Acoustical Society of America
1996 Science Writers Award Winner

Decibel by Decibel, Reducing the Din to a Very Dull Roar


Richard Wolkomir

Published in February 1996 issue of Smithsonian Magazine
Copyright © 1996 by Richard Wolkomir. All rights reserved
Article will be removed from posting on September 16, 1998 by Richard Wolkomir

Decibel by Decibel, Reducing the Din to a Very Dull Roar

At RH Lyon Corp, noisebusting engineers tackle everything from leaf blowers to ticking clocks in their search for the right sound

The interrogation is under way. The "prisoner"--an upright vacuum cleaner--stands atop a wooden box. Cords hitch the cleaner to the wall so it cannot move. The inquisitor, an engineer skilled at getting household appliances to reveal their secrets, turns on the juice: the immobilized vacuum cleaner can only scream. It is a caterwaul piercing enough to peel your nerve fibers. And that is precisely the point. Via instruments, the engineer is asking the tied-up vacuum cleaner to confess: Which of your parts makes that obnoxious shriek?

In the end, the vacuum cleaner will spill the beans. They always do, the refrigerators and food blenders and washing machines and electric drills sent here to RH Lyon Corp, in Cambridge, Massachusetts. Manufacturers of these sonically wayward appliances turn them over to the Lyon engineers for acoustic tough love.

Richard H. Lyon, the company's founder, is an MIT professor emeritus. He is the holder of six patents, past president of the Acoustical Society of America and the author of four engineering tomes on acoustics. But in this nondescript building tucked among outermost Cambridge's malls and burgerterias, Lyon vivisects appliances. He and his ten full-time and part- time noisebusters disembowel clothes driers and pull the wires out of leaf blowers. Their business is helping companies like Ford, Raytheon, General Motors, Nissan, General Electric and Xerox to get their products to pipe down. Few manufacturers can do it themselves.

Even megacorporations rarely employ their own acoustics experts, and as one Lyon acoustical engineer puts it, "Decibels just don't mean much to washing machine engineers." Also, trying to trace a clang or a pop-pop or a wheeze to its source can make you crazy.

Richard Lyon, soft-spoken and silver-haired, cites the relatively simple case of the rattly gears. Shortly after he founded the little firm in his suburban basement in 1976, the Singer Company sent him a sewing machine: its gears rattled, but engineers could find nothing wrong with them. Lyon finally traced the noise to the sewing machine's needle. Every time the needle stabbed the fabric, it transmitted vibrations through its shaft, down a pulley belt and out another shaft to the gears turning the bobbin. The solution was to reposition the sewing machine's motor closer to the bobbin so that its mass damped the vibration.

Lyon engineers still exorcise rattles and screeches, but now they also frequently work on sound quality. Too much noise remains an issue, but they also want to know if a product is making the right noise. As Richard Lyon puts it: "Noise is as much psychology as physics."

Car buyers, Lyon points out, are apt to spurn a car if its doors shut with a tinny clank. They will favor a model with doors that solidly thud. "Sound is information," he explains. Consider the clunking sound made by some clothes washers as they open and close valves to let water in or out. Those clunks can be annoying, and some manufacturers have, in fact, eliminated them. But it turns out that the clunks tell users when the machine is changing cycles. "My wife's new washing machine doesn't clunk, so they had to add a buzzer to tell you when the cycle changes," explains Lyon.

Thus, fiddling with product sounds inevitably leads engineers into "psychoacoustics," which is the interplay of sound and mind. "A lawn mower's racket is annoying, but a skinhead motorcyclist coming down my street, making exactly the same level of noise, is much more annoying," Lyon remarks. He cites a study of Los Angeles freeway noise: "As you got farther away from the freeway, annoyance went up!" That was because neighborhoods farther from the freeway were wealthier. "People living next to the freeway had the benefit of lower housing prices, and they had to cope with many troubles besides traffic noise," he notes. "Farther back, in the luxury neighborhoods, even though the noise was muted, it seemed much more intrusive and irritating."

Of course, sound is physical too. When something vibrates, like an elm branch a woodpecker is jackhammering, it sends matching vibrations rippling through the atmosphere, the air molecules alternately rarifying and bunching. And those physical pulses can have physical consequences. The Lyon engineers once were called to an electric power plant where earsplitting vibrations from steam pipes heated to 900 degrees F were so intense that parts came flying off.

