Here is my 3phase inverter project.....
| Yes, this thing does work ( just about! ;-) I built it about 2 years ago, when a group of EV Listers set up an AC motor controller group on Deja News. Unfortunately this group died, and this was as far as I got. AFAIK, Rich Rudman was the only one to get all the way to a working, useable AC drive ( see hisMinibike from Hell ). However, since this new group ( EVTech ) has been set up, I think I'll develop the ideas further, but with a different CPU..... so for the sake of posterity, here's the PIC-based inverter..... |
| There is 1 processor module ( monster.gif ), and 3 phase-driving modules (driver.gif ). The processor module uses a PIC 17C756, running at 28 MHz. This CPU has a high-speed RISC core, 3 PWM modules, UART and single-cycle hardware multiplier block, as well as some other goodies. But, it's only an 8-bit machine, which limits how much intensive maths it can do. Also, I don't have a C compiler for it, so I had to hand-craft the code in assembler, which was a nightmare. |
| Currently, only raw open-loop V/F control is implemented, speed feedback will come later! Oh yes, not shown on the schematic is a nice 2 line LCD module, which displays useful stuff like the AC frequency that's being output, and the percentage power |
| This circuitry has successfully driven a modified 3 phase 750 watt induction motor that I salvaged. I rewound the motor with thicker wire to lower its rated voltage and increase the current. I managed to get the motor spinning at somewhere up around 5000 or 6000 rpm. There was good torque available from the gearbox that was attached ( 11 to 1 reduction ), until I tried a stall test and blew up the power stage! I have since rebuilt the power stage, so it still works. |
| CAVEATS: |
|
This circuit has NO deadtime generation, NO current limiting or desaturation detection. |
| It's easy to blow up the power stage because of this. The new improved version will use different gate drivers ( the HCPL-316J ) which will provide fault detection and feedback to the CPU. |
| Reset circuitry for the CPU is not shown - I just used a pullup resistor. In a real-world inverter, a good power-supply supervisor/watchdog chip should be used for this. Again, the new version will have this sort of stuff. |
|
|
| Above is the processor module. The 4053 chip selects either the high or low driver for each phase, this section is supposed to prevent simultaneous turn-on of both the high and low side drivers of a phase, which would short the motor power supply and let the smoke out! |
| In the software, the two IGBTs that control a phase were switched "off" during the low part of the PWM cycle, so the motor current would be carried just by the integral freewheel diodes. I don't know if it'd work better to have the phase switched either high or low but not floating, if you see what I mean.... |
| Below is one of the phase-driver modules, remember there are 3 of these. The "PSU" devices are small isolated 1 watt DC-DC converters. |
 |
| Vcc_bar in these circuits is the same regulated 5volts that supplies the CPU, the circuit takes about 1/2 amp continuous, with several amps surge during switch-on. |
| The +12v input supplies the low-side IGBT gate drivers, this must be referenced to the "Negative" power connection of the inverter. On my prototype, I used a 2 watt isolated DC-DC converter to generate this, so my power stage is totally isolated from the processor stage. I built it this way because I ran out of the small DC:DCs used for the high-side drivers ( scrounged from old PCBs in work! ). |
| In the new version, there will be six identical IGBT driver pcbs, with isolated supplies for driving the IGBTs. |
| Back to my Electric Mini Pickup...http://homepage.ntlworld.com/electric.mini/ |