Molding around a bearing

I am looking for advice on molding a nylon cover to a bearing. I have been, up until this point, molding the nylon roller cover separately and then press fitting the bearing to the cover. We have recently had a few issues with the nylon cover expanding too much over time and then separating from the bearing.
We then decided to amend the tool so that we could mold the cover directly to the bearing. the issue we now face is that the nylon is getting into the bearing. We have changed the bearing from a ZZ (metal cover both sides) to a 2RS (rubber seal both sides). the nylon is still squeezing in past the rubber seal.
Is there anyone who has experience with molding directly to a bearing? What could we do to alleviate this issue?

Silly question, If the original part is press fit, then why are you molding over the shield ? shouldn't the tool be designed to seal on the outer race of the bearing because there is no way a thin pressed in seal can handle any type of high injection pressure

The problem we were having was that the nylon was expanding over time. Initially the bearing press fitted to the cover is very tight and virtually impossible to remove by hand. After a few days the cover becomes loose and you are able to push it off the bearing with very little force. We have been using the same mold for 7 years and haven't had any issues. We then changed to a different injection molder. This is when we started having the issues with the caps expanding. We are using the same material. We have tried the molding at different temperatures and different speeds. We still had the same issues.
We then decided to try and amend the mold to take the bearing in the injection process. You are correct - there is no way that the bearing covers can handle the pressure. I was wondering of there anyone who has done this sort of thing before?

We have a four cavity bearing/POM overmold project. Like ijchan223 said, you need to seal off on the outer race, not the seal. Our overmold comes down a few thousandths around the race to keep the bearing in place.

We ran nylon wheel bearing parts that required press fit assembly in different configurations including what you are describing. We never found a successful process that would produce consistent parts after final assembly. Then we added an annealing after molding. The annealing process relieved molded in stresses, open the process window and solved the problems with failures after assembly. That maybe the step the previous molder was doing and it didn't get pass onto the new molder. Dupont plastics has really good information for annealing nylon materials in their design guides.