A Ken Kessler interview with Rich Vandersteen explains some of the DNA concept of his designs.
Richard Vandersteen doesn't look like a typical loudspeaker
designer. True, he wears glasses, but his presence suggests a
longshoreman or somebody who'd be played by Gene Hackman. And sure
enough, he tells you in a quasi-Dukes of Hazzard drawl that he's
been a construction worker, plumber, truckdriver, and electrician.
Electronics had always been a hobby, but Vandersteen formalized his
understanding by working in electronics during his stint in the Air
Force. Back in civilian life, Vandersteen entered into speaker
manufacture, producing the "baffleless" range, at least regarding
the midrange driver and tweeter, which bears his name. The speakers,
particularly the Model 2 and its variants, have become, in a decade,
one of America's most respected brands, despite RV's low-profile
marketing techniques. I met with Richard at the Las Vegas CES in
January and asked him what had got him started in loudspeaker
design.
Richard Vandersteen: I came up with this idea for a speaker without
a baffle in 1972, and started building them for people in our local
area. A high-end store opened up nearby in Visalia, California, and
they said "Let's take these things to Chicago." It was 1977. I said
"Where?" because I didn't know of the Consumer Electronics Show. I
took vacation time from my job, met something like 30 dealers and
got 200 orders for about 250 pairs. I went back, told the boss I
gotta quit, gotta find a building, because we were doing it in the
garage and it was too small for that. That's how it started.
Ken Kessler: What inspired you to throw away the baffle for the mid
and treble drive-units?
Vandersteen: In some of the experiments in the early designs, we
were flush-mounting drivers, and I was sealing them in with
silicone. My wife has a favorite record which has a saxophone on
it—it was stolen, I don't remember what it was—but it was one of her
favorite records and I used it a lot for testing. I'd blown a
midrange and had no more new drivers, so to make sure it was still
working without all of the problem of removing the silicone, which
was very difficult to get out, I just laid the midrange up on top.
My wife was around the corner in the kitchen; she said, "Richard,
what did you do to the stereo? It sounds so much more real." I said,
"Actually, it's not right." So I took the other midrange out and
laid it up on top, and then went into the garage and got a stick and
laid the tweeter on it, and that's where we came up with a tube. The
problem was this thing was real ugly; she said, "There's no way
you're going to leave those in the living room."
About three weeks later, as I was hanging up my jacket, I said,
"closet pole." Then my wife sewed up some double-knit fabric which
was in at that time—it's not any more—and made a sock to go around
the loudspeaker because, she said, "you could never sell them that
way."
Of course, three, four, five years later out comes the KEF 105 and
the B&W 801 and they did very well with that modular affair. But
that was the beginning of the Vandersteen range. It's the same
design today; I just keep refining it with better drivers. The
company finally got large enough that we could get custom drivers
made, and that helped a lot. We wind our own capacitors, and on and
on.
Kessler: The striking thing is, of course, the lack of a baffle. But
that flies in the face of tradition. As everyone knows, the baffle
isn't just there as something to hold the driver. The baffle affects
dispersion, it reinforces the sound. How do you account for
something that went completely against traditional lore for box-type
loudspeakers?
Vandersteen: Partly by experiments. If you take a perfect tweeter
and mount it in a baffle, of course you can put it in the baffle
asymmetrically, round the corners and minimize the effect of the
baffle, but the baffle is still there. However, you do have a
problem when you have no baffles in that there is no averaging; with
a baffle, when you strike your first boundary at a
quarter-wavelength, you get a 3dB increase. And this is nice; by
positioning the drivers you can be quite a bit cruder on the
crossover and still get flat power response. So we did have to get
involved in extensive use of compensating networks and phase
networks and so forth to make the drivers work without the
assistance of a baffle, without any averaging going on, to make the
speaker flat.
The crossover technology was the most difficult part of the design.
Exactly as you said, you don't have this averaging effect that comes
from the baffle. In hindsight, this is why the smaller speakers that
were and still are popular in England always seem to be more natural
when you have them out in the room: they in themselves have a lot of
their frequencies unbaffled and . . .
Kessler: . . . behave like point sources?
