| ??? 02/04/11 11:01 Modified: 02/04/11 11:06 Read: times |
#180999 - hints Responding to: ???'s previous message |
you have to design the hardware carefully so that it helps your software.
Since the PIC has 2 PWM channels you have to use them to free the software overhead. You have to rethink your algorithm so that not to produce clipping, how to do that: do lots of PWM lookup tables, lots of them stored in program memory. Another important issue is not to use zero crossing. Since you generate the PWM cycles, you know exactly where you are in the sine wave in real time. Use modern PIC18F2520 running at 10MHz with 4x PLL.It is pin for pin compatible and has more resources and double the speed. I used one TMR0 interrupt at 50usec interval. I also did load detection, if load is not present inverter is turned off to save battery, if load comes back on, inverter turns on again. Also I did battery monitoring and temperature detection with led alarms and audio alarms. I charged my customer 7500$ for giving them the hex code. Good luck. Mahmood |
| Topic | Author | Date |
| How to measure a clipped and non-clipped sinewave accurately | 01/01/70 00:00 | |
| peak voltages does not represent RMS if distorted curve | 01/01/70 00:00 | |
| RE: How to measure a clipped and non-clipped sinewave | 01/01/70 00:00 | |
| Still can't look at a single point in time | 01/01/70 00:00 | |
| RE: Still can't look at a single point in time | 01/01/70 00:00 | |
| peak-to-RMS relation In general | 01/01/70 00:00 | |
| I have posted this thread on some other forum | 01/01/70 00:00 | |
| Sine feedback for inverter | 01/01/70 00:00 | |
| Absolutely | 01/01/70 00:00 | |
| Re: Sine feedback for inverter | 01/01/70 00:00 | |
| hints | 01/01/70 00:00 | |
| Waveform comparison | 01/01/70 00:00 | |
| What if? | 01/01/70 00:00 | |
Drive and table | 01/01/70 00:00 | |
| Re: Waveform comparison | 01/01/70 00:00 |