Agree with the viewpoint that, in principle (all other things being equal, which they probably won't be in reality), there should be no difference between torque stored as inertia in a high mass platter, and torque as applied by the motor -- at least in the moment of initial transient attack. But the dynamics of how the platter recovers lost speed in preparation for the next transient event, or during sustained peaks, may well be different, and maybe could favor the lower-mass/higher-torque model. (And/or also the model of active speed sense-and-control?) Anyway, there are clear advantages of price and ease of placement in not having to get a TT with a platter that's half your body weight. A lighter platter also eases the job requirement of the main bearing. A degree of flywheel effect is certainly desirable in any TT design, but to me there's something dissatisfying about the idea of always resorting to supermassive platters.
I'm not sure I can entirely agree with Johnnantais on a few other things, but he probably knows more about it than me. Still, I can't reconcile the blanket assertion that a motor turning at 1,500 RPM is intrinsically better than a motor turning at 33 1/3 RPM -- that high rpm "by itself" tends to "smooth out" speed imperfections. If we assume the same level (amplitude) of vibrational contamination for the sake of argument, the only difference, it seems to me, will be in the frequency and its harmonics, which will also alter the spectrum of intermodulation products. That will change the sound, but I don't know which one is better, or what might constitute the "ideal" RPM or frequency to superimpose upon the music signal (none would be nice). But I doubt it's as simple as saying "higher is better", especially when only one fundamental falls within the audioband (1,500 RPM = 25Hz, while 33 1/3 RPM is only about 1/2Hz). I'm not exactly sure if all this stuff is pertinent to the topic at hand though, maybe an engineer or physicist could elaborate further.
If we stipulate (correctly, I hope) that high torque is what pushes the platter through dynamic passages, then as far as I know, in electric motors lower RPM = higher torque (larger motor diameter does too, which DD's also generally have). I have zero experience with the big idler-drivers mentioned, so I'm guessing that for their motors to be both high-speed and high-torque, they must also be relatively large and powerful (higher amp). If so, wouldn't that increase the amplitude of vibrations vs. a lower-speed, lower-amp motor with equivalent torque? Obviously there will be a lot of other variables regarding how any two motors in question are made, but I wonder in principle.
However, it is undeniably true that the faster you spin anything that's not in perfect balance, the more violence will be imparted by its shaking. Can a motor be perfectly balanced? If so, I guess we wouldn't need to have this discussion. So, adding it up, I can't buy the notion that higher motor RPM's are somehow "better" for TT performance. I've always assumed one of the strengths of DD was the low-RPM motor: high-torque, low amplitude of vibration, low frequency of vibration. Somebody please point out the error of my thinking if I'm wrong. Absent the RPM argument, I haven't yet detected the theoretical case for why idler-wheel drive should be intrinsically superior to direct-drive. That it may prove subjectively superior among certain 'tables auditioned by certain listeners is another question.
Maybe I'm way off base somewhere here, but speed performance is measureable -- you don't have to depend solely on subjective listening impressions to find out about this particular factor. Does anyone know that idler-drive tables can have better wow & flutter numbers under dynamic playing conditions than a good quartz/PLL controlled DD? Are there any constraints that prevent a DD motor from being made just as powerful as for an idler-drive, if that's what one wanted? (Magnetic interference with the cartridge, perhaps? The platter's shielding ability may be limited, but I don't know -- I thought the Rockport DD was supposed to be terrifically powerful and nearly impossible to deviate the speed of. If you pay enough, you can get a high-torque/high-mass DD table, with the electronic control to harness both.)
To me, DD does seem to have several intrinsic theoretical advantages. One is that the motive force is applied without physical contact of the platter. If you use a belt or a wheel to transmit force to the platter, it must result in some added degree of motional friction, which must produce a characteristic resonance, much like road noise in tires (okay, so it's probably more like urethane skate wheels on polished marble tile, but the principle still applies -- it's still not silent).
Another is the fact that there's only one bearing, the ubiquitous main bearing. In any design with a separately housed motor, that motor must have its own bearings in addition to the necessary main platter bearing. Then there's shafts and pulleys or wheels -- they won't be perfectly concentric. DD does away with powered driveshafts, pulleys, and/or wheels.
I tend to think simpler is better (simpler, but not too simple! ;^) A TT doesn't get any simpler, mechanically speaking, than a DD: the plinth is the stator, and the platter is the rotor. In any other arrangement, the separate motor can and will move in relation to the platter, which causes variation in platter speed. In a DD the motor can't move relative to the platter, since the platter is itself one half of the motor.
