46 CIRCUIT CELLAR® • www.circuitcellar.com

Mar

ch 2

010

– Iss

ue 2

36

igure 1—This is a wiring diagram of the unit (System 1) with theamplifier and attenuator.

2103018_Cicchinelli.qxp 2/11/2010 10:51 AM Page 46

You can build your own calibrated decibel meter with a little know-howand four main parts: an audio amplifier, a meter circuit, a display, and apower supply. With this design, you can accurately measure the outputlevel of most audio systems.

Calibrated Decibel Meter Design

I

FEA

TURE

ARTICLEby Larry Cicchinelli

have been interested in measuring various electrical properties since my earliest days working with ham

radio. My profession for over 30 years was to design andimplement test equipment, and I have been working withpublic address systems of one sort or another for almostas long. The project described in this article is an out-growth of both of these experiences.

I built a calibrated decibel meter to accurately measurethe output level of almost any audio system. The systemI finally implemented contains four parts: an audio ampli-fier, a meter circuit, a display, and a power supply (seePhoto 1). I actually built two versions, which I will detailafter I describe the circuits. Figure 1 shows the first sys-tem’s wiring. Figure 2 shows the second system’s wiring.

dB VS. dBmLet me start off with a few comments about decibels (dB)

and decibels referenced to 1 mW (dBm). Whenever a spec-ification talks about “dB,” it is a relative value referenc-ing gain, which can just as easily be positive or negative(loss). When the discussion is about “dBm,” or any other“dB” with a third letter, then there is an absoluteinvolved. In the case of “dBm,” the reference is 1 mW toa known resistance—usually either 500 or 600 Ω foraudio systems and 50 Ω when discussing RF systems.For this system, I chose a reference of 500 Ω. This yield-ed 0.707 VRMS, or 2 VPP for 0 dBm.

AUDIO AMPLIFIERThe audio amplifier contains four gain stages, three of

which have selectable gain: 0/40 dB, 0/20 dB, and 0/10 dB

(see Figure 3). The total gain of the subsystem is selec-table in 10 dB steps from 0 to 70 dB. This is sensitiveenough for most microphones. I did not implement phan-tom power in my system, but it would be easy enough toadd. One method I’ve used involves inserting a trans-former with a center tap on the primary. The center tapwould be connected to the phantom power supply line.Most professional systems use 48 V, but the 15 or 23 Vavailable in this system may work for you.

In dealing with the gain values used in this subsystem,the ratio of the resistor values is more important than

FFiigguurree 11——This is a wiring diagram of the unit (System 1) with theamplifier and attenuator.

www.circuitcellar.com • CIRCUIT CELLAR® 47

Mar

ch 2

010

– Iss

ue 2

36

wanted to put the highest-gain cir-cuit first so that a low-l evel signalwould get amplified early in order t omaintain the signal-to-noise ratio.

The remaining two selectable gainstages are similar to the first. In eachof the three gain stages, the lowervalue gain is enabled by shor ting outthe higher value resistor in the feed-back loop. Theswitches I usedwere SPDT(because I hadthem on hand), butSPST would do justas well. Each gainstage has provisionfor capacitors inthe feedback loop. Ifound that thesewere necessary inorder to eliminateoscillations due tothe high gain val-ues. I designed thecircuits so that

their absolute values. Also, most ofthe resistors I used were 1% sincethey cost the same as 5% resistors;however, even 5% values wouldyield acceptable results for mostapplications. Just to prove this withnumbers, let’s calculate the gain ofthe 40-dB stage using one set of“worst-case” 5% values. Let R6 =950 Ω (5% low) and let R8 = 105 kΩ(5% high). The gain of the stagewould then be 106K/950 = 11 1.6,which is equivalent to almost 41 dB.For most applications this would bequite acceptable.

