Paul and Sabrina's Cheap 3 Phase Inverter (AC Controller) with Field Oriented Control

[MPaulHolmes]

Holy crap! Finland 1, USA 0. THank you!!!! May Moomintroll Live forever!

nlc, I should have said I understand the process for permanent magnet synchronous too. haha. The space vector modulation was always confusing to me until a couple days ago, but I still need to understand exactly how they compute the rotor flux angle in the ACIM case. My understanding is that if you command too much torque things will get jacked up. that's my technical understanding for ACIM at this point.

[nlc]

Yesterday evening I read the ST documentation of their V2.0 library for ACIM motor. FOC principle is the same as PMAC, the only diffrence comes from the software block which manage the rotor flux position : for PMAC the rotor flux position can be established with hall, encoder or estimated with a sensorless algorithm. In ACIM, the rotor flux position is estimated with a software block which use the measured rotor speed, the measured Iqs and Ids (from measured phase current and Clarke+Park tranformation), and motor characteristics.

Thus it seems that for ACIM it's very important to have a perfect knowledge of motor characteristics. For example, the IFOC ST library for ACIM needs :

#define ROTOR_TIME_CONSTANT

Define here (in μs), the rotor open circuit time constant of the motor τ r:

τ r = Lr/rr = (Lm + Llr) / rr

#define POLE_PAIR_NUM

Define here the stator winding pole pair number;

#define RATED_FREQ

Define here (in 0.1Hz) the right-hand boundary of the constant torque region (see
Figure 6): in that region we have rated current, rated flux, rated torque, rated power

#define NOMINAL_FLUX
Define here the required magnetizing current im (positive, peak value), expressed in q1.15 format

#define NOMINAL_TORQUE
Defines the maximum value for the motor’s rated torque expressed in q1.15 format.

#define FLUX_REFERENCE_TABLE: this look-up table (256 signed 16-bit values)
provides reference values of current ids (expressed in q1.15 format), according to
increasing stator frequencies (see Section 4.5.4 on page 70);
TORQUE_REFERENCE_TABLE: this look-up table (256 signed 16-bit values) provides saturation values of current iqs (expressed in q1.15 format), according to increasing stator frequencies (see Section 4.5.4 on page 70).

[TadhBijou]

Hello, I must say I am happy to see that this project is going to get to the production phase soon.
Thank you to all of you who contributed your knowledge into it.

This thread is really inspiring to me and I will soon start my electric car modding project.

Any way, I wanted just to say hello to every body and big thanks, especially to MPaulHolmes and nlc.

[shenta]

Great Job done by NLC, I am working also on STM32 with Hub Motor, but with FOC3.0 Library. My current wave form is some what more different than ur, and it swin like crazy on some angle... Again it may be problem of my motor, I will post some picture soon.

[nlc]

I work with FOC library V3.0 too. Don't hesitate to post screenshot, now I know in detail how this library works.

First step is to have a screenshot of backemf(scope between 2 phases and you turn the motor by hand), if it is not a true sinus the FOC algorithm cannot work correctly.

[MPaulHolmes]

I just got 300v worth of Sealed lead acid batteries. That should allow for 220v AC or something like that. They should be here some time next week. 25 12v batteries. ya! Also, Rod Hower from the EVTech list is sending me a bunch of TO-247 igbts to blow up. I'm going to do a small prototype before scaling it up. Since I'm broke, the power board will be something done on the mill. The plan is to do a driver board for the low side, driver board for high side, and then control board. low side on bottom, high side stacked on that, and then control on top of that. There's a pcb company here that will do $33 per board with no setup fee, so I'm going to use them. Just one board of each one, so no screw ups allowed!

[sawickm]

Originally Posted by MPaulHolmes

The plan is to do a driver board for the low side, driver board for high side, and then control board. low side on bottom, high side stacked on that, and then control on top of that.

Hi Paul,

I sounds like your AC Controller will be packaged similar to your new DC Controller (your PCB stacking concept). Are the driver PCBs going to be compatible for the AC & DC Controllers? Could you use the same driver PCB for the AC & DC Controllers? Or are they completely different circuits? Is it compatible with IGBTs or MOSFETs or both?

How is the testing of your new DC Driver PCB going

-Mark

[MPaulHolmes]

I haven't tested the new DC driver because of money. There were unexpected expenses in Kenya with our daughter at her new high school. The drivers are identical except that the AC driver goes from -8v up to 15v, and requires a 24v supply. Even though the new controller was going to be for Ben, a guy sent me some money, basically bribing me to let him try it! haha. I figured that it's better to get it tested than not, so I ordered the rest of the parts and will be building it in the next week or 2. After trying this approach, which should work fine, I'm going to switch to a high side drive and try that out.

[skeeterb]

I've been lurking in this thread for a few days, reading the 678 posts. It's looking good Paul. Hope you and the others can get a working model finalized. I like the idea of being able to regen on braking.

[MPaulHolmes]

The software for ACIM is SOOOOOOOOO close to being done, and is done in the case of a permanent magnet synchronous motor. Just another day or so. And everything is correct this time. Mathematically right. Nothing is based on taking other code and adapting it. It's all based on deriving each formula from scratch myself and coding that. It makes sense and will work.

I thought I was heading off on a slower way of doing the space vector modulation than the Microchip code, but it turned out to be probably half the steps of theirs, but slightly longer. It just kept simplifying.