I hope this is ok to post here. I just found this posting by Mark Kelly over on AA related to this topic.
"The idler drive mechanism has only a few percent of the speed variability inherent in belt drive (due to the lossy transmission eg belt creep).
Most idlers make up for this by using motors which can't hold constant speed against variations in platter drag (like the stylus traversing the record). Anyone who says their stock Garrard holds constant speed simply isn't measuring it properly.
As against this, the short term speed stability of an idler is excellent due to the almost lossless transmission from a motor with a high amount of flywheel effect. Most people assume this is due to the motor having lots of torque but that's not the case.
The other problem is noise - an idler definitely transmits more noise than a belt and they can create extra noise all of their own if the bearings aren't perfect.
Altogether its a toss-up.
Mark Kelly"
"The idler drive mechanism has only a few percent of the speed variability inherent in belt drive (due to the lossy transmission eg belt creep).
Most idlers make up for this by using motors which can't hold constant speed against variations in platter drag (like the stylus traversing the record). Anyone who says their stock Garrard holds constant speed simply isn't measuring it properly.
As against this, the short term speed stability of an idler is excellent due to the almost lossless transmission from a motor with a high amount of flywheel effect. Most people assume this is due to the motor having lots of torque but that's not the case.
The other problem is noise - an idler definitely transmits more noise than a belt and they can create extra noise all of their own if the bearings aren't perfect.
Altogether its a toss-up.
Mark Kelly"