Sounds can deafen. OSHA, the Occupational Safety and Health Administration, has limited workers' exposures to no more than eight hours per day at sound levels of 90 decibels, four hours at 95 decibels and two hours at 100 decibels, with special limits for quick, sharp "impulsive" sounds, like explosions. Some RH Lyon noisebusting is pure decibel reduction to meet the safety standards. That can be difficult. But the company's new focus on tuning gadgets until they sound pleasing to consumers can be even tougher. For one thing, customers send out contradictory signals.

"People want quiet appliances, but at the same time they equate noise with power," says Lyon. "How do you make a quiet food processor that sounds powerful enough to handle the job?" And successfully quieting a noise, like a car's engine racket, can unmask a most unwelcome panoply of previously unheard noises, such as the heater-fan's incessant drone. "Recently I replaced my computer's hard drive after living for years with the tremendous racket it made, and now I've noticed that my wall clock has an exceedingly loud tick," comments RH Lyon engineer David Bowen, glumly. "I'm experimenting with rubber mounts and damping for the clock."

Lyon's first sound-quality customer, in the 1970s, was the Singer Company. Singer had lined up sewing machines, its own and competitors', so a sales vice president could listen to each in turn. "I want all our machines to sound like that one," he pronounced, pointing to one model.

"The result was total consternation," recalls Richard Lyon. "They didn't know what part of the sound he liked--so they put the magic machine in a closet." That machine became the standard for the ultimate in a sought-after sound. Every so often, engineers would have another try at fine-tuning their own sewing machines' sounds. "They would run the magic machine, but, of course, it became noisier with time because the oil had drained from the bearings," Lyon says.

One reason it is difficult to get sounds just right is that the overall sound may be a blend of sounds contributed by a machine's various parts. And even the sound of a single component may be a medley of sounds from its own subcomponents. Also, small transient frequencies at the initiation of a sound often provide its "personality." "Those initial transients are really the difference we hear between a piano and an oboe," explains Lyon. "Or make voiceprints of yourself saying 'bah' and 'dah' and you'll see almost no difference between the 'b' and the 'd' sound; the real difference is the position of your mouth in each case as you start the 'ah' sound."

In the early 1970s, the Environmental Protection Agency and OSHA campaigned to get the decibels down. Noise pollution was news. But enforcement-hampering budget cuts during the 1980s deep-sixed the issue. Today concern about noise is again ratcheting upward but not because of new regulations. One reason is that the "soundscape" seems noisier. Richard Lyon knows of no scientific study of noise-level changes, but he points out that, for one thing, there are more of us. And some of us do particularly noisy things, like amping up our car stereos so high that people a block away feel the rumble, or demufflering our Harleys. And we all have more motorized tools and appliances and toys to contribute to the general din.

For instance, suburbs nationwide now experience nearly continual lawn-mower snarl, backed by the buzz of weed whackers. Leaf-blower roar can be so unbearable that many towns have mounted ban-the-blower campaigns. Dirt bikes, snowmobiles, all-terrain vehicles and chain saws make the notion of a quiet day in the country as quaint as a Currier and Ives scene. America's lakes and rivers now throb with speedboat whine and the screech of water-scooters and other motorized craft. And while our automobiles are quieter inside, many are louder outside. One reason, offers Lyon, is the shift from eight- and six-cylinder engines to noisier four-cylinder ones. Also, today's radial tires have stiffer treads, which make them hum.

For the sound sensitive, there is a hopeful note: industry is increasingly concerned about noise. The reason is international competition. "Europeans and East Asians are much more sensitive to noise than Americans, and so their products tend to be quieter," Lyon observes. He cites the U.S. maker of a whirlpool pump that shipped its latest model to its British branch. The British office said it could never sell the pump in Europe because it was too noisy. Lyon then helped the manufacturer quiet the pump for the European market. Prodded by such overseas competition, many U.S. corporations are trying to tone down their products' annoying squeaks and pocketa-pocketa-pocketas.

"Who wants to be awakened by your bread maker starting at 3 in the morning, going whir-whir as it makes a turn, waits for the dough to settle back down on the hook, and then goes whir-whir again?" Lyon asks.

