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

[thingstodo]

Originally Posted by freebeard

Will this work with polyphase DC?

I think so.

What is polyphase DC? I picture AC sitting on top of a DC bias.

As I understand it, Polyphase DC is the same as 'brushless DC', where the DC is switched on and off in square waves instead of pulses. Turn on 'phase A' and the poles of the stator push the magnets on the rotor to rotate. Then you turn off 'phase A' and turn on 'phase B' to continue the rotation. Switching speed controls rotational speed as long as you have enough torque to rotate.

There are ways to micro-step, where phase A is pulsed, so that the rotation is smoother at low speeds.

Although more phases are possible, I've only ever seen 3.

Reverse works, right? Is regen in this release.

I'll be using reverse on the transmission so I don't think I asked that yet.

Regen is asking the motor to turn slower than it is, so the motor 'generates' back to the bus instead of taking energy from the bus in 'motor' mode. So I am confident that the regen will work.

How many are in the field so far.

I think Paul answered that one ... Hmmm .. I guess Paul answered the rest as well ..

As soon as Paul has run his tests, hopefully I get one to test out.

[freebeard]

Thanks. I hope that Paul doesn't take to disliking me. I hadn't forgotten Post #1062. The confusion set in because there's also a DC controller thread (as you know) and I was trying to follow along.

Apparently the DC controller has no application to Toyota Synergy Drive components. Which is why E*clipse is working on the resolver details.

Why 3-phase

My current (ha!) understanding is that DC with a commutator is like greyhound chasing a mechanical rabbit, where polyphase AC is (thanks to the magic of the Right-hand rule) an electrical vortex that drags the rotor arond with it. That leaves brushless DC and single-phase AC for me to 'brush' up on.

3-phase DC is sounding like Selsyn (synchro) motors with the master simulated in software.

Reverse would be just switching A and B, right?

Edit: e*clipse -- I looked at the page I linked to to define selsyn (self synchonizing) motors and about 80% of the way down the page there is a section on Resolvers. Quoting:

It is possible to derive resolver-like quadrature angular components from a synchro transmitter by using a Scott-T transformer. The three TX outputs, 3-phases, are processed by a Scott-T transformer into a pair of quadrature components. See Scott-T chapter 9 for details.

There is also a linear version of the resolver known as an inductosyn. The rotary version of the inductosyn has a finer resolution than a resolver.

Maybe something useful there?

[thingstodo]

Originally Posted by freebeard

My current (ha!) understanding is that DC with a commutator is like greyhound chasing a mechanical rabbit, where polyphase AC is (thanks to the magic of the Right-hand rule) an electrical vortex that drags the rotor arond with it. That leaves brushless DC and single-phase AC for me to 'brush' up on.

If you take the polyphase AC and replace the AC rotor (laminated steel with conducting bars) with a magnetic rotor (one pair of magnets for each 'pole') .. you have a brushelss DC motor or polyphase DC.

Single phase AC uses a capacitor circuit to offset the single phase at low rpm and generate torque, or uses a second coil inside the stator to perform the same function.

I *HIGHLY* recommend howstuffworks.com for going through the basics ... well illustrated and well described (unlike my descriptions)

[Quezacotl]

Originally Posted by MPaulHolmes

Just give me your email, and I'll send it to you. Mine is paulandsabrinasevstuff@gmail.com.

I sent you mail. marko@mthtek.net

[MPaulHolmes]

I found several bugs in the code. It's still not quite right, but is improving. It's easy to lose sync at the moment. However, here's a test with the 480VAC motor running from a 48v DC bus.

https://www.youtube.com/watch?v=5bRs...ature=youtu.be

[e*clipse]

Yes, basically the "brushless dc," "permanent magnet synchronized motor," etc etc - there is no standard - are mostly 3 phase motors. Your greyhound chasing the magnetic field description works pretty well. By using 3 phases, one can create a smoothly rotating magnetic field. Of course more phases could also do it, buy why bother with all the extra hardware??

