Like many B owners, I have had the seat belt problem since day one.
Every time when leaving car, it was needed to manually guide the belt into the reel because the belt would not retract itself fully. There are actually two design flaws: the reel spring is too weak, not providing enough force to pull the final part of the belt, when reel is almost full (I've changed the reel on the driver side, but there was no effect, so it's by design). Also the belt is making 90 degree turn and also goes thru velcro strap, all this is adding to pulling resistance that the reel is just too weak to overcome.
Having read all forum posts and not being convinced by advice to apply silicon oil to the belt, I started looking for the root cause of this problem.
It appeared that although belt can move freely thru the guiding slot on the 90 degree turn, still quite significant force is required to actually pull the belt through, towards the reel. It feels like the belt gets "caught" (kind of) in the place of the turn, creating resistance more than the reel can handle.
Further inspection showed, that the slot the belt passes thru before the actual turn, has small "humps" inside that make the slot narrower. These humps cause the belt to "lean" against the other side of the slot (because the belt is turning 90 degrees) and this creates the catching effect.
The rest was easy, I took a narrow stanley knife and cut off the humps on the inner surface of the slot. Also I cut a bit wider the inner edge of the opposing side of the slot, to even more reduce the catching effect.
Then after some sandpapering (120, 800 and 1600) I reinstalled the guiding slot back to the car and bingo - the belt moves without a hitch, retracting fully back into the reel, like in any other car.
Here's the guiding slot before and after the procedure:
Every time when leaving car, it was needed to manually guide the belt into the reel because the belt would not retract itself fully. There are actually two design flaws: the reel spring is too weak, not providing enough force to pull the final part of the belt, when reel is almost full (I've changed the reel on the driver side, but there was no effect, so it's by design). Also the belt is making 90 degree turn and also goes thru velcro strap, all this is adding to pulling resistance that the reel is just too weak to overcome.
Having read all forum posts and not being convinced by advice to apply silicon oil to the belt, I started looking for the root cause of this problem.
It appeared that although belt can move freely thru the guiding slot on the 90 degree turn, still quite significant force is required to actually pull the belt through, towards the reel. It feels like the belt gets "caught" (kind of) in the place of the turn, creating resistance more than the reel can handle.
Further inspection showed, that the slot the belt passes thru before the actual turn, has small "humps" inside that make the slot narrower. These humps cause the belt to "lean" against the other side of the slot (because the belt is turning 90 degrees) and this creates the catching effect.
The rest was easy, I took a narrow stanley knife and cut off the humps on the inner surface of the slot. Also I cut a bit wider the inner edge of the opposing side of the slot, to even more reduce the catching effect.
Then after some sandpapering (120, 800 and 1600) I reinstalled the guiding slot back to the car and bingo - the belt moves without a hitch, retracting fully back into the reel, like in any other car.
Here's the guiding slot before and after the procedure:
