Spiral ramp bug

I have made a spiral ramp for one of my houses, its a partially hollow cylinder which is cut to leave a sliver, then the top is twisted to form a ramp that spirals up around 180 degrees.

the ramp is set up where I can walk up it to go upstairs or under it to get to the balcony.

see photo: http://www.pbase.com/cptinrn/image/54675310

The ramp works fine when I first create it. But it eventually converts into an invisible solid that can not be walked under or up. When this happens the slice/cut dimensions are as follows.

begin 1.00 end 0.15 (I might be off on begin cut, I will verify when I get in world)

The defective ramp changes back to a proper ramp if I modify the last slice number to make it thicker, like 0.30. Ok I can live with that but what blows my mind is after a little time (few days) the ramp changes its dimensions back to the 0.15 and becomes an invisible solid again. This has happened 3 different times on 3 different houses, all in the same sim.

My partner said she believed this was a known bug. I would like to understand it better so I can avoid having to fix a bunch of broken houses. Any information would be appreciated. I can probably provide anyone who wants it and example of the defectivce ramp, I believe I saved one.

Sadly another old bug ~.~ been around awhile, only way that sometimes fix it , copy the one bugged or try remaking one :/

if at all possible, reverse the direction of the twist, this bug only happens on clockwise twist. The issue has to do with the mass functions SL uses, this is actualy a bug fix first released in 1.5. The issue is that the direction of twist effects mass in a non symetic way. Instead of LL fixing this bug, at this point in time they were less interested in fixing the sim crash issue and just spackling over it. You see the the twist could result in the shape becoming a negitive mass. Negitive mass objects on the whole aren't bad just the make things interesting. LL's approuch to fixing negitive mass objects was two fold. If the shape was hollow and it was the hollow face that was resulting in it being negitive mass, then don't apply the hollow subtraction. If the shape was negitive mass before the hollow subtraction, replace the entire bounding box with a basic cube that encompasses the prim entirely.

If you want LL to fix this, the mass calculations for twist need to be fixed.

At the time i had hounded LL to fix the mass functions, as i had created a variety of inside out prims, that by default had negitive mass.

Threadtunnelled to: /8/be/88467/1.html