Confessions of a Former Driver Designer
By Perry Sink Marshall
The audio construction hobby is driven by the desire to delve into the nitty grittys of audio systems, and control those subtle, underlying details.
If you build amplifiers, you probably appreciate the opportunity to design the whole thing from scratch. Most folks who go to the trouble of building an amp aren’t going to just buy a high power IC and slap it on a heat sink. You select every component in order to maximize performance at every level.
Building speakers is a bit different, though. Most of us don’t have the luxury of building drivers from the ground up. The closest we have to that is simply being able to choose from many, many drivers, hopefully being able to pick one that matches our priorities very closely, then having complete control over every other detail – enclosures, crossovers, etc.
I had the opportunity to get down to the bottom, however – designing drivers at Jensen in Chicago, from 1993-1995. I’m going to share some things I learned, but first a little background.
They Got Me At An Early Age
I was 12 years old when I acquired what would prove to be a “permanent” urge to have a better stereo system. My financial resources, based on allowances, odd jobs and occasional stints as a substitute paperboy, were extremely thin, and visiting all the stereo shops in Lincoln, Nebraska, the prospects looked a little bleak for anything beyond the most basic system.
But Radio Shack had a book on building speakers, which I bought, and then I also bumped into a guy at one of these shops, Steve Peters, who was apparently willing to give up a sale (my eye was on his Boston Acoustics A60’s) in order to teach me the ropes of building my own system.
80 bucks later, I had a Peerless woofer and Tonegen tweeter from the McGee Radio Catalog, some crossover parts, and a laborious job of cutting particleboard with a jigsaw.
Those speakers actually sounded fairly decent, and I was hooked. I built another pair – and sold them – and continued through high school and college, picking up extra bucks building speakers for friends.
When I was a senior in high school, I secured an arrangement with a local dealer, selling my brand in his store. A year later I was studying Electrical Engineering at the University of Nebraska, and in ’92 I left with my EE degree, with a motional feedback subwoofer as my senior project and a focus on control systems and communications.
Jensen, then in Schiller Park, Illinois, offered me a job and I moved to Chicago. I started working in the OEM division, which services the automotives.
During my time at Jensen, I was involved in designs for the 94 Ford Probe, the 95 Jeep Cherokee, the 95 Acura Vigor and Integra, the 96 Honda Civic, and the 97 Chrysler Cirrus and Stratus. I also worked on a “skunkworks” project to develop a low cost motional feedback subwoofer. (I left before that project saw the light of day, but it was fairly successful, and if someone out there wants to pursue it, it’s still lurking in my files!)
OEM vs. Aftermarket
There were two divisions of Jensen’s engineering department: OEM, which built speakers for new cars; and Aftermarket, which was responsible for the products that got sold at car stereo dealers, WalMart and everyone else. You know, Jensen 6×9’s with blue surround and all that stuff.
When I was interviewing, they asked me which department I was interested in, because they had one opening in each.
I really didn’t know. Somehow I wound up in OEM, and in retrospect, I’m glad I did.
Here’s why: In aftermarket, some product manager, who’s previous job was in the marketing department at Whirlpool or some such place, gives you a pair of speakers from Sony or Pioneer and says, “Make our next model sound just like these, because they’re outselling ours. And they have to look like this drawing right here. And you have to do it with these parts.” Blah, blah blah.
Basically, in aftermarket, I would have had both hands tied behind my back. Guys who worked in aftermarket usually had voodoo dolls of marketing people full of pins, and they tended to drink a lot. They had clippings on their bulletin board where Dilbert says, “We encourage a healthy tension between engineering and marketing.”
They did like going to car stereo shows, though. They could sit in a conversion van with fourteen subwoofers and three thousand watts of power, watching girls in skimpy bikinis and feeling their pantlegs flap in the bass.
The Hidden Opportunity In Designing Cheap Car Speakers
Ironically, OEM was almost the opposite. You’d think that Honda or Chrysler would have given us a merciless straightjacket of specifications (actually, Ford did), but mainly it just had to be cheap.
What exactly does “Cheap” mean? Shaving pennies to the tune of $3.75 per speaker, and 0.01 cents precision on pricing. Yes, it had to fit in a particular mounting hole, and it had to have a minimum 88.5dB sensitivity or something like that, but especially on the “basic” systems (as opposed to the premium systems), I had a free hand.
