An Interview with Thom Mackris of Galibier Design

LB:      Let’s turn now to speed control.  In principle, the motor system — which comprises the motor per se, the controller, and belt — has a simple job, namely rotating the platter at a constant speed.  In practice of course, that is anything but simple.  Tell us about your choices and implementation.

TM:    The drive system comprises all of the moving parts, including the bearing, lubricant, and platter, as well as the obvious components you mentioned.

In designing our new drive system, I was looking for better speed “lock” than traditional DC could provide.  I was presented with what appeared to be difficult choices in AC motors:

• Small (15-50 watt), 3-phase AC motors are no longer made
• Current production AC motors have vibration issues
• High quality, low production AC motors typically used in military applications are made by small companies with no guarantee of model continuity

The search took us to 3-phase, electronically commutated DC motors – the industry replacement for low-power, 3-phase AC motors.  These motors are controlled electronically, using positioning information fed from the motor.

The softest fabric belts we were able to manufacture are on the stiff side, and require a larger pulley in order to run the optimum belt tension while conforming to the pulley.

This larger pulley’s gear ratio increased the power requirements, but conferred the advantage of quieter operation due to lower motor speed.   A lower speed was mandated by the more robust ball bearings in these motors.  Ball bearings can get noisy at high speeds, but they have a considerably higher mean time before failure and we did not want to compromise durability.

So, slowing down the motors to 300 rpm (for 33-1/3 rpm playback), conferred the advantages of facilitating use of the stiffer belt (for better drive coupling), airborne noise reduction, and more belt traction due to the increased contact patch on the pulley.

If your readers are interested, they can read about the drive upgrade on our forum:  http://www.galibierdesign.com/phpbb/viewtopic.php?f=4&t=187

LB:      Some manufacturers use two or three motors (if not in their entire line, then in their top-of-the line model(s)), the rationale being that a single motor pulls unevenly, thereby causing the platter and bearing to “rock.”  Have you ever experimented with multiple motors?

TM:    The rocking argument always makes me smile.  (I’m glad to have brought some humor into your life — LB 🙂 It continues to rear its ugly head, whether from advocates of inverted bearing design, or those advocating Byzantine, overly complex drive systems.  Those who would make this argument are either unable to design a proper bearing, or are taking their audience for fools.

A Galibier bearing does not rock.  None of our measurements have pointed toward this, and we’ve never seen a reason to experiment with a multiple motor solution.

We’ve always found that the simplest approach (done right) has prevailed.  I realize that this sounds dismissive of complex, seemingly elegant solutions, but there’s a limit to the number of thin branches of the tree you can crawl out onto during your experimentation.  Being guided by Ockham’s razor principle has served us well.

I’ve also seen the argument that three motors cancel out the effect of speed variances between motors.  Designing a proper drive system is a more direct approach to this.

It’s possible that they’re trying to address the instability resulting from rubber (compliant) belts – attempting to mitigate this with shorter paths between each contact point (pulley to platter).

Again, (applying Ockham’s razor), the correct solution is to eliminate a chaotic system by using a non-compliant belt, and not by trying to minimize the limitations of a rubber belt, which is an exercise in futility.

LB:      How big a role –if any — do measurements play in your designs?  Do they correlate well with listening?

TM:    It’s pretty much what you’d expect – measure to establish a baseline, listen/adjust to what sounds “good,” re-measure, rinse, repeat.  There’s quite a convergence between good measurement and good sound, but we observed one interesting thing during the beta testing of our new drive system.

We measured the performance of our new fabric belts against the old standby – spliced Mylar tape.  There were slight speed fluctuation differences between the two.  The fluctuations for both were well below the threshold I discussed earlier – below the audibility threshold of the wow and flutter time domain.  Preferences for the two belts were split evenly.

Selection of platter materials was another case of working with both the numbers (measuring the speed of sound in materials), as well as listening.  The “paper” experiment I referenced in our discussion about platter resonance, is one of the many experiments we made to test the numbers and fine tune our results.

At the 2004 European Triode Festival, Pete Millet did a presentation on the sound of distortion (http://www.pmillett.com/etf_sod.htm).  An interesting takeway from this was that injecting a bit of odd order distortion into a single ended amplifier (they’re heavily biased with even order harmonic distortion – that “sweet” sound) was preferred by most of the attendees.  The attendees were a collection of talented designers.  A slight blend of odd order harmonics added a bit of excitement to the presentation.  There was a point beyond which, it was too much, however.

The point of all of this is that we design to the numbers and then tune by ear.  There’s still a lot of art involved, and knowing what music really sounds like is one of the best tools in your quiver.

People who know me, know that I am very body-centric.  During listening sessions, I watch (in myself and others), the physical response to music playback, and not just whether they’re tapping their feet.

I also observe the language people use when discussing the playback, and I know something good is occurring when people speak in present tense, performance terms – when they make comments like:  “did you hear what she just did?,” “can you believe the syncopated groove that drummer is laying down?”

We’re well beyond the old byline – the sound of music created in real space – to something far more difficult to describe and capture.  Manufacturers of late have caught on to something I’ve been saying forever – that the equipment should connect you with the intent of the composer and performer.  This involves much, much more than capturing the basic tonality of the instruments or of hearing the reflection off the rear wall in Orchestra Hall in Chicago.

A good hi-fi system should be a space/time portal, and if people aren’t effusively using performance terms (without thinking about it) like the ones above, then a system isn’t weaving a magical spell and something needs fixing.

Lynn Olson wrote an excellent, provocative essay on this topic with the title “Illusion Engines.”  I direct all of my customers to it, if for no other reason than to stimulate thought.  Some of the comments are dated (written in the early days of the triode renaissance in the US), but the principle thesis is solid.  You can find it at this link:  http://www.nutshellhifi.com/library/illusion-engines.html