| ??? 08/13/11 18:06 Read: times |
#183354 - Even Better.... Responding to: ???'s previous message |
Andy Neil said:
Maarten Brock said:
MOV SBUF, A ; <=== It is better to check before sending instead of waiting H_1: JNB TI, H_1 ; <=== for the transmission to complete. This way you can do other CLR TI ; <=== things in the meantime. RET END In fact, it is essential to check before sending - otherwise you risk corrupting the byte currently being transmitted! Even better is to put each byte being sent into say an 8 byte FIFO circular queue. Then let the occurrence of an interrupt from the TI bit getting set cause the next byte to be drawn out of the queue and get loaded into SBUF. This way only when you are sending more bytes than the size of the queue do you end up spinning in a polling loop wasting time that could be used for other activities. Short transmits in bursts can benefit greatly with this scheme. You pay a small amount of overhead checking and managing the index pointers for the queue but it is normally a small overhead compared to what you gain in being able to let a multifaceted application run on the MCU. Michael Karas |
| Topic | Author | Date |
| need help to display timer count on hyperterminal and LED'S | 01/01/70 00:00 | |
| Attention to Details | 01/01/70 00:00 | |
| divide by 10 | 01/01/70 00:00 | |
| Hundredth digit is quotient of division by 100 | 01/01/70 00:00 | |
| thanks | 01/01/70 00:00 | |
| Debugged with pen and paper? | 01/01/70 00:00 | |
| Debugged | 01/01/70 00:00 | |
| Decimal or hexadecimal? | 01/01/70 00:00 | |
| decision | 01/01/70 00:00 | |
| So take up the challenge then | 01/01/70 00:00 | |
| agreed | 01/01/70 00:00 | |
| Actually BCD!! | 01/01/70 00:00 | |
| Oh Details Details | 01/01/70 00:00 | |
| update | 01/01/70 00:00 | |
| Issues | 01/01/70 00:00 | |
| better to check before sending | 01/01/70 00:00 | |
Even Better.... | 01/01/70 00:00 |