Interviews

Bill Firebaugh's first product, the outrageous-looking Well-Tempered (tone) Arm, established him as one of audio's most innovative designers. At the 1985 Winter CES, he showed a prototype companion product—the Well-Tempered Turntable—and was producing production units by January 1987. He discusses here the WTT's unusual design features. (Readers should note that, since we have not yet tested the new turntable, this interview is not to be interpreted as an endorsement of the product.)

Holt: Bill, what prompted you to get into the turntable business?

Firebaugh: Well, after I had made my arm, and achieved what I felt to be the correct degree of stability in it, I thought it would be a rather simple matter to make a turntable. I thought, "There's nothing to a turntable. It's a platter, a bearing, a belt, a motor, and so forth." To me, it seemed to be the simplest thing in the world to make them all work together. So I started out using an aluminum platter, and I made a very nice spindle sleeve and ball-type bearing.

Holt: In other words, the usual thing.

Firebaugh: The usual thing. It took hardly any time at all to make a prototype, but it didn't work well at all. It didn't have that nice, sweet, musical sound, and I wondered, "What's going on here anyway?" After a period of consternation, I started experimenting with these sleeve-spindle, ball-type bearings, and I made maybe three dozen different varieties, trying different materials and different finishes and geometries.

Holt: Why did you assume right off that it was the spindle bearing that was at fault?

Firebaugh: Because if I turned the bearing by hand and listened to it, I could hear it go "scrape, scrape, scrape" as I rotated it. It didn't sound smooth. I could feel the roughness when I turned it. So I made lots of different versions of that conventional bearing arrangement, including ones with standpipes of lubrication. I came to understand that one of the really important issues is lubrication. If you have a spindle fitting into a sleeve and there's only a fraction of a thousandth of clearance between them, there's hardly any room for lubricant to get in there. And any small speck of stuff that gets in there really causes trouble.

So I made quite a few prototypes that had reservoirs of oil; I would drill a hole through the side of the sleeve and connect it with a reservoir of oil so it could get to the spindle through the hole. And then I added more holes to get the oil to the spindle because it still sounded kind of dry.

Then it occurred to me that, if you have enough clearance around a normal spindle that it doesn't bind, and then you have a belt pulling it to one side, it's going to tend to walk around the top of the spindle hole until the belt tension pulls it back, then it's going to go "Clunk!" and do the same thing again. And you know that's not going to help the sound. I believe that's what makes belt tensions and all those things so critical on a normal type of bearing. But I was stuck. I couldn't seem to find a way around the problem.

Holt: How did you solve it?

Firebaugh: I just waited. And while I waited, I kept kind of fiddle-diddling, and eventually new ideas came.

I started thinking about what it would take to provide stable contact between the bearing and its housing without confining the spindle so it would jam up, and I thought of a round peg in a square hole. You pull the platter toward one corner of that square hole, the way belt tension will pull it, and the spindle rests firmly against the edges of the hole at two points 90° apart. And the bottom of the spindle, of course, leans up against the other two edges of the square hole. The spindle can't wobble, yet it isn't confined to a limiting space at either its top or bottom. So I made up a square spindle-well out of nylon and it worked: there was no play at all in the spindle. I thought I had discovered this radical way of using, for the first time, the old round-peg-in-a-square-hole idea.

But then, one Saturday night—actually it was a Sunday morning, I'm a night-owl type—I thought of a simpler way of doing the same thing. If the well is only contacting the spindle at four points, why not just do away with everything except those four points? I thought about using a big spindle sleeve, with an eighth of an inch clearance between the spindle and the sleeve, and then putting in four supports that come through the wall of the sleeve, so they only contact the spindle at a couple of small points.

I had some nylon setscrews on hand, so I threaded those and put them in from outside the spindle well at right angles to each other, so the top of the spindle was seated between two screws at the top and two at the bottom. It worked, so I knew I was headed in the right direction. But there still seemed to be too much friction. There was lubricant in there, but the bearing still wasn't as smooth and easy-turning as it should have been.

So I tried Teflon and polyethylene "screws" instead of nylon, but they still seemed to be binding too much. In addition, they were greatly accentuating the surface quality of the spindle. I wanted something nice and soft and smooth that the spindle could ride against, so I tried using a sort of a setscrew with a rubber point on it. I took a metal setscrew and drilled a hole in the end, and I cut off a piece of an O-ring of the same size and stuffed it into the hole. And that was it! Boy, that was so silky and smooth! It made an immediate difference that you could hear.

I eventually wound up using metal setscrews with the end ground off flat and a 3/16" flat disc of nitrile rubber bonded to it. Nitrile rubber is the stuff oil seals for car engines are made of. Working against polished steel, those seals will last for a hundred thousand miles.

Holt: But wouldn't rubber have higher friction than Teflon against the spindle?

Firebaugh: By itself, yes. But there's maybe a shotglass full of silicone lube—Dow-Corning 200 silicone fluid—in the spindle well. With silicone fluid, this nitrile rubber against metal is extremely slippery. And furthermore, there's a fair degree of damping in the bearing, simply by virtue of the fact that you have five of these pads riding against the spindle. And it's not only damping the rotation, it's damping any vibrations that might be excited within the shaft itself.

Holt: You use one of them at the bottom, to support the platter?

Firebaugh: Yes. And the bottom one is off-center! If I centered it, it just squeezed all its lubricant out and ran dry. Putting it off-center means the part of the spindle surface that isn't on top of the rubber disc at any given moment picks up a fresh coating of oil. The lubricant is constantly being drawn into the surface between the disc and the flattened end of the spindle.

Holt: I noticed your turntable uses the drive belt flipped, so that it runs upside-down around the motor pulley. What's that suppose to accomplish?

Firebaugh: It seems to have a significant effect on flutter. I spent a lot of time studying flutter, both on a 'scope and on strip-chart recordings. I was trying different belts, and I suddenly noticed one of them was giving lower flutter than it had previously. And then I found it was because I had put it on with that half twist in it. It was a lucky accident.

I'm not sure why it reduces flutter, but I think the reason is, as the belt comes off the platter, instead of losing contact the same distance from the motor pulley, the top of the belt separates before the bottom of it.

Holt: You're attributing the flutter to vibration of the belt?

Firebaugh: That's probably it. See, the frequency at which the belt will vibrate is related to the length of it that's in free air—not contacting either the platter or the pulley. Both of those will prevent it from vibrating. With the usual belt threading, the top and bottom of the belt leave those surfaces at the same time, so the lengths of free belt are the same at the top and bottom of the belt. The whole free length of belt will vibrate at the same frequency, giving a fairly high-Q resonance. But if we put a half twist in the belt, its top and bottom edges leave the platter and pulley at slightly different places, so the belt resonance is more distributed in frequency. It has a lower Q, so vibration at any single frequency is much reduced.