NOT TO BE FORGOTTEN
Jul 1, 1999 12:00 PM, Glen Ballou
We have all heard about the great newly available measuring instrumentsthat are used to measure audio circuits, instruments like FFT (Fast FourierTransform) and TEF (Time-Energy-Frequency) machines, which measure audiosystems more accurately and quickly than those instruments offered a fewyears ago. We should not, however, forget many of the those olderinstruments; they were good then and are still good today. For instance,how do we measure distortion, frequency response, resonance, loudspeakerrattles, phase differences and impedance, for example? Much of this can betested with state-of-the-art test gear, but they can also be tested withthe simple, old-fashioned gear.
How can we measure an amp's characteristics? To measure frequency response,we can insert a sine-wave signal into the input and measure the outputsignal. If the input signal is the same level at all frequencies, thefrequency response will be the variation of the output level. A sine-wavegenerator works well for the input signal. Most sine-wave generators todayhave low impedance output, so the input impedance of the device-under-test(DUT) does not affect output. If we sweep the input of the DUT with a sinewave from the lowest to the highest frequency desired, we can measure theoutput of the DUT with a voltmeter, wattmeter or dB meter. If the internaloutput impedance of the DUT is low compared to its specified outputimpedance and the impedance of the test instrument, the DUT will notrequire loading. If we do not know the internal impedance of the DUT, weshould load the output with the recommended impedance.
If we are measuring the frequency response of a power amp rated at 8 V, andwe do not know the internal impedance of the amp, we should load it with an8 V load. Remember, however, a 100 W amp requires a 100 W resistor if weare measuring it at full load. We might want to measure the amp frequencyresponse at more than one power level (at 1 W, one-half power and fullpower). Although measuring at 1 W or half power can usually be done withoutmonitoring the output with a scope, when measuring at full power, we mustmonitor the waveform. If the amp has a rising response at some frequencyother than the reference frequency, we may not see the true frequencyresponse on the meter because the amp will clip. Under this condition, theresponse will appear better than it really is. If we want to determine thepower frequency response in dB referenced to our reference frequency whenwe measured the output in volts, we can use the equation:
dBvariation = 10log10[(Vout)[superscript]2/(Vref)[superscript]2]
To determine the voltage response in dB, use the equation:
dBvariation = 20log10[(Vout)/(Vref)]
Output in negative numbers means that the output is less than the referencevalue.
With a distortion analyzer, we can check distortion of the DUT using thesine-wave generator as the signal device. Most harmonic distortionanalyzers tune out the fundamental frequency and measure whatever isremaining. This turns out to be odd and even harmonics and noise. It isimportant that the sine-wave generator has low distortion and low noise;otherwise, the readings will include the source and will consequently bemeaningless.
If we have a two-channel oscilloscope, we can put the input to the DUT onchannel 1 and the output on channel 2. If we superimpose the input on theoutput and set the levels so they are the same height, we can see thedistortion but cannot give it a value (see Figure 1). Still, this can beuseful. The waveform in Figure 1 shows clipping on the positive half of thesignal and crossover distortion.
We can also see phase shift and polarity reversal with the two-channelscope and sine-wave generator (see Figure 2). This can be useful if we wantto keep the input wave and the output wave the same polarity, which isimportant when micing and amplifying live music.
Sine-wave generators are also useful for finding resonance, standing wavesand loudspeaker rattles. If you suspect a circuit is unstable and sometimesgoes into oscillation, apply a sine-wave generator to the input and anoscilloscope to the output, being sure to load the output with the load youwill be applying in the final installation. Sweep the DUT from the lowestfrequency to the highest frequency you feel is appropriate and monitor theoutput. Remember that the system may go into oscillation or start ringingat a frequency well above the audio spectrum, so run the sine-wavegenerator as high as possible, even into the RF bands, if possible.Shocking the system by hitting the input with an instant signal is also agood idea. We can accomplish this by setting the sine-wave generator at theproper output level and removing and reinserting the input connector. Ifthe DUT is stable and properly designed, this will not hurt it, but canfind an unstable condition. If, however, the DUT is unstable or notdesigned to withstand transients, we could destroy it. If we do destroy itwhile testing it, we should be satisfied; it is always better to findproblems before the devices is installed.
Standing waves in rooms occur because of the room's dimensions. All roomshave standing waves, but some are more obvious than others. If the room ishighly damped, that is, it has a lot of absorption in it, standing wavesmay not be obvious. Standing waves, however, are most annoying at lowfrequencies where absorption is more difficult to achieve. A room with twosets of opposing walls with heavy absorption and one set of opposing wallswith little absorption exhibits a definite echo. The echo or standing waveis always there, but in the above case, it is obvious because the wavesthat normally help to mask it from the other two parallel surfaces areabsorbed. The unwanted echo has not increased; it has always been there-itwas just masked. This is why a highly reverberant room sounds rather smoothand echo free; each reflection masks the other reflections. Fuzz one set ofparallel walls, and we are apt to hear two distinct echoes. Fuzz two setsof parallel walls, and we will hear one distinct echo.
Even in a relatively dead room, we can find spots where a particularlow-frequency wave is strong. This is not the place we want to place a micin our sound system unless we want guaranteed feedback. How do we know itis there? The sine-wave generator comes to the rescue again. Sweep thesystem with a sine-wave generator and either stand at the spot where themic is to be placed and listen, or better, use a sound level meter (SLM) inthat spot and monitor the output. If you do not have an SLM, you can placea mic in the spot (preferably the mic you will be using in the finalinstallation) and look at the output with an oscilloscope or a voltmeter.Move the meter just a foot or two (305 m or 710 m) and notice that thestanding wave goes away. Do not, however, get overconfident because a newwave may appear at that new spot. It will be a frequency that has awavelength related to the new distance. The frequency can be calculatedusing the equation:
Frequency = 1,130 ft/s(wavelength)
where: 1,130 ft/s (344 m/s) is the speed of sound and wavelength is thelength of one cycle of the frequency in question.
