Fullscreen HTML5Med Res (??MB)HTML4

Build This Sound Level This Sound Level Meter adaptor will measure sound pres­sure levels from below 20dB up to 120dB with high accuracy. It connects to any standard digital multimeter and has inbuilt filters for A and C-weighting. Noise can have a huge affect on the quality of our lives. A reliable measuring instrument is a must for those interested in finding out just how much noise is in their environment. Just how much noise is present at any time is very subjec­tive. If you are confined to a soundproof room for a period of time, even the sound of a pin dropping will seem quite loud. But if you are in a normal home or office environment, the dropping of a pin is likely to be completely inaudible. And even the sounds of people on the telephone or using computers may be completely drowned out if a semi-trailer passes down your street or a jet flies overhead. The above examples show just how exceptional our ears are in responding to the possible range of sounds in our environment. In fact, we could expect to experience a sound pressure range of about three million to one. Because of this huge range of values sound pressure levels are usually expressed in decibels, a loga­rithmic ratio where 20dB (decibels) is equivalent to 10:1; 40dB is 100:1 and 60dB is 1000:1, all compared to a reference level. The overall 3,000,000 to 1 range can then be expressed as 130dB (20 log 3,000,000). Since the dB is a ratio it must be referenced to • • • • 82 Silicon Chip’s Electronics TestBench Main Features Connects to any digital multimeter Calibration method uses loudspeaker & pink noise source A and C weighting plus flat (unweighted) filters Slow, Fast and Peak response By JOHN CLARKE Meter a particu­lar pressure level of 20.4µPa (micro Pascals). Usually sound pressure levels are quoted as so many dBSPL, indicating that the 0dB reference is 20.4µPa. On the dBSPL scale, 0dB is virtually inaudible, 30dB might be the sound level in a quiet rural area with no wind while a noisy home kitchen might be 80dB or more. Heavy traffic can easily be 80-90dB while a suburban train in a tunnel can produce 100dB. Electric power tools or pneumatic drills can easily run at 110dB and some can go into the pain level at 120dB. Measuring SPL The S ILICON C HIP Sound Level Meter is designed to produce accurate readings of sound pressure which are displayed on a digital multimeter. It Fig.1: this graph shows the differences between A and C-weighting and flat (unweighted) responses in the Sound Level Meter. comprises a handheld case with a short tube supporting the microphone at one end of the unit. Flying leads with banana plugs connect to the multi­meter. A slide switch provides A-weighting and C-weighting filt­ers to tailor the measurement readings. A-weighting is called for in many measurements to Australian standards although it is not really appropriate for louder sounds where C-weighting or a flat response (unweighted) can give more meaningful results. Fig.1 shows the differences between A and C-weighting and flat (unweighted) responses in the Sound Level Meter. Slow and fast response times are provided as well, so that sudden noise can be filtered out, if need be. A “peak detect” facility has also been included which will give an indication Fig.2: the block diagram of the Sound Level Meter. IC4b controls the gain of IC2 so that the output from the full-wave rectifier is constant. IC4b’s output is atten­uated by IC3b and fed to an external multimeter. Silicon Chip’s Electronics TestBench  83 Fig.3: apart from the use of a VCA (IC2), an unusual feature of the circuit is the use of IC5 to evenly split the 18V supply. This has been done because the negative rail is subjected to a higher current drain than the negative rail, which would shorten the life of battery B2. of the noise waveform shape. If there is no or little difference between the peak and the fast reading then the noise waveform can be assumed to be relatively sinusoidal. If, however, the peak level is greater than the fast reading, then the noise waveform has a lot of transient bursts. These may result in a low average value as shown on the slow 84 or fast re­sponse settings but are easily captured by the peak detect cir­cuitry. The cost of the Sound