??? 08/20/06 16:28 Modified: 08/20/06 16:30 Read: times |
#122661 - Yes, but is it a million or a dozen? Responding to: ???'s previous message |
I think you're being unreasonable.
I once designed and presided over the construction of a test system, using >500 single-chippers and a microcomputer to tell each of them when to start. It lived in a 6-foot rack and we had to use PLL's to ensure that every MCU had access to the same clock despite the fact this was essentially a single board. It cost a great deal of money to build that system and was a difficult task because of signal propagation issues. Those MCU's operated from a 5 MHz clock. The nominal distance, worst-case, from the most remote MCU to the control processor was >>10 feet if routed on the board, so we routed clocks and controls over a harness consisting of coax cables of the same length. Your 805x's will probably want to operate from something on the order of 12 MHz. If your communication is going to happen via the built-in serial port, then you must have the phase of the clock exactly identical at each MCU, at least to within a reasonable margin, say 10%, of that 12 MHz. That's 8 ns variation, worst case, over the entire are on which your hardware will be situated. How do you propose to manage that without first putting an upper limit on the number of MCU's? RE |
Topic | Author | Date |
Mulitple 8051 from a single clock source. | 01/01/70 00:00 | |
buffers and so on | 01/01/70 00:00 | |
A definite "maybe" | 01/01/70 00:00 | |
Thanks. | 01/01/70 00:00 | |
How many and how far apart? | 01/01/70 00:00 | |
As many as possible I guess. | 01/01/70 00:00 | |
As many as possible I guess. | 01/01/70 00:00 | |
Yes, but is it a million or a dozen? | 01/01/70 00:00 | |
Can not relay on timing. | 01/01/70 00:00 | |
If you want to know what's going on in each MCU | 01/01/70 00:00 | |
Possibilities | 01/01/70 00:00 | |
what's the problem? | 01/01/70 00:00 | |
systolic networks![]() | 01/01/70 00:00 |