All of the op-amps in the subsys-tem are National SemiconductorLM833s, which I’ve used in variousaudio systems for the past 30 yearsand find that their specs hold upwell. The first stage has 0- dB gainand simply converts a balanced sig-nal to an unbalanced one. The inputsignal can just as easily be an un bal-anced signal. The second stage has aselectable gain of either 0 or 40 dB. I

each can have a capacitor , but youmay find that the lower gain stagesdo not need it. With the specifiedvalues, the system has a bandwidthof at least 20 kHz.

Note that both the 40- and 20-dBstages have offset potentiometers.These allow you to adjust f or theinput offset voltage of the op- amps.They should be adjusted with no sig-nal input and with the higher gainselected for the stage being adjusted.You should do one stage at a time:select the higher gain value andadjust the potentiometer for 0 V atH6.

The last op-amp in the circuit pro-vides another 10 dB of gain for drivingthe display amplifier. I wanted a signalwith a higher amplitude than 0 dBm todrive the rectifier circuit. With all threegain switches set to 0 dB, a 0-dBm sig-nal applied to the input will present a10-dBm (2.23 VRMS) signal to the inputof the display amplifier.

Notice that there is an “extra” op-amp stage at the output driving H7.This is an auxiliary output that canbe used to drive whatever you maywant, perhaps a set of headphones,without affecting the main output.The two resistors in the circuit (R25and R26) set the gain to 2. However,you can just as easily change the val-ues to whatever you think appropri-ate—within reason! One way to dothis would be to replace R26 with avariable resistor which would allowyou to adjust the gain. If you do this,I suggest that you include a fixedresistor as well in order to keep from

Photo 1a—System 1 is the unit that has both the selectable gain amplifier and a step atten-uator. b—System 2 is the unit with just the attenuator .

a)

b)

Figure 2—This is a wiring diagram of the unit (System 2) with justthe attenuator showing the provision for having the amplifier as aseparate unit.

2103018_Cicchinelli.qxp 2/11/2010 10:51 AM Page 47

Bo

oks

Learn more about C# programming and .NET

C# 2008 and .NETprogrammingThis book is aimed at Engineers and Scientists

who want to learn about the .NET environ-

ment and C# programming or who have an

interest in interfacing hardware to a PC. The

book covers the Visual Studio 2008 develop-

ment environment, the .NET framework and

C# programming language from data types

and program flow to more advanced concepts

including object oriented programming.

240 pages • ISBN 978-0-905705-81-1• $47.60

Prices and item descriptions subject to change. E. & O.E

The world of electronics at your fi ngertips!

Elektor ShopBOARDS, BOOKS, DVDs AND MORE AT WWW.ELEKTOR.COM/SHOP

Microsoft

approved!

Several case studies included

PIC cookbook for Virtual InstrumentationThe software simulation of gauges, control-knobs, meters and indicators which behave just

like real hardware components on a PC’s screen is known as virtual instrumentation. In this

book, the Delphi program is used to create these mimics and PIC based external sensors are

connected via a USB/RS232 converter communication link to a PC. Several case studies of

virtual instruments are detailed inclu ding a compass, an oscilloscope, a digital and analogue

thermometer, a FFT-based Frequency analyser, a joystick, mouse-control panels and virtual

displays for cars and aircraft. Full source code examples are provided both for several differ-

ent PIC’s, both in assembler and C, together with the Pascal code for the Delphi programs

which use different 3rd party Delphi virtual components.

264 pages • ISBN 978-0-905705-84-2 • $47.60

This book and more

are available at

www.elektor.com/booksConnect your mouse into new embedded

applications

Mouse InterfacingThis book describes in-depth how to connect

the mouse into new embedded applications.

It details the two main interface methods,

PS/2 and USB, and offers applications guid-

ance with hardware and software examples

plus tips on interfacing the mouse to typical

microcontrollers. A wide range of topics is

explored, including USB descriptors, a four-

channel, millivolt-precision voltage refer-

ence all with fully documented source-code.