Level Meter has been kept low by using a multi­ meter as the display. Logarithmic conversion As already noted, the Sound Level Meter will read from below 20dBSPL Silicon Chip’s Electronics TestBench to 120dBSPL, a range of 100dB. That’s a pretty stiff requirement. The circuit has to provide a direct logarith­ mic conversion over 100dB, producing an output of 10mV per dB. In practice, the signal fed to the multimeter ranges from 200mV at 20dB to 1.2V at 120dB. This means that all readings can be made on the 2V range of the multimeter; there is no need to switch ranges. Fig.2 shows the block diagram of our sound level meter. Signal from the microphone is amplified by op amp IC1a and then fed to either the A or C-weighting filters which involve switch S2 and op amp IC1b. IC2 is a voltage-controlled amplifier (VCA) which can either amplify or attenuate the signal from IC1b, depending on the voltage at its control input. This input operates in a loga­rithmic fashion so that small control voltage changes can produce large variations in the output signal. IC2’s output is full wave rectified by IC3a & IC4a and the rectified signal fed to the Slow, Fast or Peak filters involving switch S3. The resulting DC voltage is compared in error amplifier IC4b against a 20mV reference. IC4b’s output then controls the VCA so that it produces a constant output regardless of changes in the microphone signal. As well as driving the control input of the VCA, IC4b drives op amp IC3b which modifies the signal so that it provides the required 10mV per dB, to drive the external multimeter. Circuit description Fig.3 shows the complete circuit for the Sound Level Meter. It uses five ICs, three of which are dual op amps (IC1, IC3 & IC4). IC2 is the VCA, which can be considered as an op amp with a DC gain control. IC5, a TL071 single op amp, is used to accurate­ly split the 18V battery supply; more of that later. The microphone is an electret type which is biased via a 10kΩ resistor from the +9V supply. Its signal is coupled to op amp IC1a which has a gain of 7.9 (+18dB), as set by the 68kΩ and 10kΩ feedback resistors. This gain has been selected for the specified microphone and will need to be altered if other types are used. IC1a drives both the C and A-weighting filters. These are selected at positions 1 and 2 of switch S2a respectively. Posi­tion 3 selects IC1a’s output directly for the flat or unweighted signal mode. IC1b is simply a unity gain amplifier to buffer the filters and prevent loading of the filter signal. IC1b’s output is fed to IC2 via switch S2b and a 10µF coupling capacitor. Note that in positions 1 and 3 of S2b, the 4.7kΩ and 12kΩ resistors are connected in series while for position 2, the 4.7kΩ resistor is bypassed. This allows a 3dB higher gain for IC2 when A-weighting is selected. The gain adjustment is necessary to maintain the Fig.4: waveforms from the precision full-wave rectifier. The top trace (Ch1) shows the input sinewave while the lower trace (Ch 2) is the rectified version. Note that the RMS values are slight­ly different due to small offsets in the op amps. same 1kHz signal level applied to IC2 for all posi­tions of switch S2. IC2 is an Analog Devices voltage-controlled amplifier (VCA). It has a dynamic range of 117dB, .006% distortion at 1kHz and unity gain, and a gain control range of 140dB. The DC control input operates at -30mV per dB gain change. IC2’s gain is set by the voltage at pin 11 and the ratio of resistance between pins 3 and 14 and the input at pins 4 & 6. The 100kΩ resistor between pin 12 and the +9V rail sets the bias level for the output at pin 14. This bias can be selected for class A or B operation. Class A gives lower distortion but slightly higher noise. We opted for class B bias for best noise performance. A .001µF capacitor between pins 5 & 8 compensates the gain control circuitry. Precision rectifier IC2 is AC-coupled to the precision full wave rectifier formed by op amps IC3a & IC4a. For positive signals the output of IC3a goes low to reverse bias diode D1. Positive-going signals are then summed in inverter IC4a via the 20kΩ resistor R1 to produce a negative output at pin 7. The gain is -1. Diode D2 and the 20kΩ series resistor limit the op amp’s negative excursion. For negative signals