The Toyota Synergy drive is basically a "buried permanent magnet synchronized motor" where they optimized the affect of the reluctance torque created by the shape of the steel laminations used to contain the permanent magnets. Look into "switched reluctance" motors to get a feel for what I'm attempting to describe. The buried permanent magnet design has the additional benefit that the magnets are held in place by the rotor laminations. The problem of keeping the magnets from flying off the rotor at high speed has been eliminated.

By optimizing the relationship between the reluctance torque and the direct torque created by the magnetic field and the neodynium magnets, Toyota was able to reduce the BEMF at high speed, which allows a lower DC bus voltage.

That is why the FOC feature of this controller is so important for the Synergy drive, vs a standard brushless dc drive. One CAN control that crucial relationship for much better performance.

Regarding selfsyn/resolvers - I suppose there is a completely analog way to do what I'm doing, which is a hybrid between digital and analog. I'm providing the carrier frequency for the resolver and filtering the resolver's output so the digital AtoD converter can do something useful with it. This may be possible with completely analog circuitry, but I'm not that good. It would take a really experienced, wise, old-school EE to figure that one out. Oh, yea - and the controller is digital, so eventually the analog signal will have to be converted to digital for the controller to do anything useful with it. As far as the resolution of the resolver, I have NO idea what he's talking about. The resolver is an analog device, and from that perspective all it's "resolution" is dependant on the quality of the carrier frequency and decoding.

- E*clipse

Originally Posted by freebeard

Apparently the DC controller has no application to Toyota Synergy Drive components. Which is why E*clipse is working on the resolver details.

Reverse would be just switching A and B, right?

Edit: e*clipse -- I looked at the page I linked to to define selsyn (self synchonizing) motors and about 80% of the way down the page there is a section on Resolvers. Quoting:

Maybe something useful there?

[e*clipse]

Oh yea...

I've almost crawled out of the Unicorn Corral....

I did manage to get DesignSpark working, and I'm currently untangling the rat's nest to make a PC board.

I did find some AWESOME IGBT's to drive this thing with.

A few Silicon Carbide IGBT's in the 1200V > 1700V range are becoming available. Cree makes a half bridge that looks like it would directly connect to this driver board! Also, GeneSiC is making some individual switch moduels for the SOT-227 format. All these are in the 100A to 300A range, and include a SiC diode.

This is really exciting. The SiC losses are about as low as I can imagine, and they seem to get more efficient as you push them to higher switching frequencies. Cree is talking about 99% efficient. You almost don't need a heat sink.

- E*clipse

[e*clipse]

Hey Paul!

Cool to see the motor running. This is REALLY helping me figure out how to (safely, slowly) get the BLDC motor running.

What (exactly) connector did you use for the QEI port? I'd like to make my board directly connect, no hassles.

I'm going to use the extra (index) pin for the motor temperature. The 'Yota motor has a second two pin connector for the motor's temperature sensor. It will probably need 12V - I have to figure that out.

- E*clipse

[e*clipse]

I would just like to do a little post giving an update of what's happening with this project.

First, and most important, Paul is in an epic debugging battle with field oriented control and an AC induction motor. Induction motors are voodoo to me; I cannot even imagine how Tesla came up with one in the first place.

I am glad it's not me. I am also VERY grateful of the experienced folks on the EV tech list who are helping with this. It's a huge deal, and their contribution and help with this cannot be overstated.

For my little part, I'm trying to cram all the components onto a circuit board approximately 1" by 3". It's not that I actually like tiny stuff - it just has to clear other stuff that already exists on the motor control board. It's tempting to use smaller components as a solution, however, I've worked with TSSOP components (before I new better) and they are a PIA for anyone to hand-place/solder. I have had to compromise, and the board will be entirely surface mount parts. I'm using SOIC IC's and 0805 resistors and capacitors. That will be enough of a challenge for my poor eyesight and clumsy paws...

- E*clipse

[MPaulHolmes]

i FORGOT! hERE'S THE 5 PIN CONNECTOR FOR ENCODER:

171856-1005


from mouser.


back to screaming kids now!! haha