To cure such whir-whir problems, RH Lyon engineers usually start with a "sound audit." The aim is to determine what noises each component contributes to a product's overall sound. "We've taken away this vacuum cleaner's suction fan," says Peter Hammer, a biomedical engineer who works part-time here, interrogating appliances. Right now he is "questioning" an upright vacuum cleaner. Hammer explains that he separately records and analyzes the sounds produced by each of the machine's moving parts. He is about to listen to the motor. To do that, he disconnects other parts, such as the suction fan. But that creates complexities. "Without the suction fan attached, the motor would speed up and change its sound," he points out. To compensate, he has hooked the vacuum cleaner motor to another motor under the wooden podium upon which the vacuum stands. The second motor runs backward, pulling against the cleaner's motor. He can vary the pull to simulate the suction fan's drag, whether the vacuum is roaring over a carpet or running its furniture-cleaning attachment.

Hammer's workshop is filled with intriguing and arcane clutter. Cottony material lies heaped in a corner. Plywood panels lean against the walls, which are hung with mats. The aim is to randomly mix the sounds coming from the vacuum cleaner, neutralizing echoes and reverberations to get a clear reading.

Hammer has attached a microphone to the top of a camera tripod and pointed it at the vacuum cleaner. "It's approximately where the ear of a person operating the vacuum cleaner would be," he says. He slowly turns up the juice. The vacuum whines like a Boeing 737 warming up (loud), taxiing (even louder) and taking off (you put your fingers in your ears). On a computer monitor, Hammer eyes a jagged graph line, a snapshot of the sounds emanating from the vacuum cleaner at any given moment. With the motor spinning 400 times per second, one peak stands at 80 decibels. "The next phase," Hammer notes, "will be to break down the overall sound produced by this vacuum cleaner's motor into its parts, like the sound contributed by its brushes or its cooling fans--even components have components."

Once the engineers have graphed the sounds produced by the vacuum cleaner at various operating speeds, and broken those sounds down into the contributions from each component, and broken those sounds down, in turn, into the contributions from each component's parts, they feed the data into a computer. Right now acoustic engineer David Bowen is summoning up on his office monitor the data accumulated thus far from the vacuum cleaner.

"This is the motor's sound," Bowen indicates, producing a graph on the screen. Simultaneously, the computer's speakers issue a hissing roar, the noise in question. Bowen fingers the keyboard, and the screen's graph shifts, showing the noise made by air flowing through the vacuum. The speakers emit waterfall noises. Again he changes the graph, and now the noise sounds almost exactly like foghorns. "That's the sound of the agitator on the carpet," he explains. Finally, the computer emits an eerie, high-pitched hum, the noise of the vacuum's main suction fan.

Like a pianist at a keyboard, Bowen can play with the sounds, electronically altering them to see what would happen if he changed a certain part on the vacuum cleaner in a certain way. RH Lyon may arrange for a palette of such sounds to be judged by a carefully selected "jury" of homeowners who use vacuum cleaners.

"The jury would hear the vacuum's baseline sound," says Bowen, resummoning the hissing roar. "Next, we might raise the airflow sound a couple of stages and raise the motor sound." He punches keys, and the noise from the speakers changes to a Bronx cheer. It shifts to a starship going into warp drive. The jury would evaluate the various sounds. "We're trying to get a combination of higher perceived power but also a more acceptable sound," adds Bowen. The vacuum cleaner's manufacturer will receive a report detailing suggested design changes, such as altered blades on the suction fan or a reconfiguration of a nozzle.

Sometimes, despite the engineers' little victories--a quieted juicer, a muted hedge trimmer--the war against noisiness seems to falter. Richard Lyon cites refrigerators. When he was a child in Evansville, Indiana, his father, a high school dropout, worked as a self-taught analytical chemist for the Servel Refrigerator company. Its gas-powered refrigerators had no moving parts. "You can't get any quieter than that," states Lyon. Such silent refrigerators now are rarities. Quietness has been sacrificed to a quest for more interior shelf space.

In a refrigerator, a liquid--usually Freon--is pumped through the appliance, where it evaporates into a gas as it absorbs heat from the refrigerator's insides. The gas is then compressed and forced through a big exterior coil, where it relinquishes its collected heat and becomes a liquid. As it reenters the refrigerator, it vaporizes again into a gas, beginning another heat- extracting run. In older models, the condensing coil is behind the refrigerator; in newer models, to make room for deeper shelves, the coils are underneath. But down there, insufficient air moves past the coil to cool it by convection. "So they added fans," says Lyon. As a result, refrigerators are noisier.