Another thing: I don't know if this absolutely has to be the case, but as far as I know only DD's incorporate electronic sensing and control of platter speed directly, rather than control of motor speed (but typically without sensing, I believe) with a somewhat flexible linkage in between it and the platter. I know which arrangement seems like it would be better to me, but I'm open to arguments. Again, though, the results can be measured.
Along those same lines, there's the issue of how correct speed is established in the first place. With a PLL sensing system in place, it's easy to add calibration to a quartz crystal reference, a very much higher and more constant frequency than (and totally independent of variations in) the AC powerline. Who wouldn't want that if you can have it? And again, the results are measurable.
I've often seen an argument against PLL-controlled DD's that usually goes like this: They are constantly "hunting and pecking" for the correct speed, but never settle on it. I have never understood this. Any electric motor operates in what is termed a "kick and coast" fashion, dependent on the number of poles. More poles would seem to be obviously better than fewer poles, but I don't know that the number of poles possible is in any way linked to or limited by drive method. Anyway, it seems to me it's primarily this kick and coast phenomenon, dynamic stylus drag aside, that's primarily responsible for any TT not constantly rotating at exactly the correct speed, no matter how it's driven. Why this is blamed on implementing a PLL is something I want to know.
Another thing I want to know is why DD is often portrayed as constituting a "rigid coupling" between the motor and the platter (never mind that this doesn't make any semantic sense, since in a DD the platter is in fact part of the motor). At times I almost get the feeling that some audiophiles who haven't owned a DD visualize it as simply consisting of a belt-drive type motor -- meaning a self-contained unit with a housing and a protruding driveshaft -- with the platter stuck on the end of the shaft instead of having a pulley and belt in between. (If that describes anybody reading this, go to the website Viridian linked above and look at the platter-off pics of the SP-10.) Anyway, "rigid coupling" seems to imply that the platter can't "slip", which of course is 180 degrees opposite of the truth -- any DJ knows that only in a DD can the platter be freely spun manually when not under power, or manually deviated with precision from constant speed when under power, from which it will rebound when released. ("Rigid coupling" also implies vibrational transfer, which again to me is a conception misappropriated from the paradigm of separately-housed, self-contained motors physically linked to platters and plinths by compliant couplings.)
As I understand it, in typical audiophile belt-drivers, the elasticity of the belt, combined with the inertia of the massive platter, is supposed to mitigate the kick'n'coast speed variation from the motor. Of course this can't simply be "gotten rid of" -- the elasticity and inertia combine to spread its effects out in time, reducing the amplitude peaks, effectively averaging the variation in speed. Well, platters in DD's also have mass. The "slippage" and subsequent rebound that I described when a drag or an energy input is applied, isn't that functionally equivalent to belt elasticity? It seems to me that in a correctly designed DD, the PLL sense-and-control system, combined with the platter inertia and the natural ability of the platter to smoothly and infinitely vary from constant speed and then rebound without introducing mechanical friction, can constitute exactly the same kind of "averaging" mechanism that smoothes the kick and coast impulses in a belt-driver. The difference is you don't need a supermassive platter or the attendent pitfalls of a belt/pulley/separate-motor system to do it when you've got active speed control working for you.
I know many audiophiles regard the notion of any "servo" or "feedback" type of operation as something they're allergic to, whether it's negative feedback in amplifiers or servo control of subwoofers. As the saying goes, you can't correct something that's already happened. Same for many active vs. passive questions (though not always). I also know many engineers would argue with this attitude and say it's not that simple or universal a truth, that there can be well-implemented applications for feedback-type mechanisms that don't harm sonics in unintended ways.
I can't comment on all that stuff, or its applicability to DD TT design, with any authority (though I don't recall anybody saying that adding an outboard speed controller to their belt-drive TT made it sound worse). Yeah, I own a DD [KAB-modded SL-1200], but like I've said before, I lack the comparitive expeience to make pronouncements about relative superiority. But I also note, with no real satisfaction, that probably no audiophile, no matter how experienced, has ever had the opportunity to compare two turntables whose only difference was method of drive. As we all know, there's a host of other factors which affect TT sonics besides drive type. I also acknowledge that measurements, such as I touted above, often don't tell the whole story sonically speaking (which simply means we need other, better measurements to correlate with what we hear).
More important though are, as I see it, two questions: The one Drubin asks, i.e., do any of the DD's available in the moderate price range (current or restored) warrant consideration over the entry-level and next-tier belt-drive audiophile standbys (none of which are terribly massive due to their reasonable cost)? And the one Macrojack wants to know: Is it time for high end TT makers and audiophiles alike to reconsider the relative merits of the DD option -- could it be exploited to make even better tables than are generally available right now (and if so, at an attainable cost)? I admit I don't know the answer to the first, but feel the second has got to be a yes if at all possible.
