I may be wrong, but didn't all dinosaurs evolve from bipedal or semi-bipedal archosaur ancestors, with reversion to a full quadrapedal stance in the large herbivores a secondary adaptation?
Yes, seems to have happened at least four times independently, with Sauropods, the common ancestor of Stegosaurs/Ankylosaurs, Ceratopsians, and Hadrosaurs.
Were there any predatory dinosaurs that were not bipeds?
No
Why not? If quadrapedalism provided a better frame for predation because of balance, tripping, etc., why didn't some of the readadpted quadrapeds become predators in the Mesozoic?
If you're asking why theropods didn't become quadrupeds, I already answered that - they couldn't turn their palms to face the ground.
If your asking why the "herbivorous" dinosaurs didn't evolve carnivory - well, some of them probably were somewhat omnivorous actually. But the strict hunter niches were already occupied by theropods, and barring mass extinctions, something like a total trophic switch from herbivory to carnivory is uncommon (the reverse is actually very common - there were probably four or five groups of at least partially herbivorous theropods).
Is it possible that bipedal theropods with forearms freed from locomotive needs and able to evolve into full-time grasping/slashing claws or wings outweigh the balance/tripping issues?
You shouldn't overvalue arms just because they're so central to us. I'm not aware of a single large carnivorous theropod which could even bring its hands to its mouth for example. Now, some taxa did have heavily muscled arms, like Allosaurus, and probably did use their arms to hold onto struggling prey while their head did the dirty work. But others, like Carnotaurus, basically lacked functional forearms entirely, and did just fine.
As for smaller species, for the most part they probably folded their long arms tightly when they ran, like birds, and you should remember that even larger "raptors" probably had feathers all the way down their fingers (a birds primary feathers actually *begin* at the wrist).
You're right insofar as once the arms were freed of locomotion, they could do any number of things, and were by default used to grasp. But this is just a classic example of a trait being re-adapted by natural selection. The grasping hand was an effect of being bipedal basically, not the cause.