| ??? 01/27/07 22:54 Modified: 01/27/07 23:04 Read: times |
#131606 - Well, that\'s a horse of a different color Responding to: ???'s previous message |
Erik Malund said:
Well, let's see ... you said 5kbits/half-millisecond, 10 bits per byte ... that's a 10 Mbps rate ... and it looks like asynchronous format to me ... unless you also generate a clock. Is that what you've done?
I did not say 10 bits per byte I said AND I QUOTE "let me "roughconvert" that to 1 byte per microsecond. If you need more precise, feel free to do the 'proper' arithmetic" That wasn't very clear, nor was it in any way apparent from your comments up to this point. Initially you said 5000 bits per half-millisecond. I was just giving you a break, since that meant that you don't have to fetch and isolate 20% of the bits. So, is it the case that you're sending both clock and data, each on separate lines, but still at your initially claimed rate of 5000 bits per half-millisecond, which is 10000 bits per millisecond, which is 10 Mbps, or is it 4000 bits per half- millisecond or 8000 bits per millisceond, which is 8 Mbps, as one would calculate from your "one byte per microsecond" revision? That's a 20% reduction, you know, and not a very close approximation of the initially claimed rate and why would you make such an assertion? ... What happened to the 10 Mbps rate you initially claimed? Since it's likely you have to set/reset the clock bit for each bit, how do you find time for that? That has to be done out of phase with the data transitions, else you can get into skew or metastability troubles. How do you avoid that? Where do you get 10 bits per byte if it's not async format? You said you're bit-banging. Are you clock-banging too?
if that is what you want to call it, then yes. px.y = data px.z = 1;clock high px.z = 0;clock low this is 6 total cycles = 32 cycles per 100us leaving 18 cycles for byte fetch, so, of course in your sloooow world it can not be done but with ~50 cycles every 100us it is possible with careful coding. Erik Now let me get this straight ... you send a constant 8000 bits per millisecond, from power-up to end of the universe or power fail, AND you do a bunch of other stuff, including processing interrupts from the internal UART at 480 kbps, that happen WHILE you're bit-banging? As for the careful coding ... how about showing us? Surely there's no secret involved in bit-banging clock and data. RE |