A wealth of clear illustrations enhance this

highly readable text.

256 pages • ISBN 978-0-905705-74-3 • $42.80

48_49.qxp 2/9/2010 10:32 AM Page 1

Kits &

Mo

du

les

Elektor US4 Park StreetVernon CT 06066USAPhone: 860-875-2199Fax: 860-871-0411E-mail: sales@elektor.com

CD

/DV

D-R

OM

s

OBD2 Analyser NGThe compact OBD2 Analyzer in the June

2007 issue was an enormous success — not

surprising for an affordable handheld on-

board diagnostics device with automatic

protocol recognition and error codes ex-

plained in plain language. Now enhanced

with a graphical display, Cortex M3 proces-

sor and an Open Source user interface,

the next generation of Elektor’s standalone

analyser sets new standards fora DIY OBD2

project. The OBD2 Analyser NG is self-

contained and can plug into any OBD diag-

nostic port.

Kit of parts including DXM Module, PCB SMD-

prefitted, case, mounting materials and cable

Art.# 090451-71 • $135.50

Software Defined RadioSD radio receivers use a bare minimum of

hardware, relying instead on their software

capabilities. The Elektor SDR project (by Burk-

hard Kainka) demonstrates what’s achie-

vable, in this case a multi-purpose receiver

covering all bands from 150 kHz to 30 MHz.

It’s been optimised for receiving DRM and

AM broadcasts but is also suitable for listen-

ing in to the world of amateur transmissions.

The designer’s aim for this project was to

create a recei ver displaying high linearity

and phase accuracy. Development was

focussed on the characteristics that were

most important for a top-notch DRM recei-

ver and the end result is a receiver with remar-

kable interference rejection characteristics!

Art. # 070039-91 • $139.60

Elektor is morethan just your favorite electronics magazine.

It’s your one-stop shop for Elektor Books,

CDs, DVDs,Kits & Modules

and much more!

www.elektor.com/shop

Hot product!

More than 100 Elektor articles included

DVD LED ToolboxThis DVD-ROM contains carefully-sorted

comprehensive technical documentation

about and around LEDs. For standard mo dels,

and for a selection of LED modules, this Tool-

box gathers together data sheets from all the

manufacturers, ap pli cation notes, design

guides, white papers and so on. It offers sev-

eral hundred dri vers for powering and con-

trolling LEDs in different configurations,

along with ready-to-use modules (power

supply units, DMX controllers, dimmers,

etc.). In addition to optical systems, light de-

tectors, hardware, etc., this DVD also ad-

dresses the main shortcoming of power LEDs:

heating.

ISBN 978-90-5381-245-7 • $46.00

110 issues, more than 2,100 articles

DVD Elektor 1990through 1999This DVD-ROM contains the full range of

1990-1999 volumes (all 110 issues) of Elek-

tor Electronics magazine (PDF). The more

than 2,100 separate articles have been clas-

sified chronologically by their dates of publi-

cation (month/year), but are also listed

alphabetically by topic. A comprehensive in-

dex enables you to search the entire DVD.

The DVD also contains (free of charge) the

entire ‘The Elektor Datasheet Collection

1…5’ CD-ROM series, with the original full

datasheets of semiconductors, memory ICs,

microcontrollers, and more.

ISBN 978-0-905705-76-7 • $111.30

Bestseller!

LED Topwith Special EffectsIf you fit a line of LEDs on a circular PCB and

power them on continuously, they generate

rings of light when the board is spun. If you

add a microcontroller, you can use the same

set of LEDs to obtain a more interesting

effect by generating a ‘virtual’ text display.

The article also describes a simple tech-

nique for using the Earth’s magnetic fi eld

to generate a synchronisation pulse. The

potential applications extend from rotation

counters to an electronic compass.