My mission: To make four $3.75 speakers sound good in a Honda Civic, with pre-defined mounting holes and locations, and whatever radio Fujitsu Ten sells them.
To a die-hard audiophile purist, a pretty dreary proposition. But to a guy like me who loves a challenge, however, who looks at things in terms of ‘doing the most with what you’ve got to work with,’ being proud if the results come out sounding “surprisingly good,” then it’s a lot of fun.
The way I figure, any idiot should be able to come up with a nice sounding system if he’s got five thousand bucks and Krell amplifiers. But the mark of a skilled engineer is in finding combinations of materials, budgets and design concepts that deliver excellent results with scarce resources—i.e., five bucks instead of five thousand.
Most of these designs were based on full-range speakers, not 2 ways or 3 ways. I was intrigued by that, too. When I was 17, I got my hands on E.J. Jordan’s Manual on driver design, where he explains the rationale behind his 50mm module, a driver which covers the range of 100 Hz to 20 kHz. This manual introduces clear thinking and rationale for full range driver design, and provided great background for me.
When I showed up for work on February 1, 1993, I had a lot of theoretical ideas about how drivers work, or should work. I was about to put all my theories to the test of real measurements and practical economics.
Driver design is a bit mystical in the eye of the hobbyist, but there’s not much that’s really “mystical” to me, after building hundreds of drivers from the back plate on up, measuring them and listening to them.
Not that I know everything, by any means. But the reality is, real drivers are made of real materials, which have known, measurable properties, and there’s a reason for everything you hear, whether real or imagined.
Parts is Parts
At Jensen we had a sample parts room – pegged walls and shelves of back plates, front plates, pole pieces, magnets, baskets, cones, spiders, voice coils, dustcaps, whizzer cones, terminals, lead wire, dozens of versions and variations of each different part.
In the lab we had a dozen different kinds of adhesive, and a staking machine for securing front plates to housings, etc.
You could walk into the parts room, grab a handful of parts, and have a working speaker in 30 minutes. Just down the hall was an anechoic chamber and a chart recorder, where you could plot frequency response, impedance, and distortion. Another room had an automated Thiele/Small parameter measurement setup.
Then of course, you could take your new speaker, attach it to one of our handy-dandy, pre-cut templates, go into the listening room and mount the template in the wall, and listen to your new speaker. You could match levels with an SPL meter and compare it to other speakers, and do so behind an acoustically transparent curtain if you so desired. Can’t leave out the part about listening!
You can learn a lot about speaker design, making a dozen variations of a single design and measuring them all. And I did.
One of the most basic observations I made was that even though one might develop a “mystical” perception about a certain name brand, design preference or material, when you get right down to it, it’s all physics.
Voice coils have very specific mechanical and electrical parameters, and characteristic quality control characteristics. Cones come in a variety of profiles, material combinations, and mating surround materials and geometries. You can wax eloquent about the musical qualities and everything else, but it all comes from very tangible mechanical device, which responds to electrical input in accordance to well known physical laws.
Inside The OEM Speaker Business
Inside the speaker biz, speaker components come in two categories: Software (cones, spiders, voice coils) and Hardware (magnets, baskets, back plates).
Most of my work was with mass produced car speakers (i.e. 100,000 to 500,000 units per year was typical), especially 4-6” round and 6×9’s. Cost was always an issue, as was tooling. As a designer, I was strongly discouraged from “tooling” a new part. Expensive, time consuming, and prone to failure.
For your reference, the table below shows the various parts of a typical small, large volume production speaker, a rough estimate of their costs, and also of the tooling costs to create each part:
|Part||Per Piece Cost||Tooling Cost (Dollars)|
As you can see, the total “Bill Of Materials” cost for a cheap, high volume speaker like this is $2-3. Most parts could be used “off the shelf” without additional tooling, so add in the labor to put the speaker together (at Jensen it was 5 cents per five second operation), and you can sell them for $4 to $7. When I was working there, Jensen made 50,000-70,000 speakers per day, mostly with unskilled workers and relatively little automation.
How good were these speakers? Some were not too bad, in my humble opinion. We made full-range 6×9’s with whizzer cones that were +/-3dB from 55Hz to 17KHz on axis. They sold for about six bucks each. If you take the stock speakers out of your Ford Escort and measure them, you might be surprised at what you see.
