| Frieder Ferlemann 01/18/10 05:29 Read: 249 times Old Europe |
#172543 - PWM IRQ code in C, lookup table, 32 byte Responding to: Michael Karas's previous message |
If P1 (or whatever port you connected the servos to) is free to use, you can also use a table lookup.
Thus you have no performance penalty within the IRQ for connecting up to 8 servos:)
#include <8052.h>
static unsigned char ms_tenth_countdown;
// dynamically setup table.
// Indexed top down.
// setup glitchlessly!
unsigned char __pdata table_P1_over_time[180 + 1];
void Timer0_ISR (void) interrupt 1 {
// assuming P1 is only used for servos.
P1 = table_P1_over_time [ ms_tenth_countdown ];
TF0 = 0;
if (!--ms_tenth_countdown)
ms_tenth_countdown = 180;
}
void main(void){}
Now you'd be at:
0000 371 _Timer0_ISR:
0000 C0 E0 380 push acc
0002 C0 00 381 push ar0
0004 C0 D0 382 push psw
0006 75 D0 00 383 mov psw,#0x00
384 ; servo_lookup.c:13: P1 = table_P1_over_time [ ms_tenth_countdown ];
0009 E5*00 385 mov a,_ms_tenth_countdown
000B 24r00 386 add a,#_table_P1_over_time
000D F8 387 mov r0,a
000E E2 388 movx a,@r0
000F F5 90 389 mov _P1,a
390 ; servo_lookup.c:15: TF0 = 0;
0011 C2 8D 391 clr _TF0
392 ; servo_lookup.c:17: if (!--ms_tenth_countdown)
0013 D5*00 03 393 djnz _ms_tenth_countdown,00103$
394 ; servo_lookup.c:18: ms_tenth_countdown = 180;
0016 75*00 B4 395 mov _ms_tenth_countdown,#0xB4
0019 396 00103$:
0019 D0 D0 397 pop psw
001B D0 00 398 pop ar0
001D D0 E0 399 pop acc
001F 32 400 reti
I doubt that switching compilers would help much in this case. What's worse: coding directly in assembly code won't help much either:^) |
| Topic | Author | Date |
| Controlling multiple servo motors | Yash Chitalia | 01/17/10 14:47 |
| executing code takes time | Erik Malund | 01/17/10 15:27 |
| PWM code | Steve M. Taylor | 01/17/10 15:35 |
| Thanks...now, about the PCA timer.. | Yash Chitalia | 01/17/10 16:06 |
| If SDCC Then Keil | Michael Karas | 01/17/10 16:15 |
| PWM IRQ code in C, bitops, 45 byte | Frieder Ferlemann | 01/18/10 05:19 |
| Optimization having no effect. | Yash Chitalia | 01/18/10 11:36 |
| PWM IRQ code in C, lookup table, 32 byte | Frieder Ferlemann | 01/18/10 05:29 |
| I haven't understood the table lookup concept very well... | Yash Chitalia | 01/19/10 00:55 |
| table lookup | Frieder Ferlemann | 01/19/10 02:39 |
| PWM IRQ code in C, lookup table, 30 byte | Frieder Ferlemann | 01/19/10 03:07 |
| Compiler specifics | Andy Neil | 01/19/10 04:35 |
| absolutely-locating xdata at 256 byte boundary | Frieder Ferlemann | 01/19/10 05:00 |
| I see. | Andy Neil | 01/19/10 06:34 |
| An example code.. | Yash Chitalia | 01/20/10 02:20 |
| Keil: Specifics and non-specifics | Andy Neil | 01/20/10 02:50 |
| Not really working... | Yash Chitalia | 01/20/10 03:55 |
| Enough power? | Per Westermark | 01/20/10 09:45 |
| RE: running 32 servo motors | Andy Neil | 01/20/10 10:34 |
| The chip sounds perfect... | Yash Chitalia | 01/18/10 11:47 |
| not worth it | Erik Malund | 01/18/10 12:10 |
| Servo control will become difficult then... | Yash Chitalia | 01/18/10 12:42 |
| output | Andy Peters | 01/18/10 12:52 |
| HCT chips have ... | Erik Malund | 01/18/10 14:12 |
| Eval a concept with a single PCA and 3:8 HCT multiplexer? | Frieder Ferlemann | 01/18/10 14:36 |
| Check their website | Andy Neil | 01/18/10 14:44 |
| Here is the code... | Yash Chitalia | 01/18/10 15:28 |
| "sbit MTRS=P2^0;" is not OK here | Frieder Ferlemann | 01/18/10 16:11 |
| Keil syntax? | Andy Neil | 01/18/10 17:36 |
| "#define" not working.."sbit" works | Yash Chitalia | 01/18/10 23:32 |
| Yes, you were correct originally! | Andy Neil | 01/19/10 01:22 |
| Seconded | Andy Neil | 01/18/10 12:45 |
| Servo_0.1.zip | Murray R. Van Luyn | 01/18/10 15:02 |
links | James Hinnant | 02/18/10 21:10 |
| Eventually feed the guys at http://servomaster.sf.net ? | Frieder Ferlemann | 01/18/10 15:12 |
| servomaster.sf.net FLOSS | James Hinnant | 02/04/10 11:48 |
| Maybe of some use... | James Bowmaster | 02/10/10 14:40 |
| Whoops... found a bug | James Bowmaster | 02/10/10 16:35 |
| What bug? | Andy Neil | 02/10/10 16:41 |
| Bug | James Bowmaster | 02/10/10 18:47 |
| Improved version uploaded | James Bowmaster | 02/12/10 21:26 |