Kit of parts incl. SMD-stuffed PCB and pro-

grammed controller

Art. # 080678-71 • $67.80

48_49.qxp 2/9/2010 10:32 AM Page 2

50 CIRCUIT CELLAR® • www.circuitcellar.com

Mar

ch 2

010

– Iss

ue 2

36

adjusting to an infinite gain!

DISPLAY DRIVERThe display driver (see Figure 4)

uses LM3916 and LM3915 devicesto drive the LEDs. The L M3916has outputs that are close to thoseof the standard VU meter mark-ings: +3, +2, +1, 0, –1, –3, –5, –7,–10, and –20 dB. The onl y onemissing is the –2-dB indication.This subsystem does not use the–20-dB output of the LM3916. Th eLM3915 is used to extend therange to –40 in 3-dB steps startingat –13 dB.

I found that the rectif ier circuitI used has excellent linearity fromclose to 0-V input up to a signalwhich yields about a 7-VDC out-put. Because of this, I calculatedR2 and R3 such that the compara-tors in the two LED drivers haveclose to a 7-V reference value.This is the voltage necessary toturn on the highest LED. Theresistors that control the gain ofthe rectifier (R6, R7, and R8) werechosen so that R7 can be adjustedto display 0 dB with a 10- dBm inputsignal. One easy way of calibratingthe unit is to apply –1.7 VDC to thejunction C4, R6, and R12 an d thenadjust R7 until the 0-dB LED tur ns

on. This may get you close enou ghso you don’t need to use a knownaudio level.

The rectifier circuit and the gain cir-cuit of IC3B were both derived from

the LM3916’s datasheet. The reasonfor the 16 dB gain for IC3B is so thatwhen a –3-dBm signal is presented tothe rectifier it will cause the LM3915to turn on its most s ignificant LED

Figure 4—This is the audio rectifier and display driver .

Figure 3—This is a schematic of the amplifier ’s circuit board with selectable gain from 0 dBto 70 dB in 10-dB steps.

2103018_Cicchinelli.qxp 2/11/2010 10:51 AM Page 50

(see Table 1). Notice that thereis 16-dB difference between thetop LEDs of the two devices.

LED BOARDThe LED board is nothing more

than a small PCB for mountingthe LEDs (see Figure 5). At onetime, I thought about usingLED bar arrays, but they areexpensive and it is hard to findthem with appropriate colorpatterns. I designed the displaydriver and LED boards so theycan be connected using a rightangle pin header with 0.1″ spac-ing. You can solder the headeronto both boards or solder theheader to one board and use asocket on the other. I used aright angle header on the LEDboard and a socket on the dis-play board. I suggest you mountthe connectors first withoutsoldering so you can see how the ywill assemble together.

Be careful when installing theLEDs. You need to ensure that youinstall them in the cor rect order aswell as orientation. On the LEDsspecified in the BOM, the cathodelead has a more pronounced bend.Theoretically, you can install the LEDson either side of the board. I installedthem on the front—the side withoutthe copper traces. Note that there is asymbol etched in copper showing theLED orientation on both sides of theboard. Also, the position of LED20(the power indicator) is etched on thebottom side; this will make it easier foryou to determine which color LEDs go

where. See Figure 6 for System 1. Youcan see in Photo 1B that I replaced D6with a green LED simply because Iran out of orange ones (see Figure 7).

SYSTEM POWERThis subsystem develops three DC

voltages: 15 V, –15 V, and 23 V (seeFigure 8 and Figure 9). The LED driv-ers have two modes of operation.One mode essentially puts the LEDstrings in series for each of the driv-ers individually. The other mode h aseach LED drawing its cur rent sepa-rately. I chose the f irst mode becauseit enables both strings of LEDs todraw a constant current as long as atleast one LED is on and should cause

less noise. However, this moderequires a supply voltage highenough to handle 10 LEDs inseries.

