Hi Jean (from another Jean), seeing a big VPI turned into an idler-wheel project warms the cockles of my heart. Absolutely stunning, beautiful work, and good thinking outside the box too! And of course I really like the direct coupling ;-)!! Glad to hear you heard the improvement. Now even Harry is getting into the Idler Game, so kudos to him too for trying, hearing and accepting!! It would be nice if you could eliminate belts altogether from that project, as what you have now is a sort of Super Thorens TD-124, already better than a pure belt-drive as you have discovered, but in comparison with true idlers (Garrards, Lencos) having lesser dynamics, transient speed and bass (the Thorens however would be improved by substituting mylar tape for that rubber belt), but it is already sounding very complicated to add in yet another platter to add to the drive system to accomplish this. But it is very interesting to see the sonic development of the idler-wheel principle in steps as you go along: the increased torque as you add rotational mass and apply the idler principle, the increased clarity and focus as you eliminate the suspension, the increased speed stability as you eliminate the compliant belts, and so on. It all bears out the idler wheel principle as it was developed decades past, and the seriousness of stylus force drag, which is what it is all about :-)!! Fabulous and constructive experiment.
So the other issue with idler-wheel drives is simple torque: there exists nothing in belt-drive motors today to equal those massive powerful idler-wheel motors of yore (which spun at 1500-1800 RPM), but, in keeping with the spirit of the experiment, you could tear a page out of the British Voyd book, creators of the multiple-motor approach now being adopted by certain high-end manufacturers: add more motors to increase torque. I don't know how this would factor in cost-wise. Voyd's 'tables went up the ladder as they added motors, their top having three motors to create their signature powerful sound at the time. You could try adding a second VPI motor, thread-drive makes this workable, as they come already handy in separate pods, and see what happens. I look forward to your continuing experiments and experiences with idler-wheel drive principles - torque, proper use of inertia, proper coupling, grip - keep it up!! Maybe some day you'll actually try a properly-rebuilt Lenco or Garrard as Reference and let us know your findings ;-). Your experiment is a very constructive exercise which increases our understanding of drive systems in vinyl replay. And keep your ears open for such intangibles as magic, musical excitement, PRaT, gestalt, harmony, which, since they are heard by audiophools the world over, aren't so intangible after all :-)!!