I have a cusomer that wants me to furnish a bunch of, as he describes it, "complementary 5-volt PWM signals" over twisted-pair cables. Rep-rate will be around 100 KHz, so edge accuracy is reasonably important.
I may just be able to convince him to go RS485, but it is interesting to consider how to generate a real 5-volt complementary drive with as close as possible to zero skew between the positive and negative signals.
I vaguely remember an old TTL true/complement buffer. 265?
A pair of HC xor gates is superficially appealing, but takes a lot of parts.
John