| ??? 03/13/08 02:01 Read: times |
#152201 - Not right Responding to: ???'s previous message |
(float)SR = ((S1 + S2)/0x02); // wrong
SR = (float)((S1 + S2)/0x02); // do math then convert to a float SR = (((float)S1 + (float)S2)/0x02); // do math as float SR should be a 16 or 32 bit number. On an 8052 float is big and slow. Plus you get float rounding errors as a bonus. (ie 1 / 3 * 3 may not equal 1). If you need speed 16 variables may be the answer. Or you can add straight out of the A2D. S1 to 16 can be 8 bit but to get 16 bit addition you must cast unsigned int SR; unsigned int S1,S2; SR = (unsigned int)S1 + S2; someone will correct me if I am wrong, but I think you only need to cast one of them |
| Topic | Author | Date |
| Math not functioning with proper headers? | 01/01/70 00:00 | |
| horrible method | 01/01/70 00:00 | |
| Lots of issues | 01/01/70 00:00 | |
| Code Op at level 8 | 01/01/70 00:00 | |
| Try multiple steps | 01/01/70 00:00 | |
| ah hah moment arrived......omg | 01/01/70 00:00 | |
| 01/01/70 00:00 | ||
| Here is the block which now works 100% | 01/01/70 00:00 | |
| Operation question in C | 01/01/70 00:00 | |
| Because | 01/01/70 00:00 | |
| Simplify | 01/01/70 00:00 | |
| what's the point ? | 01/01/70 00:00 | |
| What's the point | 01/01/70 00:00 | |
| simpler, but | 01/01/70 00:00 | |
| Not right | 01/01/70 00:00 | |
| use shift right instead of divide | 01/01/70 00:00 | |
| If you're lucky... | 01/01/70 00:00 | |
| Use a rount trip buffer | 01/01/70 00:00 | |
| Bad names | 01/01/70 00:00 | |
| System use names ???? | 01/01/70 00:00 | |
| Common naming convention ? | 01/01/70 00:00 | |
| ISO/IEC 9899:1990... | 01/01/70 00:00 | |
| Comments | 01/01/70 00:00 | |
| same effect as a "circular array" | 01/01/70 00:00 | |
| think I like this the best | 01/01/70 00:00 | |
Use a loop with deglitching and averaging | 01/01/70 00:00 |