMS. I suggest also using a shielded enclosure to further reduce noise pick-up.
Use of a ground plane can lead to abuse because ground loops exist in ground planes as readily as they exist in wire connections. Therefore, a physical circuit layout that roughly follows the layout of the schematic diagram is recommended to avoid undesirable cross talk among stages. It is a good idea to place the muting transistor as physically close to the output connector as possible in order to obtain the greatest muting attenuation.
Use low drift parts wherever possible, For resistors, the ones marked "1%" on the schematic are rated at 50 ppm/°C or better. The TL431C is specified to be accurate to within 1% and stable to within 50 ppm/°C.
The alignment of the circuit consists of setting the biasing of the noise generator Zener diode and then adjusting the amplitude of the noise signal fed to the clipper to obtain the required ratio of peak noise to RMS noise.
The author uses a triac-controlled soldering iron and he found that the soldering iron added noise to the measurements made during alignment. Assure that no interfering noise sources are present during alignment.
To set the biasing of the noise generator Zener diode:
1. To align the circuit, attach a regulated +12 volt and -12 vol power supply.
2. Use an oscilloscope or RMS voltmeter probe to pin 7 of U2 and adjust the 10k pot to obtain the maximum noise amplitude.
3. Allow the circuit to operate for 20 minutes to assure it is stable.
4. Check and if necessary, readjust the 10k pot to obtain the maximum output on pin 7 of U2.
To set the ratio of peak noise to RMS noise:
5. Adjust the 20k pot to obtain the largest signal amplitude possible from pin 1 of U3.
6. Measure the peak-to-peak signal amplitude on the Clipped Pink Noise output of the circuit, which is pin 7 of U3. Divide this value by 4 to obtain the desired RMS noise amplitude. (See note below)
7. Using an AC coupled true RMS voltmeter, adjust the 20k pot to obtain an RMS voltage on the Clipped Pink Noise output of the circuit that is equal to the desired RMS noise amplitude.
Note about step 6 above: Why divide by 4? IEC 268-5 and IEC 268-7 call for a peak-to-RMS ratio between 1.8 and 2.2. The peak-to-peak noise is measured at the output so that any DC offset on the output is ignored. The peak-to-peak voltage divided by 2, which gives the peak value. Since the desired RMS value is 1/2 the peak value ±10%, the RMS value is found by dividing the peak by 2 again. Dividing peak-to-peak by 4 provides the same value in fewer steps. The 20k pot may be used to obtain other peak-to-RMS ratios if desired.
The generator provides a wide band audio white noise signal that is flat past 20 kHz, a pink noise signal per IEC 268-1, and a clipped pink noise signal per IEC 268-5 and IEC 268-7.
Make sure that the generator is used with a fairly clean +12 and -12 volt power supply. Supplies based on an LM7812 and LM7812 should be sufficient since the generator was designed to reject low frequency power supply variation and drift. The use of switching power supplies is not recommended but they can be used only with the greatest attention to switching-related noise.
Since all three signals on the output are connected directly to the outputs of op amps, care must be taken to not drive long cables with the outputs. The author's intention is to place the noise generator inside an enclosure with the required power amplifier, keeping the capacitive load to a few tens of picofarads.
The mute input mutes the clipped pink noise output only, and does so when held near ground. The input may be controlled by a switch to ground or the output of a logic circuit that is either an open collector (or drain) type output or one with active sinking and sourcing as long as the positive voltage (when not muting) is between +2.5 volts and +12 volts. Note that the open circuit voltage on the mute pin is approximately 2.6 volts.
Find updates at www.projects.cappels.org
HOME (More Projects)
Contents ©2012 Richard Cappels All Rights Reserved. Find updates at www.projects.cappels.org
First posted in August, 2012You can send email to me at projects(at)cappels.org. Replace "(at)" with "@" before mailing.
Use of information presented on this page is for personal, nonprofit educational and noncommercial use only. This material (including object files) is copyrighted by Richard Cappels and may not be republished or used directly for commercial purposes. For commercial license, click here.
Liability Disclaimer and intellectual property notice
(Summary: No warranties, use these pages at your own risk. You may use the information provided here for personal and educational purposes but you may not republish or use this information for any commercial purpose without explicit permission.) I neither express nor imply any warranty for the quality, fitness for any particular purpose or user, or freedom from patents or other restrictions on the rights of use of any software, firmware, hardware, design, service,information, or advice provided, mentioned,or made reference to in these pages. By utilizing or relying on software, firmware, hardware, design, service,information, or advice provided, mentioned, or made reference to in these pages, the user takes responsibility to assume all risk and associated with said activity and hold Richard Cappels harmless in the event of any loss or expense associated with said activity. The contents of this web site, unless otherwise noted, is copyrighted by Richard Cappels. Use of information presented on this site for personal, nonprofit educational and noncommercial use is encouraged, but unless explicitly stated with respect to particular material, the material itself may not be republished or used directly for commercial purposes. For the purposes of this notice, copying binary data resulting from program files, including assembly source code and object (hex) files into semiconductor memories for personal, nonprofit educational or other noncommercial use is not considered republishing. Entities desiring to use any material published in this pages for commercial purposes should contact the respective copyright holder(s).