That’s an interesting subject. You can take a 5” speaker that’s got a smooth response from 80 Hz to 6, KHz and replace the dustcap with a whizzer cone, and now it goes out to 15 or 20KHz.
I’s almost as good as adding a tweeter, but not quite. Let me explain. A whizzer attaches directly to the voice coil, and because of its smaller size and (usually) higher density, it responds to frequencies that the woofer can’t cover.
Whizzers are highly resonant, with most of the damping, I believe, coming from the air itself. Whereas tweeters can often have a smooth response curve, whizzers always have a bit (often a lot) of jaggedness, and tend to sound “phasey.” You just don’t get the clarity you do with a tweeter. Also, the polar distribution looks like a bunch of bananas. But it’s much better than no whizzer at all, and the cost increase is perhaps only ten or fifteen cents.
To my knowledge, nobody has successfully “modeled” the behavior of a whizzer cone. Everyone I know just fiddles with a design, with various sizes and shapes, until something seems to work.
Make a Silk Purse Out of a Sow’s Ear?
I always thought that someone should take a pair of those six dollar 6×9’s, which were really more accurate than most people might guess, put them in a set of beautiful, heavy, rosewood, wool-filled transmission line cabinets with elegant cloth grilles, and market them as minimalist, exotic high-end speakers for $3500 a pair. No crossovers, or any of their associated evils! Imagine the harmonic integrity and all that jazz, however you want to phrase it.
Match them up with a single ended triode amp and some really expensive cables, and somebody would be in audio purist heaven.
(Well, there’s a business idea for somebody! I promise you it has an excellent chance of succeeding, too.)
A Day In The Life of an Acoustical Engineer
There was certainly fun stuff, like brainstorming concepts and listening to test systems in vehicles. But the majority of my time was spent building prototypes, batches of samples, and doing lots of testing and statistical analysis.
One of the benefits of working at a high volume OEM is that everybody’s happy to send you samples. You call them and explain that your company makes 70,000 speakers a day and they pay attention.
Actually, especially with the cone, spider and voice coil manufacturers, they know who you are, and it’s perfectly normal to call up their samples guy every other week and rattle off some stuff you need, and get some custom made parts within a week or so, possibly with a whole range of variables in the sample lot.
(One of the mechanical engineers was fired when someone figured out that the samples he was requesting from a machine shop were actually engine parts for his motorcycle.)
Why Acoustical Engineers Who Design Cheap Car Speakers Bash High End Audio Magazines
Most of the guys I worked with had the same geeky attraction to speakers that I did, so we spent plenty of time discussing audio theories, lore, and strange inventions, and of course, bashing Stereophile and Absolute Sound.
As I said before, there’s something about designing drivers that teaches you ‘harsh reality.’ In the end, there is no mysticism. There’s only physics, and the attempt to relate it to the subjective experience of listening to music. There are things that make a lot of difference, and things that make only a little difference.
Here’s my quick-and-dirty list of things you can change in your system to make it sound different, and hopefully better, in relative order of how much those things will actually influence the sound:
- Your listening room
- Your crossover design and cabinetry
- Your drivers
- How much clean power you have available
- Inductors (replacing ones that saturate with ones that don’t)
- Signal sources, amps, preamps, D/As, etc.
- Speaker stands
One of the things that became clear to me as I worked on audio systems was that items on the top of the list are difficult to change, require genuine insight into problems that most people haven’t studied, and are difficult to “sell” in the sense of swiping somebody’s American Express card and providing immediate gratification (with the notable exception of simply buying new speakers off the showroom floor).
However, the items on the bottom of the list are easy to swap and have the highest profit margins, even though they hardly make any difference at all. So those are the things that get the most attention in the press. Somebody will always be happy to lighten your wallet with something that makes more of an imaginary than real difference, simply because it’s profitable.
So when our engineers (many of whom own excellent home audio systems) would go to a show like CES, we would spend most of our time at the high-end portion, thoroughly appreciating the guys that were making real progress, doing real research, and certainly taking notice of the best-sounding rooms. And we also didn’t miss any opportunities to make light of voodoo and magic tricks!
You may have never thought about this before, but nearly every part in a speaker is secured by adhesives. The only typical exception is the attachment of the basket to the front plate. This is typically “staked” but even then, it is still damped with a bead of glue between the two parts.
