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

[thingstodo]

Originally Posted by P-hack

Interesting re: dual motors. To me the simplest is to just one set of igbts and one controller in slip mode and two motors in parallel. Read both motor speed signals and take the max speed for determining slip (prevents wheelspin runaway and encourages the slower one to speed up). Set 0% throttle to not regen and coast around the corners when you can. The latter fits well with "good" driving, traction pie habits anyway, you *shouldn't* be braking or accelerating much while turning and saving your traction for changing direction.

Interesting .. running AC motors in parallel while in FOC.

Extending that to 4 AC motors for AWD ... 4 speed sensors.

I'm not sure about using the maximum motor speed, though. If you use the maximum speed, you're going to have rated slip on the fastest motor, higher than rated slip on the slower turning motors. I think that gives the slowest turning motor the highest torque. So it will speed up, and the highest speed motor will also speed up? Not sure, just thinking out loud.

I think selecting the lowest motor speed for the slip is the way to go. It would effectively give you 1 wheel drive (on the slowest rotating wheel) during normal driving of the car, and if the slowest motor begins to lose traction, the other motors will begin to exert torque as the vehicle slows down just a bit. Or maybe I'm way off base?

[e*clipse]

Originally Posted by P-hack

Interesting re: dual motors. To me the simplest is to just one set of igbts and one controller in slip mode and two motors in parallel. Read both motor speed signals and take the max speed for determining slip (prevents wheelspin runaway and encourages the slower one to speed up). Set 0% throttle to not regen and coast around the corners when you can. The latter fits well with "good" driving, traction pie habits anyway, you *shouldn't* be braking or accelerating much while turning and saving your traction for changing direction.

What do you mean "slip mode"? If A motor like the one in your car unsyncs from its rotating magnetic field, it looses ALL torque. It might work fine for an induction motor, but definitely not for a BPMSM.

Also, there are plenty of "good" drivers who would disagree with you about when to brake or accelerate. You'll find them on most racecourses.

[P-hack]

Originally Posted by thingstodo

Interesting .. running AC motors in parallel while in FOC.

You would probably have to sacrifice FOC with one controller and multiple motors that aren't bolted together , which doesn't look like a deal breaker, since you can adjust your driving style (FOC seems most helpful in lighter loads AFAIKT and seems to lose any advantage at heavier loads, you can regain some starting torque with slip tuning).

Haven't completely thought through which sensor to use for slip, maybe average with individual limits? I don't know of a good way to make the slower wheel not have more torque than the faster, except to stop feeding the motors (or don't feed it much, just keep your speed up).

Originally Posted by e*clipse

What do you mean "slip mode"? If A motor like the one in your car unsyncs from its rotating magnetic field, it looses ALL torque.

I believe we are discussing induction motors here, right?!? slip is the relationship between the rotating magnetic field speed and the rotor speed.

Originally Posted by e*clipse

You'll find them on most racecourses.

By manageable, I mean that you may have to think about your driving style for the first day, not necessarily win races.

Remember, I'm talking about the simple approach that I can think of for multiple induction motors, not the fanciest thing you can imagine that will make your breakfast for you approach.

[thingstodo]

Originally Posted by P-hack

You would probably have to sacrifice FOC with one controller and multiple motors that aren't bolted together ,

With FOC you track frequency, but it does not appear to be part of the control strategy, per se, after you start rolling (maybe 3 or 4 Hz). As far as I can see, you control Id and Iq. The frequency is adjusted to generate the torque. You keep track of the frequency and use that to limit rpm.

which doesn't look like a deal breaker, since you can adjust your driving style (FOC seems most helpful in lighter loads AFAIKT and seems to lose any advantage at heavier loads, you can regain some starting torque with slip tuning).

Hmm. As I understand it, running without FOC (I call it Volts per Hertz, or V/Hz), is not as efficient. In FOC, the voltage used for a given frequency is a function of the torque required at that frequency. So the current required determines the voltage applied. Lightly loaded uses lower voltage. Heavily loaded boosts the voltage as required.