If we measure the distance from the source to the mic, we can use that asthe wavelength of the frequency in question. This does not have to be astraight line; it may be a reflection off a wall, floor or ceiling, or allthree.
We just looked for a feedback type of frequency. What if we are installinga home theater or stereo system? The predominant problem is now the lostfrequency. Low frequencies have a tendency to go through house walls andwindows rather than reflect back into the room to fill it with beautifulbass. A signal lost is a signal lost forever; no amount of EQ will everbring it back.
So what are we to do? Never place the best seat in the house at the spotwhere the signal is at a minimum. Sweep the system with a sine-wavegenerator to be sure there is not a big hole where the master's ears willbe. If there is, move the chair or reposition the loudspeakers until asweep is as smooth as possible. Do not try to boost the system with EQ; itwill not bring the lost signal back, but it will apply two bumps, one oneach side of the lost frequency.
If we want to measure the frequency response of our complete system with asine-wave generator and an SLM, we will have to warble the tone to averageout the standing waves and other anomalies. Warbling can be accomplished bymoving the frequency dial back and forth while sweeping the generator. Thisis the way measurements were made before pink noise and RTAs. This may notbe as accurate as today's method, but if all we have is a sine-wavegenerator and an SLM or voltmeter, it will give us a reasonable frequencyresponse of the room and the system.
How many times have we had a loudspeaker that rattles or an object in theroom that rattles? With a sine-wave generator, it is easy to locate. All weneed to do is sweep the system and listen for the rattle. If it is anobject in the room, maybe we can tighten it or move it to where it is notaffected by the loudspeaker output. If it is in the loudspeaker, it may bedue to poor construction, a loose driver, a loose or resonant grill, orpossibly the loudspeaker cone is going into resonance because we are tryingto use the loudspeaker below its cutoff or at too high a level. At anyrate, it gives us information we can use to correct the problem. As we cansee, the sine-wave generator, which has probably been sitting on the topshelf collecting dust, is still a useful tool.
Most of us today have a pink-noise generator (PNG) and a real time analyzer(RTA), which we probably use to measure and EQ rooms, but there are manyother things for which we can use these devices. We can measure thefrequency response of our electronics by inserting the pink noise into theinput of a DUT and measuring the output with an RTA. This is a quick way tocheck out our electronics and test our filter sets-do they combine, or arethey 24 independent filters?. Also, are they narrow or wide band, 12 dB or24 dB? Do they cause distortion, or are they impedance sensitive?
The PNG and RTA are also useful for measuring impedance. With, for example,a Gold Line DSP-30 with the impedance program, we can measure the impedanceof loudspeakers or anything else directly. If we do not have a DSP-30, wecan make a black box to work with our RTA (see Figure 3). If we feed theDUT with our PNG and put the black box between the DUT and the RTA, we cansee the impedance of the DUT over the entire audio spectrum on the RTA. Thedevice is calibrated for 4 V, 8 V, 16 V, 600 V and a 70.7 V system. All weneed to do is set the calibrate switch to the impedance we want to measure,insert the pink noise, set the level to any reference point on the screenthat we want, switch to read and look at the response on the RTA (seeFigure 4). Each +-6 dB represents doubling or halving the impedance. Simpleand useful, be sure to take it with you when installing a 70.7 V system; itcan save time and amps.
Another handy gadget to carry in the toolbox is a polarity checker. Withthis simple device, we can measure the polarity between the high-frequencydriver and the low-frequency driver of a loudspeaker, or we can check eachloudspeaker in a 70 V system or cluster to assure ourselves they are all inpolarity. Remember, one out-of-polarity driver can ruin a good system.
Every so often, we find ourselves with a signal generator that has too muchsignal for our input or an output from our DUT that is too high orimproperly matched to our test instrument. This is the time for the in-linedevices made by Shure, Sescom and others .
These devices can do many things. Having the same male or female connectorson each end, they can change gender. They can also reverse polarity. Thehot pin and common are reversed; for instance, pin 2 and pin 3 are reversedon an XLR connector. Lifting grounds is possible; pin three floats on anXLR connector. The device may be a line-level to mic level attenuator, a 10dB, 20 dB, or 40 dB attenuator, allowing it to change level. If we have toreduce our signal by more than what one device can do, we can connect theunits in series. Be sure to observe impedance were applicable. The unit maygo from low to high impedance or visa versa, but impedance changing alwaysattenuates the signal. Usually accomplished with a built-in transformer,the device can also go from unbalanced to balanced or visa versa. Someunits, equipped with battery-powered single-frequency sine-wave generators,can insert a signal into a circuit. Lastly, they can change connector type.If the test equipment uses one type of connector, and the DUT has one ofthe many audio connectors that are out in the field-RCA male or female, 1/4inch plug or jack-we can use an adapter that will correct our interfaceproblem. Some come internally prewired, but just as many come withoutthrough connections so we can wire them to satisfy our needs. Of course, wecan make our own adapters using connectors and wire, but if we calculatethe cost of our labor, we will probably determine it is cheaper to purchasethem rather than try to make them and store them.
Always have an ample supply of these useful gadgets in your toolbox.Whatever the job, whatever the test, there is always more than one way todo it, so never get hung up on using only the latest test gear. Much of theold standby gear of yesteryear will not only do the job, it will do itbetter.