Vandersteen: Yeah, exactly. The closer you can space that driver
group, the better, the more realistic the speaker will sound in a
room. And that's for two reasons: If you have a large source, even
if it works as a true piston—let's assume 6' tall and 1' wide—you
can, by designing it properly, make the sound hit the ear of the
listener perfectly in phase. But you'd have a problem. The way the
ear/brain works, as long as it has clues as to its environment, it's
very good at separating your living room from the Steinway in that
room. It should do the same with hi-fi. If your source is too large,
you can design that source, with dynamic drivers or planar drivers
or whatever, to hit the listener properly, but what gets all balled
up are the phase differences and the time delay in all the room
reflections. Even if you damp the room, no material is 100%
effective in damping, and no material I've found is totally linear
in its absorption characteristics. You end up with false clues.
For instance, if you put a squarewave into a Quad '63 (which behaves
in a manner akin to a point source), you can get somewhat of a
squarewave out of that speaker at the listening position. You could
turn the microphone to the rear wall, side walls, up to the ceiling,
anywhere, and all of the reflections within that room will be pretty
close in shape to the original pulse that came directly from the
speaker. You cannot do that with a multi-unit dynamic loudspeaker.
You may have the individual drivers focused with delay and
everything precisely on the listener so that you can make the
squarewave hit the listener, but when you start looking at the
reflections in the room, it's hopelessly messed up. There will be no
recognition of the original pulse at all.
So I looked at perfecting that idea too, making sure that group was
close, and that it loaded the room in a realistic way, so that all
of that reflective energy would come to the listener in phase, or in
sync.
Kessler: But you mentioned earlier that the biggest problem with
baffleless driver mounting was compensating for phase and so on via
the crossover. In taking care of the benefits of having a baffle
which you sacrifice by not having a baffle, didn't you just trade
one set of problems for another? In other words, you sacrificed a
mechanical cure and replaced it with an electronic cure. Aren't they
equally complex, equally flawed?
Vandersteen: They are equally complex, but the electrical cure is
much easier to control. You aren't really trading one problem for
another because when you get that first near-field reflection, the
ear/brain combination has a very difficult time separating the
source from the garbage. It's such a small period of time, and the
ear is not so much an amplitude device as it is a time device. I
think this goes back to prehistoric times, when you had to know
where the bear was coming from. For instance, if you phoned me from
England, I would recognize you, even though the telephone is a
horrendous amplitude device. The ear uses time signatures and phase
and so on; that's how we recognize what's realistic and what isn't.
Of course, we haven't accomplished perfection. Eliminating the
baffle eliminates all of these earlier errors, because sound
radiates from a transducer like ripples from a pebble dropped in a
pond. When those rings first strike the baffle board, even if the
board is padded with something or the corners are rounded (that
helps minimize the problems of the baffle because it splays and
averages the diffraction and keeps it from focusing on the
listener), it is still a problem because there is a lot of time
distortion in that diffraction. I have test equipment that can
measure that very easily. You put in a source, hit it again with the
same source, then vector the two together to get the difference of
what came from the drivers and what came from the baffle. What comes
from the baffle is not linear at all. It's a mess.
Kessler: Yes, but you do compensate for that. You know what its
characteristics are when you design a speaker with a baffle.
Vandersteen: You can accomplish flat amplitude response; it's easy
to do that. But you can never correct for a time distortion.
Kessler: But there are other time distortions created by not having
the rigidity of a tweeter and a midrange driver mounted on a baffle,
which has to be greater than free-standing supports.
Vandersteen: We have a free-standing support, obviously, which is
actually a small box. It doesn't matter, it could be a column, but
the boxes are small enough, in order to make them very strong.
Because of their columnar effect, you don't have to worry about the
resonances in these materials because all these parts are different
sizes. I use different thicknesses of wood, different damping
material on the inside, so that if you assume everything has a
problem, at least they never add on top of another. It's kind of
like the random theory. But as long as those structures are less
than a quarter-wavelength of the driver frequencies, you can say
it's baffleless even though theoretically there's a very small box
there. The old LS3/5A, for example, with the exception of
frequencies over 5k and the sharp edges around there, was basically
a baffleless speaker in my opinion. So is any other mini-monitor, at
whatever frequency of which the dimension of the baffle is less than
a quarter wavelength. Which goes quite high up in the midrange on a
mini-speaker. And that, I think, is why, even though they have a lot
of other problems—a small speaker's distortion is very high—they're
still pleasant from a musical standpoint because they have minimized
the time distortions created by baffle reflections.