In the unit I built, the 23 V i sactually developed by stackingan 8-V regulator on “top” ofthe 15 V. Each regulator gets itsinput from a half-wave rectifi-er. The regulator subsystem ispowered by a 24-VAC “wallwart.” Anyone building thepower supply can just as easilyuse a 24-V regulator instead theof stacking method I used.There is a jumper (H3) on theboard that allows you to selecteither method. Connecting pins1 and 2 is for a sin gle 24-V reg-ulator. Connecting pins 2 and 3is for stacking an 8-V regu latoron top of the 15-V regulator .You do not have to install aheader in this position. You

should just insert a wire between theappropriate points and solder it inplace.

All of the regulators are linearbecause I did not want th e noiseassociated with switchers. I al sobuilt the regulators in a separate boxbecause I did not want th e 120 VACin the same box as a system capableof 70-dB gain.

SYSTEM ASSEMBLYThe first version I built (see Figure

6) has all circuits, except the powersupply, in one aluminum box. Thissystem can be used to measure low-level signals with a practical lowerlimit of about –90 dBm. If you tr y touse the maximum available gain ofthe system you will probably havequite a few LEDs lit just due to noise.I have found that I can reliably use thesystem with up to 60-dB gain selected.

The system has two inputs: lowlevel and high level. The low-level sig-nals are routed to the 70-dB amplifierwhile the high level signals are routedto the voltage divider. The five-posi-tion switch allows you to select whichof the two sources are being measured.With this arrangement, you can meas-ure signals from about –90 to +40 dBm(70 VRMS) referenced to 500 Ω—a rangeof 130 dB.Figure 5—The LED circuit board

www.circuitcellar.com • CIRCUIT CELLAR® 51

Mar

ch 2

010

– Iss

ue 2

36

Table 1—This table shows the signal levels into the twodisplay drivers and the LED indication values.

Signal Level (dBm) Indication (dBm)13 312 211 110 09 –1

LM3916 7 –3

5 –5

3 –7

0 –10

–3 –13

–6 –16

–9 –19

–12 –22

LM3918 –15 –25

–18 –28

–21 –31

–24 –34

–27 –37

–30 –40

2103018_Cicchinelli.qxp 2/11/2010 10:51 AM Page 51

52 CIRCUIT CELLAR® • www.circuitcellar.com

Mar

ch 2

010

– Iss

ue 2

36

2.23 VRMS is applied to the input ofthe display driver. This is about 0.6 Ω,which is 28 dBm into 8 Ω. However,30 dBm is equivalent to 2.83 VRMS

into 8 Ω. The adjustment range ofthe circuit will easily allow you toset this as the “0 dB” indica tion. Youcan then use this circuit, withoutmodification, with calibrated rangesof 30, 40, 50, and 6 0 dBm into 8 Ω.An interesting point to note is that30 dBm is equivalent to 0 dBW – asignal level referred to 1 W. Whatthat means is that the display willshow values of 0, 10, 20, an d 30 dBW.

FILESSeveral file types are included: .dch

(the DipTrace schematic), .dip (theDipTrace PCB), .gbr (Gerber), .drl(drill), and .xls (BOM). The main filenames should be fairly obvious as towhich circuit they apply. The Gerberfiles and drill files can be sent to aPCB manufacturer to make the boards.I have had good success with this forseveral of my designs using this soft-ware. You can also use the PCB filesand print them full size if you want toetch your own. Although I have notmade the SilkScreen files, you can doso from the PCB files. There is a free

version of DipTrace avail-able on their website.

The dbMeterBOM.xls fileon the Circuit Cellar FTPsite is a combined BOM ofall four circuits individuallyas well as a consolidatedBOM. However, I did notadd the quantities in theconsolidated BOM. TheFrontPanelLabels.dip filehas the front panel labelsfor the LEDs and the fiveposition switch I used. Thiswas created using the Dip-Trace printed circuit pro-gram. By using variousPrint options, I was able toprint both a negative and apositive image, as can beseen on the pictures of thetwo systems. The systemswork as expected, but myconstruction skills are notthe best.

