The E3 and E2 have clamping diodes on the serial port's input and output pins. As such anything above 5V or below 0V is effectively shorted out. This is meant for ESD protection but I suppose the diodes will effectively clamp +/-12V RS232 to 0- 5V TTL. So long as the PC doesn't supply too much current and warm the diodes up too much! John Chadd Cliff Lawson wrote:

I'd have thought the protection was vital. It's not so much whether the PC can detect the "weedy" 0V..5V swing coming out of an E2/E3 but whether the UART input buffers on the OMAP/Sharp can stand getting -12V shoved up them coming from the PC (I'm assuming that PCs still transmits true RS232 even if they're a bit more lenient about what they can detect?)

(Or maybe my h/w designing colleague did splash for sufficient inbound protection on the board - I can't remember off hand)