Kessler: What are the aspects of a baffleless design that prevent it
from attaining the levels of openness and absolute transparency of,
say, a dipole, or a panel-type loudspeaker?
Vandersteen: That's a very good question. Panel speakers have a high
end that is exaggerated in a way that's kind of opposite to a lot of
recordings, but in the midrange—this is just my opinion—in the
midrange, they have a genuine realness. Every one of them. Even the
bad designs have a genuine something about them that just sounds so
damn real. This has puzzled me for 17 years, but I think I've
discovered what it is—they have no diffraction problems behind them
until they hit the rear wall. So you don't get those minor time
distortions you get from a baffle, when its energy is reflected and
delayed and sent back to the listener no longer in phase or correct
in amplitude, because you have to have a problem in the driver
offset by a problem in the crossover and/or the baffle. You get too
many wrongs trying to make a right, and I basically don't believe in
that.
It turns out that the internal diffraction problems of dynamic
drivers are far greater or as great a problem as the problems with
the baffle, because there's as much energy coming off the rear of
the cone as the front of the cone, and here you have a big magnet
and a basket and everything, with all of this crap delayed only two
inches—in time a few milliseconds—before the sound is reflected
back. We probably pioneered minimizing the external diffraction
problems involving time and phase—people have said that they enjoy
what it does for us, and I think it's real—so the natural extension
was to work on our own drivers where we have no reflective surfaces
behind the radiating diaphragm. We're applying for a couple of
patents on it; we're using very, very powerful magnets I bought in
England. They're very expensive, but that's the secret, because the
structure can be very, very small. And the basket is configured in
such a way that it has no reflective surface; it's like a spider's
web.
Now, this vectoring thing I told you about where you can look at
what's coming from the cone: you can vector it out and look at
what's coming from elsewhere. I measure it with the microphone in
front of the cone, using the Crown TEF machine. You can eliminate
the sound that came immediately from the diaphragm, take it out of
the picture, and look at what's left a very small period of time
later. It turns out it's only about 20dB down and is not linear:
it's a lot of garbage.
Drivers have always been designed to be flat to take into account
the negative effects that come back from all this structure, so
without them, the cone itself has to be flat, without these
reinforcements and cancellations and so forth going on at certain
frequencies. It's been a very difficult four-year project, we've got
an incredible amount of money involved in it, but it's finished now;
the Vandersteen 3A's here, and that midrange driver is in it.
Kessler: Would you classify this as the first dynamic driver with
all the beneficial properties of a planar driver?
Vandersteen: We don't know. The market will tell. I think this
midrange approaches some of what drove me crazy for so many years
about the planar speakers. We have more of that now. Do we have as
much in all areas? Probably not. But then the planar speakers have
other problems, so you're talking about tradeoffs here. Maybe we've
made it close enough in that area so that some of the other
advantages of a small point source are going to sway some people.
It's not a problem—we sell a lot of speakers—but when I listen, I
always miss that midrange; midrange I could hear in a planar design
so dreadful no one would give it a second listen. But if I listened
around all that, it still had that midrange character I wanted.
Kessler: Because this midrange unit is a relatively new type of
driver, doesn't it have its own signature that causes
discontinuities?
Vandersteen: Our tweeters have already had that effect incorporated
in them for the last three years. With a woofer, although it won't
work in a two-way, in a three-way you can crossover out of them low
enough to where these negative effects from the physical structures
are happening at a higher frequency than you're using the driver to
cover. A woofer's structure is basically transparent as long as you
get out of it soon enough. Now at some point you get into the same
problem as in the midrange frequencies, at 3–4kHz; a well-made,
cast-aluminum basket, 8" woofer, any one of many good ones
available, will have the same problem. But that's the whole beauty
of a four-way or a three-way design—you can get out of it before
that frequency.