D1 conducts and IC3a acts as an in­verting amplifier with a gain of -1 to sum into IC4a via R5. Negative-going signals are also summed in IC4a via R1. Since the voltages across R1 and R5 are equal but opposite and the value of R5 is exactly half R1, the net result of the sum into IC4a is a negative output with an overall gain of 1. So for positive signals applied to the full wave rectifier the gain is -1 and for negative signals the gain is 1. Thus IC3a and IC4a form a precision full wave rectifier. The 10kΩ and 5.6kΩ resistors at IC3a’s and IC4a’s non-inverting inputs minimise any offset voltages in the op amps. Fig.4 shows the oscilloscope waveform of the precision full wave rectifier. The top trace shows the input sinewave while the lower trace is the rectified version. Note that the RMS values are slightly different due to small offsets in the op amps. The switched feedback across IC4a provides filtering of the rectified signal as well as gain control. In the ‘slow’ setting of S3a, the 20kΩ resistor sets the gain and the 470µF capacitor controls the response. Similarly, for the ‘fast’ setting of S3a, the 100µF capacitor sets the response. In the ‘peak’ position of S3, diode D3 charges the 10µF capacitor to the peak value of the waveform while the 12kΩ resistor sets Silicon Chip’s Electronics TestBench  85 Fig.5: follow this diagram when installing the parts on the PC board and take care to ensure that all polarised parts are correctly oriented. Note that REF1 and a number of capacitors must be laid flat on the PC board (see text). the gain. This is lower than the 20kΩ value used in the other S3 positions so that the output at the wiper of S3b is the same as for the slow and fast settings when a sinewave is applied. VR1 allows precise adjustment of this calibration, providing a divide by 4.6 to 1.8 range. VR2 is the offset adjustment. Error amplifier If, after reading the circuit description so far, you are unclear about its operation, do not despair. Let’s summarise what really happens. Op amp IC4b, the error amplifier, is really the The filter signal at the wiper of S3b is monitored with error amplifier IC4b. This has a gain of -100 (ie, it is an inverting amplifier) and compares the rectified signal from switch S3b against the -20mV reference at the non-inverting input, pin 3. IC4b’s output drives pin 11 of IC2. The -20mV reference is derived from the 2.49V reference REF1 via 560kΩ and 4.7kΩ resistors. REF1 is an LM336-2.5 preci­ sion reference diode which has facility for a small amount of adjustment although it is not used here. REF1 is also used to provide a calibration offset for op amp IC3b. IC3b attenuates the logarithmic DC control voltage for IC2 to convert its nominal 30mV/dB calibration to 10mV/dB. 86 The big picture CAPACITOR CODES ❏ ❏ ❏ ❏ ❏ ❏ ❏ ❏ ❏ ❏ ❏ Silicon Chip’s Electronics TestBench Value 0.56µF 0.22µF 0.18µF 0.15µF .047µF .0027µF .001µF 100pF 33pF 12pF IEC EIA 560n 564 220n 224 180n 184 150n 154 47n 473 2n7 272 1n 102 100p 101 33p   33 12p   12 heart of the circuit. It continually adjusts the control voltage fed to IC2 so that the negative DC voltage fed from the wiper of S3b to its pin 2 is always very close to the -20mV at its pin 3. In fact, VCA IC2 does not really operate as an amplifier for most of the time. For example, when a signal of 120dBSPL is fed to the microphone, the output of IC1a and IC1b is close to clipping; ie, around 14V peak-to-peak or 5V RMS. This is heavily attenuated by IC2 so that around 30mV RMS (see Fig.4) is applied to the input of the precision rectifier, IC3a. Actually, it is only for signals of around 20mV or less from IC1b that the circuit involving IC2 has any gain; the rest of the time it is attenuating and the actual degree of attenua­tion depends on the size of the signal coming from IC1a. Typical­ly, the control voltage delivered by IC4b ranges from about +3V, corresponding to maximum attenuation in this circuit, to about -1V, corresponding to maximum gain. Hence, IC4b makes sure that its two inputs are very simi­lar, and in doing so, it produces a control voltage which happens to be 30mV/dB. This is then further attenuated by IC3b to produce an output of 10mV/dB which can be read out as a measure of the sound pressure level. Looked at this way, the output