Over the last couple of weeks I have been playing with two models of the Pandora's Box 3. I initially purchased one of the newer 'bootleg' models, had some problems with it and stumbled upon this thread while doing research/troubleshooting. Thanks to this thread I decided to look for the original PS4-styled version and found a cheap deal on eBay.
Having used both now I've been able to draw some comparisons. Interestingly, the PS4 style model runs pretty smoothly with no framerate stuttering on most of the games. However, this model has some nasty screen tearing. On the second model, I noticed very little to no screen tearing whatsoever. However, it seems like every 15 seconds or so the framerate dips. So, it will run at a nice 60fps for a few moments, then it dips, and then it returns to normal. It constantly does this--a few times a minute at least--but it doesn't appear to be related to anything happening on screen (it will occur regardless the amount of action on screen). A friend and I were thinking about it and our best guess is that the second model has some form of forced V-Sync enabled, which eliminates the screen tearing but causes the games to stutter instead. It's a shame there's no way to disable it or tweak the emulator.
One massive advantage the PS4 model has over the later one is that it has a headphone jack. I'm thinking about hooking up a cheap pair of speakers to it so I can easily adjust the volume from game to game. Or, maybe I will run a pair of headphones with external volume control out of the cab. The volume leveling in each of the games is all over the place, so having a headphone jack is a very nice option.
On a side note, my "problem" with the bootleg version was that it didn't play nice with my monitor. After a few minutes of play, I'd get some
crazy vertical rolling. It turns out my monitor just needed to be adjusted, despite being perfectly fine with the Final Fight PCB the JAMMA harness was previously connected to. Something to keep in mind for anyone that has screen issues with these.