We have had to redesign the woofer, but that was done because we
were using the same basic woofer in our 1B, which is a little
two-way. The reflection of the backwave is a great problem in a
two-way. That was dealt with acoustically, not as well as in the
midrange driver, but enough to keep it from nagging us on a
three-way design. And it helped the two-way, but it just minimized
the problem rather than solving it.
Kessler: What is the next stage in the refining of the baffleless
system?
Vandersteen: The drivers have literally to be meshed to 0.1dB to
make the concept work, because you don't have the averaging with the
baffle anymore. So the crossovers and the drivers have to be
fanatically matched to one another, and that is a production
problem. It's also a design problem in that I have to find a faster
way to do it. I'm very involved in computers right now. We've always
done this, but I would like to match the drivers even more
accurately to one another without having to jack the price up.
That's something the Dutch in me won't allow to happen.
Kessler: You're now at a stage where you have models covering just
about every level—though none of your speakers are actually budget
designs—all the way up to the high end. Now, aside from production
problems, which are refinements and have nothing to do with design,
I believe that the next step in the development of the Vandersteen
family has been working with subwoofers.
Vandersteen: Powered subwoofers. The only way to do it.
Kessler: Don't you find it a hindrance in the sense that you're
supplying the amplifier, and that it forces the end-user to match
whatever amplifier they want to buy with the characteristics of the
amplifier you're using?
Vandersteen: We've got a unique solution for that. Our subwoofer is
bizarre in that it has a first-order passive filter, nothing but a
capacitor between the amp and preamp, rolling off the main amplifier
and the main speakers at 6dB per octave, below 60, 80, or 100Hz—your
choice. When I was developing the subwoofer, I found that the thing
had to be quick, and it had to be floor-mounted so that you always
knew its acoustic environment—the one thing we all have is floors in
our homes. This thing is slot-loaded and floor-mounted. That makes
things very predictable in low frequencies. That's one advantage. It
causes some problems, but we can use a feedforward concept to solve
those. The other thing I discovered was that many times the reason a
subwoofer wasn't seamless was because you had different amplifiers,
or different amplifier characteristics. And that was a big problem.
So that delayed the subwoofer for three years.
We took an input circuit from an oscilloscope, which is a
dual-differential floating input, so it doesn't matter if an amp's
bridged: ground is not really ground. Our subwoofer samples at the
amplifier output terminals. So you bi-wire, or if you're already
bi-wiring your speakers, you tri-wire off the output terminals. The
input impedance of the subwoofer/amplifier is about 4 Megohms. So
the amplifier hasn't the foggiest notion the subwoofer is even
there. It's like putting an oscilloscope across the terminals to see
what's happening. It then samples that. Now remember, the amplifier
is being rolled off below 80Hz by the passive filter, but those
frequencies aren't gone, they're just attenuated as they go down.
The subwoofer input circuit knows that this signal's been attenuated
and restores it, meaning that you pay a 6dB penalty in noise. This
is not a problem in the deep bass, at least not with the circuitry
we're using.
This technique means that the amplifier playing your woofers sounds
like you've got a great big version of whatever amplifier you had in
your main system. The first negative comment we get is, "Wait, I've
got a tube amp and it's not so great in the bass. You know, I don't
really want that character carried on into the bottom." But a tube
amplifier into a four Megohm load has a damping factor of how many
thousand? It's only when tube amps have to deal with the mass and
the movement of woofers that they take on that character. Their
signal is state-of-the-art until it has problems dealing or
interfacing with a large woofer.
In this case it has a four Megohm resistor to deal with, so it's no
problem. Then you go through a high-current situation and there're
three floor-mounted 8" drivers in there, a 300W amp, and it goes for
it. So it will take on the little phase nuances and so forth that
one amp has from the other. We don't know how to measure all these
things, but we all know they sound different. But this thing will
take on whatever character of amp you put on the top, minus some of
the negative things. Those characteristics of an amplifier that are
caused by a design problem are not passed on into the bass because
that design problem has been removed. It's very innovative; I think
it will be very widely copied in the future.