But refrigerators are about to get even louder, Lyon predicts. That is because Freon-today's standard refrigerant--is accused of harming the ozone layer and will be phased out. Whatever replaces Freon will probably need higher pressures for liquefaction. That extra effort will mean more decibels.

Lyon is a veteran of these "win some, lose some" racket wars. In 1946, when he was in high school, he took a Lee De Forest radio-repair course that hooked him on acoustics and ultimately led to a PhD in physics from MIT. He became vice president of a Massachusetts sound-and- vibration consulting firm, working on such projects as quieting nuclear submarines and hushing the Saturn V's rocket-shaking lift-off roar. In 1970 he returned to MIT as a professor of mechanical engineering. Physicists had by then ceded acoustics to engineers, believing all basic discoveries had been made. On the side, Lyon developed his small acoustics company. And in June 1995 he retired from MIT to devote his full attention to RH Lyon Corp and the war against electric-carving-knife buzz and can-opener whir. For appliance vivisectionists, these are busy times.

"This makes too much noise, unpleasant noise," announces one RH Lyon engineer, Richard Cann, in the inflections of his native London. He is holding up a weed whacker. From the same bench he picks up a partially dissected leaf blower. "This fellow is the most unpleasant. It's now banned in about 150 communities," he continues, contemplating the eviscerated blower as if it were a dead adder. "It has to be lightweight to ride on your back, which leaves little room for noise-control materials," says Cann, adding that the gasoline-powered tool--essentially a vacuum cleaner in reverse--blows out a 250-mile-an-hour gale. "You've got the wind whooshing through a nozzle, plus a noisy fan, and manufacturers worry their products will not be usable."

Cann, a Cambridge University-educated engineer, has been a part-time RH Lyon noise detective since 1977. He has now turned his attention to a belt sander, which he has under sonic surveillance. The power tool hangs from strings, suspended just above the floor. "That's where you might hold it to give it a try at the store," he explains, adding that he also will probe the sander's sounds when it is actually sanding a board. "The manufacturer wants to know why it's so noisy, and what changes they should make on their new model." Inside the tool, Cann notes, a high-speed motor drives a belt, which turns a speed-reduction gear. The gear then turns the sanding belt. Because the motor generates heat, the sander's metal casing is slotted for ventilation. An internal fan draws in cooling air and blows it out, while a second fan blows sawdust through a channel inside the sander and out a slot at the back into an attached bag.

All three processes--belt driving, motor cooling, dust extraction--are driven by the same motor; all make noise. Cann must identify which processes are making which noises, pinpoint objectionable sounds and decide what to do about them. That means sleuthing. "For instance, look at this belt that runs inside the sander," Cann suggests. The belt is ridged to engage cogs on the pulley that turns it. "Each time a ridge of the belt fits between cogs on the pulley, a little puff of air squishes out, creating a noise. The pulley has ten teeth, so the belt is making ten air puffs every time the motor turns, creating an amazing amount of noise." Meanwhile, the sander's fan blades produce whirs and buzzings. "You also get turbulence as the fans move the air--and with the grating of the sander against wood, altogether it sounds like a small jet taking off," says Cann.

One key question is how the sander radiates its sounds. Do all the sounds come from the vibrating of the sander's metal shell? Or do some noises emanate from inside, through the sander's ventilation slots? To find out, Cann has puttied up all the ventilation slots to see if that makes a difference. But, as usual, the test creates its own complexity. "We have to test the surface vibration before and after puttying up the vents to be sure that the putty itself is not changing how the shell of the sander vibrates," Cann points out. "Our other concern is that, with the vents puttied up, will it melt down? So we have to do our tests quickly."

That is what Cann now does, aiming a microphone at both unputtied and puttied sanders and letting them roar. The sounds turn up as frequency-graph squiggles on a nearby attached computer, with spikes indicating particularly loud noises. "That one's about 100 decibels for the user," says Cann, pointing. Because he knows the specifications of the sander's various moving parts, he can mathematically deduce which parts are making which spikes. By comparing the puttied and unputtied sanders, he can determine which sounds waft from the ventilation slots. And by puttying only some slots, he can begin to pinpoint which parts make which noises. "Look at the graphs," he says. "The air holes are a major part of the sander's problem, and you can't fix it without dealing with that--you can see that with the puttied sander we've reduced the noise by 10 decibels."