voltage read by the external multimeter is almost just a byproduct of the overall circuit operation. The assembled PC board is secured to the base of the case using four small self-tapping screws. Battery supply Two 9V batteries in series provide an 18V supply. The 18V is divided using two series connected 10kΩ resistors, to produce a 0V reference and this is buffered by op amp IC5. IC5’s output feeds a 100Ω resistor and two 100µF capacitors. These decouple the op amp’s output and ensure that it has a very low output impedance at all frequencies of interest. The result is a dual-tracking supply which is nominally ±9V. Now why go to all that trouble when we could have used the midpoint of the two 9V batteries to do the same thing? The reason is that there is more current drain from the negative rail in this circuit and so the negative 9V battery would normally be discharged faster than the positive 9V battery. This would be a problem because the circuit require more negative output swing. By using the op amp split supply RESISTOR COLOUR CODES ❏ No. ❏  1 ❏  1 ❏  1 ❏  3 ❏  1 ❏  1 ❏  1 ❏  1 ❏  6 ❏  1 ❏  2 ❏  9 ❏  1 ❏  1 ❏  1 ❏  2 ❏  2 ❏  2 ❏  1 Value 2.2MΩ 560kΩ 180kΩ 100kΩ 68kΩ 33kΩ 22kΩ 24kΩ 20kΩ 18kΩ 12kΩ 10kΩ 8.2kΩ 6.8kΩ 5.6kΩ 4.7kΩ 3.9kΩ 150Ω 100Ω 4-Band Code (1%) red red green brown green blue yellow brown brown grey yellow brown brown black yellow brown blue grey orange brown orange orange orange brown red red orange brown red yellow orange brown red black orange brown brown grey orange brown brown red orange brown brown black orange brown grey red red brown blue grey red brown green blue red brown yellow violet red brown orange white red brown brown green brown brown brown black brown brown 5-Band Code (1%) red red black yellow brown green blue black orange brown brown grey black orange brown brown black black orange brown blue grey black red brown orange orange black red brown red red black red brown red yellow black red brown red black black red brown brown grey black red brown brown red black red brown brown black black red brown grey red black brown brown blue grey black brown brown green blue black brown brown yellow violet black brown brown orange white black brown brown brown green black black brown brown black black black brown Silicon Chip’s Electronics TestBench  87 REF1 is mounted on its side as shown in Fig.5, to allow room for the battery to lie on top of the PC board. For the same reason, the .001µF capacitor near IC2, the 0.18µF capacitor near VR2 and the 100pF capacitor near VR1 should be inserted so that they lie flat on the board. The electrolytic capacitors must be oriented as shown. Insert and solder LED1 at the end of its leads to allow it to protrude through the front panel when assembled. Insert trimpots VR1 and VR2 and cut the ‘A’ PC stakes slightly higher than the trimpot height. This will prevent the batteries pressing on the trimpots and altering the set values. This battery holder was made by soldering several pieces of double-sided PC board Now fit the assembled PC material together. The three smaller pieces fit into the integral slots moulded into the board into the base of the case lid of the plastic case. and secure it with four small self-tapping screws. Wire up the method, the current drain from the per tracks. Repair any faults before 9V battery clips and multimeter leads two 9V batteries must always be the assembly of components. Begin by as shown. Prepare the two wires for same and the battery life will be exinserting the two links and all the switch S1. tended. For the same reason, LED1 is resis­ tors. The accompanying table Fit the Dynamark adhesive label to connected across the full 18V supply can be used as a guide for the resistor the lid of the case and drill and file via a 10kΩ resistor. colour codes. Alternatively, use your out the holes for the switches and multimeter to check each resistor as LED. Attach S1 with the screws and Construction it is installed. connect its wiring. Next, insert and solder in the PC The S ILICON C HIP Sound Level The rear end panel can be drilled stakes. These are located at all external Meter is housed in a plastic case to accept a small grom­met. Pass the measuring 150 x 80 x 30mm and wiring points, the ‘A’ positions and for multimeter