Speakers endure horrendous stresses, and it’s pretty impressive that they stay together as well as they do. I can assure you that a great deal of time and attention is invested in this vital design ingredient. Jensen had a full time engineer devoted to this purpose, and in the time I was there, he reduced the total number of adhesives used in Jensen’s manufacturing from 27 to only 5. Very impressive, considering the complexity and diversity of requirements.
Here is a brief list of constraints that a designer must deal with when choosing an adhesive:
- How fast does it cure? Adhesives made with volatile, nasty chemicals cure very fast and form very strong bonds, but the fumes might erase your DNA. Water is safe but cures much more slowly. By the way, US environmental laws limit the use of the nastiest adhesive chemicals and therefore eliminate some otherwise attractive choices at the outset. For that reason, foreign-made speakers may have performance advantages, like higher power handling.
- Is it waterproof? Do ultraviolet rays break it down?
- Is it brittle or does it have problems as it ages?
- What temperature range can it withstand?
- What materials will it stick to (and not stick to)?
- Will it damage some materials as it cures?
- Are there chemicals that accelerate the curing process?
- Does it have to cure in an oven?
- Does it leak and drip into tiny spaces?
In general, the following are somewhat typical of the kinds of adhesives, which are used for various surfaces:
Surround to Basket, Spider to basket: Water or solvent – based glue, somewhat resembling “elmers glue” in consistency.
Magnet to Metal: Vile smelling, solvent-based adhesives, with two parts: The adhesive and the accelerating agent. The glue only cures when the agent is mixed with it.
Cone to Surround: Varies widely according to the types of materials used. Hot melt is not uncommon.
Voice coil to cone: This one’s critical. Typically cyanoacrilate (“Superglue” like Loctite) or else 2-part Epoxy. Epoxy is great, but harder to use. “Superglue” is easy to use and cures quickly with accelerator, but has problems at high temperatures.
Plastic cones, especially polypropylene, typically don’t bond well with much of anything, unless you use special techniques. It’s possible to peel the cone away from its adhesives with not much difficulty with most polypropylene woofers. Poly introduces power handling problems. Jensen developed a special technique that worked really great, but if I tell you about it, I’ll have to kill you.
By the way: Guys with PhD’s in Chemistry just hate it when you call it Glue. Epoxy, Adhesive, Hot Melt, whatever you want to say. But please don’t denigrate the profession by just referring to it as glue.
My Personal Invitation to a Special “Meeting”
During the summer of ‘95, orders were slow, production was down, and rumors of “downsizing” were floating around. One day I got a tap on the shoulder and I followed someone into the conference room.
An odd assortment of people from various departments were collected there, and Herman, the division manager, was pacing nervously and breathing heavily. He was sort of wheezing and he looked real uptight.
Herman managed to communicate to us that the company was experiencing a slowdown and they had to turn some people loose, so… best wishes to all of us, and someone would be discussing terms of severance with us, and all that fun stuff.
I had discussed this possibility with my fellow engineer Giles Davis just a day or two before, and I figured that this was as good a chance as any to go experience something else besides building speakers.
So I left the conference room. Somehow, everybody had found out in the meanwhile what was happening to all of us. They were all wierded out, with sort of a somber, deer-in-the-headlights, “gee-I-still-think-you’re-a-good-guy and I hope the unemployment line ain’t too long for ya and oh yeah, what if this were me” look.
I met with this VP who pensively explained my severance plan. 7 weeks pay, turns out – a great vacation if you don’t have to find another job! I said, “You know, on the day I was born, it’s not like some other guy was born on the same day with the duty of giving me a job. It’s been great working here, and thanks for the opportunity. Sorry for the circumstances, but I’m sure things will work out fine.”
And an hour later I had cleaned out my desk, and I was driving home at 11:00 in the morning, ready for my next adventure.
Acoustical Engineer Gets Sex-Change Operation
I hit the ground running. And it couldn’t have been more than three hours before the recruiters started calling. Speaker design jobs in Indiana, Michigan, California, Florida… the list of places where I could be living and working in just a few short weeks was plentiful.
But my wife was expecting, and we didn’t want to start over after being in Chicago for 2 1/2 years, so we decided to stick around. I made the rounds with the various audio-related firms in Chicago and came up empty handed.
