I've read some good reports about various kinds of thread, string, fishing line, dental floss, etc. I also remember one VPI owner who modified his table to accommodate a wider belt, used 1/4" wide mylar and reported good results.
The difficulty in implementing this kind of belt on a VPI (correct me if I'm wrong) is that the motor capstan and/or platter rim are grooved for a fairly narrow belt. This limits your material choices to whatever will fit, unless you modify your table.
If your platter rim is flat, getting a local machine shop to make you a taller motor capstan might be viable. For a 1/2" belt you need a capstan working surface height of about 9/16" to 19/32", very slightly convex and with lips at top and bottom to stop the belt riding off. The O.D.at the center would be the same as your existing capstan, obviously. Material choice probably isn't super-critical, but the surface should not be highly polished.
The key characteristics in a good belt material are:
1. Lack of elasticity. The more resistant the material is to stretching and rebound the better. Rubber is terrible (sorry). Mylar is dimensionally very stable. At the speeds and loads experienced by a TT belt it wouldn't be far wrong to call it "perfectly" stable, depending on the thickness.
2. Lack of "slipperiness", ie, sufficient friction to prevent slippage when a big transient tries to slow the platter. This is tricky to get right. The biggest risk of slippage is at the very short belt/capstan interface. The belt/platter interface is long enough to make slippage less likely there. The optimum material is very much table dependent. On a Teres motor capstan materials like string, thread, etc. result in quite audible slippage, though they're still better than anything elastic. Mylar will also slip if it's too thick and/or too smooth to make the curve around the capstan, and that's a function of your capstan diameter, material and surface.
Not much of an answer, sorry. But I encourage you to experiment. It can't hurt and you'll certainly learn something from everything you try, as we have.
The difficulty in implementing this kind of belt on a VPI (correct me if I'm wrong) is that the motor capstan and/or platter rim are grooved for a fairly narrow belt. This limits your material choices to whatever will fit, unless you modify your table.
If your platter rim is flat, getting a local machine shop to make you a taller motor capstan might be viable. For a 1/2" belt you need a capstan working surface height of about 9/16" to 19/32", very slightly convex and with lips at top and bottom to stop the belt riding off. The O.D.at the center would be the same as your existing capstan, obviously. Material choice probably isn't super-critical, but the surface should not be highly polished.
The key characteristics in a good belt material are:
1. Lack of elasticity. The more resistant the material is to stretching and rebound the better. Rubber is terrible (sorry). Mylar is dimensionally very stable. At the speeds and loads experienced by a TT belt it wouldn't be far wrong to call it "perfectly" stable, depending on the thickness.
2. Lack of "slipperiness", ie, sufficient friction to prevent slippage when a big transient tries to slow the platter. This is tricky to get right. The biggest risk of slippage is at the very short belt/capstan interface. The belt/platter interface is long enough to make slippage less likely there. The optimum material is very much table dependent. On a Teres motor capstan materials like string, thread, etc. result in quite audible slippage, though they're still better than anything elastic. Mylar will also slip if it's too thick and/or too smooth to make the curve around the capstan, and that's a function of your capstan diameter, material and surface.
Not much of an answer, sorry. But I encourage you to experiment. It can't hurt and you'll certainly learn something from everything you try, as we have.