leads through the gromuses a PC board coded 04312961 the eight switch terminal locations for metted hole and attach the banana and measuring 67 x 120mm. The S2 and S3. plugs to it. microphone is held inside a copper Next, the ICs can be inserted and Microphone mounting tube which protrudes from the front soldered in. Take care with the oriof the case. This is done to prevent entation of each and make sure that An 80mm length of 12.7mm copper sound reflections from the case from IC5 is the TL071 (or LF351). Diodes tube is soldered to a 12 x 30mm piece upsetting the read­ing. D1-D4 can now be inserted, taking care of 1mm thick copper sheet (or PC to ensure that they are also correctly board). The copper sheet becomes a Fig.5 shows the component layout oriented. Switches S2 and S3 can be flange for easy attachment to the front for the PC board. You can start construction by checking the PC board mounted by soldering their pins to the end piece of the box. Drill holes in top of the PC stakes. for any shorts or breaks in the copthe flange and front end plate to allow Fig.6: this is the set up used for calibrating the Sound Level Meter. It relies on using a speaker of known sensitivity. Most manufacturers quote sensitivity figures for their loudspeakers. 88 Silicon Chip’s Electronics TestBench it to be secured with two screws and nuts. Also drill a hole central to the flange and end plate for the shielded cable to pass through the tube. The tube and flange can be painted if desired. Connect the microphone using shield­ed cable and attach some heat­ shrink tubing around its body. Shrink the tubing down with a hot air gun and insert the wire and microphone into the tube. Leave the microphone flush with the end of the tube. The flange can be attached to the end plate of the case with the screws and nuts. The shielded cable is clamped with a solder lug attached to one of the screws. The batteries are held in place on the lid of the case using three pieces of double-sided PC board (73 x 5mm) which are inserted in the integral slots. Two pieces of double sided PC board, measuring 30 x 15mm, are soldered in place between the transverse pieces so that they provide a snug fit for the battery and clip assemblies. Check that the lid will fit onto the base of the case. Voltage checks Switch on and connect the red multimeter lead from the Sound Level Meter to the common input of the multimeter and then measure voltages on the circuit with the other lead of the multimeter. Check that there is +9V at pin 8 of IC1, IC3 and IC4; at pin 7 of IC5; and at pin 2 of IC2. There should be -9V at pin 4 of IC1, IC3, IC4 & IC5 and at pins 10 & 16 of IC2. REF1 should have -2.49V at its anode and pin 3 of IC4b should be -20mV. LED1 should also be lit. Connect both output leads from the sound level meter to the multimeter. Performance ‘A’ response .......................................... -18dB at 100Hz, -10dB at 20kHz (see Fig.1) ‘C’ response ......................................... -5dB at 20Hz, -13dB at 20kHz (see Fig.1) Overall flat response (input versus multimeter reading) .................. -3dB at 28Hz and 50kHz Log conversion accuracy at multimeter output ................................ <0.5dB over a 100dB range from 0.550V RMS to 5.5µV input level Temperature stability ............................ <10mV (1dB) change per 30°C Slow response time constant ............... 9.4 seconds Fast response time constant ................ 2 seconds Peak response ...................................... 1.5ms attack; 120ms decay Power ................................................... 12-18V at 32mA Microphone Performance (ECM-60P A version) Sensitivity �������������������������������������������� -56dB ±3dB with respect to 0dB+1V/µbar <at> 1kHz Microphone response .......................... within ±3dB from 50Hz to 3kHz and ±6dB from 3kHz to 8kHz. Above 8kHz and below 50Hz unspecified. Maximum SPL ..................................... 