How important is low W & F performance anyway?

I think that we are not really asking the right question. Personally, unless a TTs W&F is really bad, I don't notice it as speed error per se. (Perceptible pitch changes in say piano's decay would be considered to be really bad)   But as the TT's figures improve, what I do hear is a solidifying of the sound along with improvements in attack and a lengthening in decay time. A subtle improvement in focus, to use a visual analogy. It becomes easier to hear, see,  into the music. 

Further W&F figures aren't that relevant as they are virtually all measured during steady state operation, a test tone for example. Music is not steady state so it places dynamic loading on the drive thru stylus drag. How it responds to this is much more important. 

Years ago I owned a Goldmund Studio, in an attempt to improve it, I built a regulated power supply based around a LM317 3 pin regulator chip. The change was disappointing. I scoped the supply voltage while playing a record and noticed that it was being modulated in sync with the beat of the music. ( music was muted)  This was a revelation leading me to come up with two explanations.

1) The output impedance of the LM317 supply was high, making it a bad choice for this application 

2) Much more importantly.... the stylus drag was slowing the platter. The JVC motor's feedback loop was sensing this and drawing current in sync with the music's beat in an attempt to hold speed steady. 

I don't recall how heavy the platter was. Certainly it was, for a DD, a fairly high inertia design. 

So how does a TT measure when playing a dynamic piece of music. Now that is a real test of its speed accuracy. (If we hear changes in pitch, the TT is basically broken.) What we are looking for is stability and focus of the music itself. Solidity of the notes, just like the real thing.

Cheers.  

@elliottbnewcombjr 

"The tolerance of the housing and bearing shaft were so fine that an air spring was created, it took several minutes for the air to get out and the platter/shaft to descend fully. I’ve never forgotten that precision."

Even inexpensive bearings work that way. I remember first setting up my Revolver Rebel turntable decades ago, after squirting the premeasured lubricant from the syringe the manufacturer provided and dropping the spindle into the bearing housing, there was very long period of settling of the shaft, might have taken up to an hour if I recall and the manual cautioned not to touch anything while this was happening since it might damage or score the assembly.

All Rega turntables go through this during the manufacturing process and the mating parts arrive sealed and protected from foreign matter prior to assembly. These bearings can remain untouched for decades without any maintenance at all.

You should all take a lesson from @richardkrebs.

@faustuss 

this video, assembling a Linn TT in the Bedrock Plinth (it’s not a plinth, merely a frame).

https://www.youtube.com/watch?v=SqP4eHBSlI4&t=2s

watch and listen at 5:15, see the platter and shaft drop like a stone, AND listen, a clunk, what the hell is that?

The motor has four shaded poles, so it is not quite synchronous, instead slipping slightly with the braking effect and stylus drag.  Personally I think stylus drag is tiny compared to the eddy current brake, which could explain why these decks are valued for their ’drive’ when playing heavily modulated passages.

I’ve seen this rationale applied before in the past and it is counter-intuitive to me.  The concept is if you load the drive/platter with torque (either through an eddy brake or thick bearing lubrication), then small dynamic changes in torque will be "swamped"  by the static torque load and will therefore go unnoticed.  If the speed is stable with whatever static load is placed on it, the system is in balance and dynamic loads will still disturb that balance proportionate to the amplitude of the dynamic load, not in ratio of the dynamic to static load.  The only way this makes sense is if the torque load/rotor angle is not linear, which it is not-torque is proportional to the sine of the rotor/field angle, but the relationship is very nearly linear at the lower end of the torque curve.  So unless you are applying a static load that is very near the stall torque (close to 90° rotor/field angle where the slope starts to flatten) it’s difficult to see how this could have any benefit.  I suspect the improvement in sound that some report with thicker bearing oils is due to mechanical damping and not to speed stability.

@phoenixengr. I tend to agree with you.  Viscous pre-loading of the drive can be beneficial as it damps the bearing and by extension the platter. But perhaps more importantly, it damps any tendency of motor to oscillate about the synchronous speed when presented with a sudden change in load, torque demand. 

Cheers.