??? 07/07/04 21:55 Read: times Msg Score: +1 +1 Good Answer/Helpful |
#73757 - RE: Hi speed serial commn revisited Responding to: ???'s previous message |
With a 11.0592 crystal and using serial interrupt , and a ring buffer for temporary storage ,is it likely that if the code size is big enough the serial interrupts come rapidly enough to overwrite the buffer ?
absolutely, just figure - you have ~100usec ~50 instructions per byte. If there is sufficient RAM avlbl on chip ( 89C51RD2xx ) , how does one optimise ( calculate ) the size of the ring buffer to avoid this ? I would always have a ring buffer large enough to hold one burst. Calculation can be quite hairy e.g. how often does a timer interrupt delay the process, e.g. usually the processing time for a length byte will not be the same as for a data byte etc When J Guy says that 18.432 MHz will give the same baud rate and be faster , what does it mean ? 115200 baud should be the same whether 18.432 or 11.0592 . Or is it just that he means the code will be processed faster so that buffer overwrite does not happen . so that buffer overwrite is less likely to happen. If a board is working well with a 12 MHz chip , will there be any problems if I pluck it out and replace with a 20 MHz , asuuming there are no timing implications on the software side ? usually no problem, but of this is not a (4 layer) PCB but components wired together, I would not trust it. Also things, like decoupling caps and the distance they are from the chips, become more important the higher the frequency. To run at 115k, I would, at least, use a 6 clocker at 18.432 MHz Erik |
Topic | Author | Date |
Hi speed serial commn revisited | 01/01/70 00:00 | |
RE: Hi speed serial commn revisited | 01/01/70 00:00 | |
RE: Hi speed serial commn revisited | 01/01/70 00:00 | |
RE: Hi speed serial commn revisited | 01/01/70 00:00 | |
RE: Hi speed serial commn revisited | 01/01/70 00:00 | |
RE: Hi speed serial commn revisited | 01/01/70 00:00 | |
RE: Hi speed serial commn revisited![]() | 01/01/70 00:00 |