V/Hz is very simple but a bit wasteful. In the case of a constant torque load, like a car, you have a straight line from 0 speed to rated speed. An example helps. 1800 rpm rated at 60 Hz, 230V. 0 speed = 0V. 900 rpm = 115V. 1800 rpm = 230V. So if you are driving at 1800 rpm but are running at anything below maximum load (max load you should be accelerating) you are using too high a voltage and the motor has to dump this energy as heat.

If you are cruising at 60 mph, using 20 HP, and you have a motor that can do 30 HP at rated voltage .. you are wasting 50% of your battery capacity in the motor and you have to use even more energy to cool your motor and your controller.

FOC is quite important, IMHO. I know that our industrial controllers (VFDs or ASDs) - used at the day job - do not allow multiple motors for FOC. But they are not using encoder feedback. They are measuring back-EMF from the motor and determining rotor position with some space vector math.

If a workable theory - not the right name - workable set of rules? can be hammered out, I'd like to test multiple motors in parallel ( likely on a bench) and see if something can be empirically worked out to give all the motors some torque. Perhaps the slip of each motor would need to be characterized separately? We need to sum the required currents so that the controller knows what current, at what phase, to supply.

I have a couple of 5 HP, 230V, industrial motors that I can bench test with, should the opportunity arise.

Haven't completely thought through which sensor to use for slip, maybe average with individual limits? I don't know of a good way to make the slower wheel not have more torque than the faster, except to stop feeding the motors (or don't feed it much, just keep your speed up).

Keep thinking - maybe we can all work out something!

[P-hack]

I don't know at what point one controller makes sense though, maybe if you have a lead on 2 or 4 small motors that are just enough power for your needs (leads have to be the same length) and are on a tight budget. But if you have 4 regular motors, you are going to need so much battery that the extra power sections aren't gonna be noticed in the budget (and your controllers might need a controller, conceptually anyway, if you are trying to do stuff like traction control).

Originally Posted by thingstodo

.. They are measuring back-EMF from the motor and determining rotor position with some space vector math.

You have multiple rotor positions here (using current sensors), if you want FOC on multiple motors, you need multiple switches (IGBTs) and signals at a minimum so that the stator coils fire at just the right angle for each motor (though you still need some slip with FOC, I think it is mostly when changing speeds that it targets the highest torque vector). But a slip algorithm can also set slip for max torque, just not on the very first pulse.

[thingstodo]

Originally Posted by P-hack

... maybe if you have a lead on 2 or 4 small motors that are just enough power for your needs (leads have to be the same length) and are on a tight budget.

Fitting the available motors into the engine bay with the transmission, etc is sometimes a challenge. The Siemens motors that are surplus from Azure dynamics, for example.

With a small single-speed gearbox per motor, you may open up some options on different builds, or make the builds possible.

IMHO cost is always a factor, but not necessarily the over-riding factor. If something will do the job (safely, power is OK, etc) why not use it?

Why do the motor leads have to be the same length? At my day job, we are running 4 motors in parallel, each with a gearbox, for 4 wheel drive. There are separate overloads and contactors on the output side of the controller for each motor. Leads vary from 16 feet to the nearest motor to about 60 feet for the furthest motor on the opposite side. They've been running (intermittent duty) since 2006. They are only 3 HP, and the big iron thing they are moving (called a tripper) only moves 200 feet per minute ... but that's how fast it's supposed to move

There's not much access (to the motors), and my boss won't let me run tests on them. The controller is in V/Hz, no encoders or feedback to the controller. A DCS gets an encoder feedback from one wheel, so it can send out the stop and start commands.

But if you have 4 regular motors, you are going to need so much battery that the extra power sections aren't gonna be noticed in the budget (and your controllers might need a controller, conceptually anyway, if you are trying to do stuff like traction control).

If there are 4 motors, I won't need full power from any of them (I don't drag race). It would be nice to have one controller, but complexity COSTS MONEY in development time and troubleshooting PITA.

Multiple power stages on one controller would allow for separate control and be a bit more complicated in the one controller. Multiple controllers would need, in my opinion, to exchange a bunch of data to keep things running smoothly, perhaps with an overall controller as you suggest. But it sounds complicated.

