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

[thingstodo]

Continuing on from this morning (I had to go to work)

I edited the last post and put a link to an explanation of how induction motors work that I think is pretty good. i duplicated it here

http://www.explainthatstuff.com/induction-motors.htm

** Sorry for the long post **

Originally Posted by e*clipse

A analogy may be how an unloaded ac induction motor responds when it is first plugged into a 60hz wall socket. Initially the rotor is at zero rpm and the field is rotating at full speed. The current (and torque) is at a maximum because the slip is at a maximum.

That does not quite match my understanding. There is a torque curve on the motor that is related to slip. 100% rated torque is at rated slip. As the slip rises to somewhere between 3.5X and 5X rated slip you reach the 'knee of the curve' at maximum torque, anywhere from 4X to 7X rated current. To get the fastest acceleration out of an induction motor, you need to keep the slip as high up that curve without going over the top. Torque after the peak drops on a curve to about 50% of the rated (depending on motor design) as a minimum and then rises as the slip increases until at 0 rpm is maybe 65 - 70% of rated torque.

As the motor speeds up, the slip decreases and the torque decreases.

Again, not quite how I understand it.

Eventually the motor gets close to the field speed, but will never match it because there always needs to be enough slip to keep the rotor spinning and overcoming the drag in bearings, the fan, etc.

And by the load it is driving

The torque curve starts with HUGE spike (limited by winding resistance, inductuance, etc.), then settles to near zero, while the speed curve gently approaches a value close the the field speed. All this happens relatively slowly, as it generally takes a couple of seconds for an induction motor to come up to speed.

A portion of the spike comes from the additional current that the motor absorbs at 0 rpm. Modern high efficiency motors can take 10X rated current for 10s of ms. And more current means more torque, even with the lower torque curve.

So, in a way, the torque is somewhat speed related - I'm trying to clarify - I'm NOT saying that speed is the control request.

OK.

The problem Paul is running into is that at high speeds, control gets more difficult and rough. I'm guessing here, but physically what may be happening is that as the motor spins faster, it generates more back-EMF. So maybe the stator field has to spin proportionally faster than the rotor to compensate and produce the desired torque.

As I understand it, the back-EMF limits the current, which limits the torque. So if you can't push the amps, you don't maintain the torque. The idea is that if you take a motor wound for 100V and drive it at 4X the speed, you need 400V instead. I have not been able to test that so far. I guess I'm not creative enough

As for the control getting more difficult, I don't know. Getting more rough ... that's not what I see in commercial controllers. Is it possible the feedback is getting more coarse as you have fewer samples per motor rotation? Above perhaps 20 Hz output, I've never seen a motor run rough for electrical reasons ... bearing issues, alignment issues, load issues - sure. But not controller issues.

The difference between the bus voltage and the back-EMF probably doesn't reduce in a nice, linear fasion.

I thought that it did. The impedance of the motor winding is related directly to frequency?

With both IC motors and electric motors, the torque reduces as the speed increases. Electric motors are cool because up to a certain speed, the torque is constant. However, after that point, all motors see their torque drop in a decaying exponential. So, Vd and Vq can be approached from a linear perspective as long as the torque output can be constant.

On the induction motor, I've seen graphs in presentations on VFDs (controller sales guys) below rated speed called Constant Torque and above rated speed called Constant Horsepower, as the torque degrades linearly in an ideal motor, minus the wind resistance, bearing friction, etc. The graphs look linear to me ..

Thus, for an induction motor, the changes in Vd and Vq need to be proportionally larger when operating at a speed where back-EMF becomes significant. In other words, this is a very non-linear problem, and attempting to solve a non-linear problem with a constant value solution will be nearly impossible. It does seem that it might help to give the "motor men" some parameters to work from depending on operating conditions may help the situation where one or the other attempts to grab all the resources...

I think Vq, or rather Iq, needs to remain at the desired setpoint and Vd must be adjusted according to available voltage. But that's only my opinion until I see some data ... data usually changes my first opinions to better opinions.

One thing I am pretty sure about is that shooting for a really fast convergence is a nice, interesting challenge, BUT it can lead to it's own problems. I think Paul was able to get the motor to converge in about 1ms with a 5amp step. While it's cool that the control system can get the motor to converge that fast, it's probably not necessary. There is soooo much inertia in an automobile that requiring a 1ms convergence to a step input would require HUGE amounts of torque. F=ma.

1 ms convergence is perhaps a bit aggressive. At 8 Khz carrier frequency, that's only 8 pulses ... right?

In order to avoid twisting the transmission input shaft, perhaps something in the order of 100 - 200 ms would be a better target? The ICE world ramps up in 500 - 750ish ms for a more normal car .. perhaps 100 - 200 ms if you rev it and pop the clutch?

It would probably help a lot to figure out some reasonable ballpark numbers for mass and desired accelleration (not just of the car, but how the motor/drivetrain responds with inertia) This would help produce some reasonable specifications for the control loops (like convergence time) to shoot for. Perhaps this could be broken down with different specifications over the speed range, depending on the motor's torque output curves.

