The assumption is made due to:
1. JXD advertising the device as 1.6 GHz.
2. RockChip advertising the chip as stable up to 2 GHz.
3. The fact that every time the device boots with the original firmware, it tries to reach 1.6 GHz and fails.
All this leads us to believe that the second batch is the broken one.
Also, you're correct. On most devices, EMULATION wouldn't suffer a huge hit, with the exception being PPSSPP due to it not being optimized for multicore. But, higher end Android device already do suffer a hit, which will be compounded by the fact the SoC has a lower maximum frequency.
Also, when the S7800B was first released, there were many problems due to the faulty recharger JXD supplied. This is still true, but people have published workarounds and willgoo has supplied replacements - so you hear less and less about it. Also, the light leakage was so easily solved it shouldn't even be an issue, merely an oversight (it's loosening 4 screws that were overtightened).
The most notable issues were the faulty touchscreen and controls on JoeSixPack's unit, a unit with a missing back camera, and a unit that was fried due to the charger. 3 units out of thousands shipped. Crashes were rare, and I don't remember hearing about one.
So yeah. The S7800B's launch wasn't perfect (add it to the 2 month delay), but I'd rather have a light leaking device with a faulty charger than one with voltage leaks - at least I can take care of the first one.