The next step was to explore sales opportunities. Yes, despite the Hatfields-and-McCoys style of affection that Engineers and Marketers naturally share, I always felt like I could, and should, live with one foot in each world. A good bit different from working in an engineering department, though, no doubt about that!
So I started exploring that venue, and within two weeks I had three job offers. I accepted a position as a manufacturer’s representative, selling industrial automation components. That “controls and communications” specialty from school was just as applicable there as it was to audio systems, actually; just a totally different industry and application of that knowledge.
I still work in the controls industry, and now I’m up to my eyeballs in the nuances of industrial networking. I was once an analog type of guy. Now I guess I must be digital.
Audio Regains its Hobby Status
Speaker Builder is all about doing it for the love, not for money. That’s the very essence of the term Audio Amateur. And I have to admit: Doing speakers every day for a living may be more interesting, in some respects, than doing something more “utilitarian,” but also, doing it every day takes some of the fun out of it.
I also have to admit that ignorance can be bliss; while it’s very rewarding to learn the subtleties and nuances of the art and science of speaker design, romance can also be displaced by reality.
Actually, I think the intrinsic conflict between romance and reality is at the heart of the “high end audio” debate that I alluded to earlier, as well as driving the intrinsic conflicts that occur between engineers, and the sales people who sell what the engineers design.
I’m a member of PSACS (Prairie State Audio Construction Society) in Illinois, which is a motley collection of manic audio constructor hobbyists like you and me. Once the speaker design job was in my rear view mirror, those PSACS meetings started to seem much more interesting, and tweaking those speakers in my living room slid back up a few notches in my priority list.
The knowledge I gained at Jensen was not only intellectually satisfying, but it even earned me the distinguished title of “Official PSACS Speaker Design Wizard”—an honor that I hold with great reverence and pride, you understand.
Post Script to My Jensen Story
About three years after I left Jensen, things were going well there, but as I understand, the company got into a precarious position where it had lots of great OEM designs in the pipeline, but at the end of 1998, only one paying customer. This customer cancelled its current project and Jensen OEM couldn’t keep the lights on until the next project materialized. So they closed their doors in the spring of 1999.
To everyone’s credit, I must say that nearly everyone who worked there when I did, still keeps in remarkably close touch, particularly with e-mail. Much better contact, in fact, than I’ve seen in any other company I’ve worked for before or since. It’s a rowdy bunch of folks, and once in awhile there’s a Jensen reunion at some bar in Chicago, and I can assure you that a good time is had by all!
Jobs In The Audio Industry
I’m sure a few of you are wondering if you could get a job like that—designing speakers or perhaps some other component, for some audio manufacturer. Let me give you a few angles on that subject.
First of all, I should say that getting a speaker design job in a small, “high end” company is rather difficult, though not impossible. Very commonly, the owners of the company are responsible for that job, and though they may get a steady trickle of resumes, they are unlikely to give up that job to an outsider.
Jobs like the one at Jensen – pragmatic design jobs for large, consumer driven companies are much more plentiful and, in my opinion, offer greater technical and educational resources anyway.
Jensen had an extensive R&D department that, in reality, was doing more ground breaking work and real research than 95% of high-end companies which are much more “prestigious” than Jensen. Sounds like an outrageous statement, but it’s really true. You just have to also understand that this R&D was ultimately focused on economic benefits, not philosophical ideals.
Jensen employed lots of mechanical engineers, who often did not have audio backgrounds, as well as electrical engineers who handled the acoustical aspects of projects and were more likely to be “audio geeks.” In fact, I’d say nearly all were audio enthusiasts. An EE degree, or a physics degree, are both taken pretty seriously by those who are hiring for such jobs.
As a final note, the audio world is in the midst of a massive collision with the computer world. This is a hot area with a multitude of possibilities for someone with a background in Computer Science, Computer Systems Engineering, or Digital Signal Processing. MP3 and all that stuff will not only transform the distribution of recordings, they will also change the nature of our audio gear!
So if you want to work in the audio industry, the diploma is a big help, and be sure and subscribe to the Journal of the Audio Engineering Society. It will beef up your technical knowledge. Speaker Builder’s sister publication, Voice Coil, is the loudspeaker industry insider publication, and if you’re at all interested in this, get a subscription. It should be interesting even to amateurs.
And finally: Don’t let that subscription to AudioXpress lapse. You don’t ever want to forget where you came from!