It is painstaking work. But backers of an experimental new technology called active noise control (ANC) promise to make silence easy. The idea is to wipe out a noise by generating a diametrically opposed sound: where the unwanted sound's time graph has a peak, the anti-sound has a valley, and vice versa. When the opposed sound waves intersect, they cancel each other out and all is silence. At least, it is in theory. In practice, ANC would require equipping an offending machine--let's say a noisy air conditioner--with microphones to pick up the unwanted noise, a computer chip to analyze it and design the anti-noise, and speakers to send out the anti-noise on its silencing mission.

"It's extremely oversold," states Richard Lyon. Most noise, he points out, is not a single sound but a multi-source cacophony. He doubts it will ever be practical to analyze changing sounds from several sources and, in a twinkling, to generate canceling counternoises. He says ANC works best where noise funnels through a defined point. He cites protective earphones for factory workers. Also, he mentions that some companies are working on ANC car mufflers. But a stumbling block has been getting the ANC equipment to survive more than a few months in a car's harsh underworld.

RH Lyon engineers deal with car noise the old-fashioned way. One company, for instance, asked Lyon to determine how much noise was emanating from beneath an American luxury car's chassis. Certain European luxury cars had underbody shields--did that make them quieter? The Lyon engineers fabricated an underbody shield to test the theory. They filled the car with instruments and an "acoustic head," which is a dummy head with microphones where a human's ears would be, so that it "hears" what a person in the same location would hear. Then they repeatedly drove the car up a New Hampshire mountainside, turned off the engine and coasted back down at 70 mph, with the electronic ears pricked up. "Did you ever try to drive a big luxury auto with the power steering turned off?" asks Richard Cann, remembering the project's rigors. It did lead to a conclusive finding: no noise came from below. No shield was in fact necessary.

Sometimes, instead of trying to eliminate noise, the acoustic engineers exploit it. For General Electric, for example, RH Lyon developed a diagnostic system that senses surface vibrations on a diesel locomotive's engine block. The data--properly processed--reveal combustion pressures inside each of the diesel's 16 cylinders. It is somewhat like a physician checking a patient's health with a stethoscope. Occasionally, the engineers actually become noisemakers. For RCA, they developed a new home theater system in which the speaker sitting atop the video screen acts as a ventriloquist. Sounds emanating from the speaker are manipulated so that they seem to come from the mouths of the actors on the screen below.

Usually, however, the engineers' goal is silence. There was the Malaysian surgical glove factory, for instance. To make the gloves, workers repeatedly dipped porcelain models of a human hand into liquid latex, which then dried into gloves. Air jets then blew the new gloves off the porcelain hands. But that meant 250 gloves blown off 250 porcelain hands every 30 seconds. The sound was literally deafening--125 decibels. RH Lyon engineers used a specialized camera to study the problem: they photographed shock waves moving through the air as the air jets blew the gloves off. They discovered that altering the air jets' nozzles would reduce the noise by 20 decibels, enough to allow those workers equipped with ear protectors to safely stand nearby.

RH Lyon's case list sounds Sherlockian. What mysteriously vibrated the Japanese luxury car? And why did that new windshield-wiper motor hum? There was the clunking elevator in the millionaires' condominium building. The Smithsonian Astrophysical Observatory x-ray telescope that shivered. The portable generators for pleasure boats that whined in the night. "There was a noisy computer hard disk, too--the solution turned out to be drilling holes in the right places," remembers Richard Cann.

Such work can lead to sensitive ears. Cann has discovered that by repositioning a newspaper that he is holding up, he can make his wristwatch seem to tick louder or softer. And not long ago, at lunch in a noisy Cambridge cafe David Bowen suddenly looked distracted. A new roar, muted by distance, had added its note to the din. Abruptly Bowen brightened: "Dishwasher!" he announced triumphantly.

Richard Lyon, like most people, dislikes screeching dental drills. Old-fashioned dental drills, he notes, were driven by a cable at low speed. Modern drills are air-powered, and they produce an unnerving high-pitched whir. "Everybody assumes it's from the air, but that's producing sounds above the range of human hearing," explains Lyon. "What you actually hear is the vibration--I figured out how to fix it, but it turned out there already was a Japanese patent."

After all these years as an acoustic sleuth, Lyon has pretty much heard it all--the clunks, wheezes, screeches, buzzes, bangs, roars, shrieks and pocketa-pocketas. A man gets jaded.

Lyon shrugs, almost apologetically. "I'm probably more tolerant of noise than most people," he says. Return to 1996 Science Writers Award Winners