On 06/09/2018 01:00, Jim Brain wrote: > > I have no idea how one would ever prove the design intent. Even if > the decap shows no special logic in the clock paths, you would still > argue that does not show design intent. > No. If the TxC clock always goes through the clock circuit that generates a clock from a crystal, then it doesn't matter too much if the chip designer thought about 115200 specifically. It would show that the chip was designed to use a crystal or clock interchangeably, with an optional divider & the guy who wrote the datasheet documented the use cases he thought of and not how the chip worked. If the /16 case tries to switch some part of the clock circuit off, because they assumed you don't want to generate a clock from a crystal, then it would be clear it wasn't intended to use external clock mode with a crystal. If this is the case then it would be nice to know how it still works & whether it's got any issues. I don't know which one is more likely, I find it interesting that the divider table would not be linear. 1, 2304, 1536, 1048, 856, 768, 384, 192, 96, 64, 48, 32, 24, 16, 12, 6Received on 2018-09-06 09:00:05
Archive generated by hypermail 2.2.0.