120dB Note: filter responses measured at VCA output with control input (pin 11) grounded. the multimeter reading is 400mV. If it is greater than 400mV, rotate VR1 slightly clockwise. Conversely, if the multimeter reading is less than 400mV, rotate VR1 slightly anticlockwise. Now measure the difference again with the 0dB/ -60dB switch. You will note that the reading will now not be 1V for the 0dB setting. However, what we are looking for is a 600mV change between the 0dB and -60dB pink noise level settings (ie, 10mV per dB). After some repeat adjust­ments of VR1 it should be possible to obtain close to 600mV variation between the 0dB and -60dB settings. Calibration now only requires the Calibration Calibration is done in two steps and a pink noise source is required for both steps. We will describe a suitable pink noise source in next month’s issue of SILICON CHIP and we assume that you will also build that or have access to an equivalent source. First, connect the pink noise source to the electret microphone input of the sound level meter. Select 0dB on the pink noise source (equivalent to 60mV RMS) and adjust trimpot VR2 for a read­ing on the multimeter of 1V DC. Now switch to -60dB on the pink noise source and check that Fig.7: check your etched PC board against this full-size artwork before installing any of the parts. Silicon Chip’s Electronics TestBench  89 PARTS LIST 1 plastic case, 150 x 80 x 30mm 1 PC board, code 04312961, 67 x 120mm 1 front panel label, 75 x 144mm 1 ECM-60P type A electret microphone (sens. -56dB with respect to 1V/1µbar at 1kHz) 3 pieces of double sided PC board, 73 x 5mm 2 pieces of double sided PC board, 30 x 15mm 1 DPDT slider switch and mounting screws (S1) 2 DP3P slider switches (S2,S3) 1 50kΩ horizontal trimpot (VR1) 1 100kΩ horizontal trimpot (VR2) 2 9V battery snaps 2 9V batteries 1 black banana plug 1 red banana plug 1 250mm length of shielded cable 1 500mm length of black hookup wire 1 500mm length of red hookup wire 1 50mm length of 0.8mm tinned copper wire 30 PC stakes 2 3mm x 10 screws and nuts 4 small self-tapping screws (to secure PC board) 1 solder lug 1 small rubber grommet 1 small cable tie 1 SSM2018P voltage controlled amplifier (IC2) 1 TL071, LF351 op amp (IC5) 4 1N914 signal diodes (D1-D4) 1 LM336-2.5 2.5V reference (REF1) 1 3mm red LED (LED1) Semiconductors 3 LM833 dual op amps (IC1,IC3,IC4) Miscellaneous 12mm diameter heatshrink tubing, solder. Capacitors 1 470µF 16VW PC electrolytic 5 100µF 25VW PC electrolytic 1 47µF 16VW PC electrolytic 3 10µF 16VW PC electrolytic 1 0.56µF MKT polyester 1 0.22µF MKT polyester 1 0.18µF MKT polyester 2 0.15µF MKT polyester 1 .047µF MKT polyester 2 .0027µF MKT polyester 1 .001µF MKT polyester 1 100pF ceramic 1 33pF ceramic 1 12pF ceramic Resistors (0.25W 1%) 1 2.2MΩ 2 12kΩ 1 560kΩ 9 10kΩ 1 180kΩ 1 8.2kΩ 3 100kΩ 1 6.8kΩ 1 68kΩ 1 5.6kΩ 1 33kΩ 2 4.7kΩ 1 24kΩ 2 3.9kΩ 1 22kΩ 2 150Ω 6 20kΩ 1 100Ω 1 18kΩ offset adjustment trimpot VR2 to be set. This is done using the setup shown in Fig.6. You will need an amplifier, the pink noise source and a woofer or tweeter with known sensitivity. All manufacturers of loudspeakers provide a sensitivity rating for their units and these are specified as a dBSPL when driven at 1W and at 1m on axis. Note that if you use a tweeter, the manufacturer’s speci­fied filter should be used when making the measurement. For example, a loudspeaker may be rated at 88dB when mount­ed on a baffle and driven from a 2.828V AC source at a distance of 1m. The loudspeaker impedance is 8Ω. Note that 2.828V into 8Ω is equivalent to 1W. Use your multimeter to measure the voltage applied to the loudspeaker and set the amplifier’s volume control to deliver 2.828V AC for an 8Ω system and 2V AC for a 4Ω speaker. Be sure to set your amplifier’s tone controls to the flat settings (ie, centred or switched off) and make sure that the loudness switch is off. Now connect the multimeter to the sound level meter (with the unweight­ ed and slow settings selected) and with the micro­phone at 1-metre and on axis to the speaker. Adjust trimpot VR2 to obtain the loudspeaker sensitivity. For our 88dB example, the multimeter should read 0.88V or 880mV DC. Alternatively, if you have a calibrat­ ed sound level meter, adjust VR2 for the same readings. Make sure that both sound level meters are set with the SC same filtering and responses. 90 Silicon Chip’s Electronics TestBench (10mV/dB) CONNECT TO MULTIMETER FILTER C-WEIGHTING A-WEIGHTING UNWEIGHTED + SOUND LEVEL METER RESPONSE SLOW FAST PEAK + OFF + ON + Fig.8: this is the actual size artwork for the front panel.