The headers and mating

Some care needs to be used in select-ing the resistors if you want to meas-ure the high-level signals. The totalresistance is 3.2 kΩ, and with 70 Vacross, it there will be about 22 mAthrough each of the resistors. Thismeans that R1 will have to dissipate alittle over 1 W. If youimplement this circuit, Istrongly recommend youuse a 3-W resistor—just tobe on the safe side. Also, R2will dissipate about 0.3 W,so a 1-W resistor would be agood idea. The remainingtwo can be 0.25 W.

Figure 7 does not havethe 70-dB audio amplifierand can be built in a small-er enclosure. It has thecapability of directly meas-uring the aforementionedhigher-level signals. It alsohas two input connectors.One is connected directly tothe attenuator while theother is connected to theswitch. This scheme allowsyou to build the 70-dBamplifier in a separate enclo-sure and connect its outputto the second input.

You can also use this

system to measure signals which arenot based on 500 Ω. Suppose you wantto monitor the output of an amplifierdriving an 8-Ω load. All you have todo is calibrate the system for theappropriate voltage values. Asdesigned, 0 dB will be indicated when

Figure 7—This is the unit (System 2) with just the attenuator .

PowerBox.dch

Power.dch

dbMeterDisplay.dch

dbMeterLEDs.dch

Display Box-2.dch

24 VAC

S13 H13

J1 S1 H1

J3

SW1

J4

H2S2

J2P2

J11

S11H11H12

H14

H15

Figure 6—The unit (System 1) with the amplifier and attenuator .

PowerBox.dch

Power.dch

dbMeterDisplay.dch

dbMeterLEDs.dch

dbMeterAmplifier.dch

Display Box-1.dch

24 VAC

S13 H13

J1 S1 H1

S3 H3J3

J4

S4H4

H6S6

SWI

H2S2

J2P2

J11

S11H11

S12H12

H14

H15

2103018_Cicchinelli.qxp 2/11/2010 10:51 AM Page 52

the connection between the dis-play board and the LED boa rd.I’ve used this method of co nnec-tion for most of my projects. It’sbeen quite useful and wor th theextra cost—about $0.16 per con-nection. I use a small pair ofneedle nose pliers to crimp thewire into the socket pins while

viewing the process using a 3× mag-nifying lens. A crimp tool would be alot easier, but it would be expensive.You do not need to use this method.You can just as easily insert wiresinto the pad locations and solderthem.

Feel free to email me if you haveany questions about this project. I

www.circuitcellar.com • CIRCUIT CELLAR® 53

Mar

ch 2

010

– Iss

ue 2

36

connectors were built using fourparts from Jameco: a pin header(160882), pin header housing(103158), socket pins (100766), and asocket (200783, only if you use thissocket on the display driver board).The total number of connectionsusing these in System 1 (see Figure 6)is approximately 35—not including

Figure 8—Power supply wiring diagram

PROJECT FILESThere is no code associated with this ar ticle. To download additionalproject files, go to ftp://ftp.circuitcellar.com/pub/Circuit_Cellar/2010/236.

RESOURCENational Semiconductor Corp., “LM3916 Dot/Bar Display Driver ,”DS007971, 2000, www.national.com/mpf/LM/LM3916.html.

SOURCESLM3915/LM3916 Display drivers and LM833 op-ampNational Semiconductor Corp. | www.national.com

Larry Cicchinelli (k3pto@arrl.net) holds a BSEE from The Drexel Institute of T echnolo-gy and an MSES from Pennsylvania State University . He has been a technical sup-port manager at Digi International (Rabbit Brand) since 2000. From 1967 to 2000,Larry worked for Ford Motor Company. He has been licensed as K3PT O since 1961.

Figure 9—Power supply schematic

2103018_Cicchinelli.qxp 2/11/2010 10:51 AM Page 53