Kessler: Have you patented it?
Vandersteen: It is a patentable concept, but I basically don't
believe in patents because, you know, who has the time to spend
enough money and the time in court to back it up? In this industry
there's not that much copying going on because the designers have a
fair amount of ego, and no one will ever do it as cheaply as we can.
I think it will be pretty obvious in the future who came up with the
idea; I'll just let the market take care of it. Patents really don't
protect anything. The US government will issue me a patent, but
they're not responsible for enforcing it. That's done in the courts.
Kessler: Where it would cost a lot of money.
Vandersteen: Well, it's not the money, it's the time and the hassle.
I've got better things to do. We're getting a patent on the concept
behind this new midrange, however, because I think that is special,
and easier to enforce.
Kessler: What is the next stage of development?
Vandersteen: At this point, we're selling a lot of product in
Europe, and I want to come up with a ".5," a very high-quality
mini-monitor of some sort. That's the present project. In addition,
for the last three years I've been doing a lot of work in
surface-mount speakers, trying to make them work in a wall—I'm very
frustrated by that. The environment just doesn't make it! I could
make one a little better than somebody else's maybe, but I'm still
investigating a way to make it really work. I guess it's because
they just don't make recordings with the mikes against the wall. The
bass has some great advantages, but in the mids and highs you get so
many time distortions that you just can't make it sound real. You
can make it sound good and pleasant enough that many people would
buy them, but I'm not really interested in how many we can sell.
Obviously I want to stay in business and make a profit, but I want
them to be something unique and special.
Kessler: Wall-mounting a speaker would seem automatically to defeat
the notion of the baffleless design.
Vandersteen: Exactly! Absorption is an option, but I haven't found a
material yet that absorbs well enough and is linear enough.
Kessler: But the non-linearity is something you can compensate for
in a known design.
Vandersteen: Somewhat, but that's almost like saying you can use
feedback, and we don't like to use feedback; look at our subwoofer.
There are many feedback servo-type subwoofer systems that go lower
than hell, that will buckle the walls, that have low distortion
numbers, but they sound like damn computers to me—they don't sound
like music. I think the reason for it is that you can't let a woofer
screw up and then say "Oh no! We screwed up! Measure how much!", and
then bring it back around and correct for it. It's a fatally flawed
concept.
Kessler: The servo subwoofers that I've heard have sounded
overdamped.
Vandersteen: Yeah, but overdamping is easy to correct—you raise the
Q, don't damp it quite as well. However, then you have another
problem: it always seems to have something not quite right. You
can't let something screw up and then decide to fix it later with
feedback, because it's after the fact—again, a time distortion. We
don't know how to measure that, don't know at what point it's too
much, or less, or not enough. One advantage tube amps have
traditionally had is that designers use less feedback in them.
Kessler: Tube amps do have lower damping factors.
Vandersteen: So what? You deal with those problems in other ways. If
a guy makes a nice impedance curve on his speaker and put a little
bit of output impedance in his formula when he designed the box in
the first place, then a low damping factor works fine. It's only the
guy who designs a speaker assuming zero output impedance who's going
to have a problem. I don't think that's wise anyway—all amplifiers
have output impedance. Even if they didn't, you still have speaker
wire you have to connect it with, so our speaker works unusually
well with tube amps with solid-conductor cable—I'm not saying I
agree with that technology, but it works well—because we put some
output impedance in the formula. We didn't assume an amplifier with
an infinite amount of damping factor and zero-output impedance.
Again, that's a problem, but you can control all those things,
especially in a subwoofer with an amp. But there're so many
advantages in low bass, knowing what woofer's going to be there,
what box, how it's going to be configured, plus this other little
trick we're playing, that it's a real advantage.
It's real important for systems to be musical. We can't turn
ourselves into pieces of test equipment and try to measure how these
things perform. You have to react emotionally. Music was Valium in
the old days. If it's done properly and if the system's working
properly, that is its function. It has been for centuries. We should
address these things from a musical standpoint. Even though I'm very
technological in my measurements, the ears are the bottom line, the
final piece of test equipment. All you do is hope enough people
agree with you that they keep you in business.