You have multiple rotor positions here (using current sensors),

How do you do position sensing using current? That would be useful. As I understand it the commercial products use Back-EMF (voltage)

if you want FOC on multiple motors, you need multiple switches (IGBTs) and signals at a minimum so that the stator coils fire at just the right angle for each motor (though you still need some slip with FOC, I think it is mostly when changing speeds that it targets the highest torque vector).

Darn - I was hoping you had thought of a way to do FOC on multiple motors. Oh well.

But a slip algorithm can also set slip for max torque, just not on the very first pulse.

Not sure I understand that one. Slip for the motor, to generate torque in the ACIM, or slip between tires on multiple outputs?

[e*clipse]

Originally Posted by P-hack

I don't know at what point one controller makes sense though, maybe if you have a lead on 2 or 4 small motors that are just enough power for your needs (leads have to be the same length) and are on a tight budget. But if you have 4 regular motors, you are going to need so much battery that the extra power sections aren't gonna be noticed in the budget (and your controllers might need a controller, conceptually anyway, if you are trying to do stuff like traction control).



You have multiple rotor positions here (using current sensors), if you want FOC on multiple motors, you need multiple switches (IGBTs) and signals at a minimum so that the stator coils fire at just the right angle for each motor (though you still need some slip with FOC, I think it is mostly when changing speeds that it targets the highest torque vector). But a slip algorithm can also set slip for max torque, just not on the very first pulse.

This thread is about a FOC controller - that's why we're talking about it; it's not a "how to control induction machines as easily as possilbe" thread.

Again, I don't know what you're talking about when you say "regular motors." Given the variety of AC type motors available, the beauty of FOC is the software can be adapted to control quite different hardware. I am going to use this controller for a motor like the one in your Prius, not an induction machine. The control hardware is more or less the same.

Really, a well engineered system would use appropriate parts for ALL components. The reason to use small motors on all wheels is not merely to find a cheap motor solution. It can provide the opportunity for better control - that's my interest. Therefore, if one used 4 motors that are rated at 50kW, a 200kW single motor with gearboxes dividing the power to all 4 wheels would be an equivelent system.

Noting that the expensive stuff tends to be the batteries and the power electronics, the same thing applies. You still need to control 200kW, whether there are 4 sets of IGBT's in parallel controlling one motor or 4 sets of IGBT's controlling 4 motors. The cost of the control electronics is relatively small, and whether one uses FOC or simply VFD control algorithms doesn't effect the cost AT ALL. Same with the batteries. They need to supply 200kW. This could be split simply by running wires to individual motor controllers or running bigger wires to one motor controller. I see no reason why multiple motors means increased battery cost. In fact, there is a really good example of a 2 motor AWD car that uses one battery pack here:
Welcome to ProEV.com
System efficiency was a high priority for the people who made this car, and they were very successful, winning closed circuit road races against ICE cars.

[P-hack]

Originally Posted by e*clipse

This thread is about a FOC controller - that's why we're talking about it; it's not a "how to control induction machines as easily as possilbe" thread.

It is specifically about AC Controllers, not BLDC, ACIM has additional requirements and procedures for FOC despite sharing the acronym with BLDC. You will always need a set of switches and rotor position for each motor, but AC has other options if you can live without FOC (and have a hype-less idea of what that means). It is worth at least knowing that when discussing multiple motors, even in an FOC controller thread.

[e*clipse]

Originally Posted by P-hack

It is specifically about AC Controllers, not BLDC, ACIM has additional requirements and procedures for FOC despite sharing the acronym with BLDC. You will always need a set of switches and rotor position for each motor, but AC has other options if you can live without FOC (and have a hype-less idea of what that means). It is worth at least knowing that when discussing multiple motors, even in an FOC controller thread.

Are you really trying to educate me about BDLD?

You make a lot of assumptions.

[cts_casemod]

Originally Posted by e*clipse

I am going to use this controller for a motor like the one in your Prius, not an induction machine. The control hardware is more or less the same.

The prius motor is a hybrid between a switched reluctance and a PMSM. How are you going to do the angle estimations?