My car is likely close to typical - 2500 lbs curb weight. 92 Mazda MX-6
The low end would be more like the WIKISPEED car - 1400 lbs, no roof, roadster
The high end ... perhaps a half ton truck or SUV? 4500 lbs or so.

Performance ... that I don't have a good grasp of. I'm looking at starting in second gear, doing one shift to fourth gear, and getting to 60 mph in maybe 15 seconds ... just grabbing numbers. I'll have to look into what that would require.

The torque required to slip the clutch on my Mazda (clutch has 360K on it) is something I need to measure.. well, measure the DC going in and calculate about what the torque would be. I'll be using a netgain warp 11 (series DC motor) for that. Using 5th gear to ensure a large load on the motor. Limit the current to 100, 200, 300 ... 1100 amps and log times, rpm, currents, voltage sag, etc for perhaps 0 - 60 kph. When it starts to slip I'm told I'll be able to smell the clutch (sounds like a bad idea, but I have not heard a better one so far)

After the test (likely a few tests to verify that the smell comes back at about the same torque) I replace the clutch and the front axles for sure since they need replacing anyway. Plus anything I break during the test

[Cyruscosmo]

Hey Thingstodo

The torque required to slip the clutch on my Mazda (clutch has 360K on it) is something I need to measure.. well, measure the DC going in and calculate about what the torque would be. I'll be using a netgain warp 11 (series DC motor) for that. Using 5th gear to ensure a large load on the motor. Limit the current to 100, 200, 300 ... 1100 amps and log times, rpm, currents, voltage sag, etc for perhaps 0 - 60 kph. When it starts to slip I'm told I'll be able to smell the clutch (sounds like a bad idea, but I have not heard a better one so far)

After the test (likely a few tests to verify that the smell comes back at about the same torque) I replace the clutch and the front axles for sure since they need replacing anyway. Plus anything I break during the test

I have measured the torque it takes to slip a clutch before on the bench. What you do is secure the flywheel you will be using to the bench with clamps or bolts or whatever. Then install the clutch disc and pressure plate and torque to spec. Then you get a splined shaft from a junk transmission and cut it off long enough for it to stick out past the pressure plate and weld a one inch nut to the cut off end.

Apply a torque wrench to the nut, pull the handle and watch the gauge.

Easy as that.

Cyruscosmo

[e*clipse]

Whooo boy, where to start. Thank you for the corrections, however I feel that the main point I was trying to express was missed.

Yes, I did get it wrong about an induction motor's torque curve. I for some reason remembered a high peak in the beginning. In reality, it has a lot to do with the motor's resistance and other design characteristics. So, I stand corrected and completely agree with your description of an induction motor's torque curve. Here are some standard torque curves for induction motors based on their duty classes (a,b,c,d, etc)


I'm talking about an uncontrolled, unloaded motor or an uncontrolled motor with a constant load. If it has a VFD, it is a controlled motor, and someone worked pretty hard to get that nice, predictable performance.

Controlled AC motors (both brushless DC and induction) can see nice torque curves like this:

Here's the Prius MG2 Buried Permanent Magnet Synchronous motor:


I'm pretty sure that curves similar to the ones above are what the VFD salesmen showed in their presentations.

Notice how curves like the one above are basically composites of many curves:


So, if you have a VFD and a motor with a base speed of 1800rpm, then you can get the characteristic AC motor curves by fixing the speed to any speed less than the base speed. Let me repeat - you are using the VFD to produce a constantly rotating magnetic field - like a 60hz line frequency - for speeds up to the motor's base speed. This is NOT using any of the VFD's feedback capabilities, whether a rotor position sensor or "sensorless" control that uses current sensors. You're merely providing different constant speed rotating magnetic fields that are NOT adjusted to provide optimal slip as the motor's speed increased. This is merely a set of tests to provide fixed conditions for the motor to respond to.

Beyond the base speed, notice that the curves don't peak as high - those peaks fall off in a decaying exponential no matter how good the control system is. Now we're getting into the nonlinear part of the torque curve I was talking about. Also note that if you put a dot at the peak of all those curves, then drew a line from dot to dot, you would end up with something like the Prius torque curve. This part of the torque curve is very non-linear, however it's an important to use this region to take full advantage of the motor's capabilities.

That's what a nice FOC AC motor controller can get you and that's why it's worth it to do all this work.

I'm going to stop here. I need some sleep.

Originally Posted by thingstodo

Continuing on from this morning (I had to go to work)

I edited the last post and put a link to an explanation of how induction motors work that I think is pretty good. i duplicated it here

http://www.explainthatstuff.com/induction-motors.htm

** Sorry for the long post **

That does not quite match my understanding. There is a torque curve on the motor that is related to slip. 100% rated torque is at rated slip. As the slip rises to somewhere between 3.5X and 5X rated slip you reach the 'knee of the curve' at maximum torque, anywhere from 4X to 7X rated current. To get the fastest acceleration out of an induction motor, you need to keep the slip as high up that curve without going over the top. Torque after the peak drops on a curve to about 50% of the rated (depending on motor design) as a minimum and then rises as the slip increases until at 0 rpm is maybe 65 - 70% of rated torque.

Again, not quite how I understand it.

And by the load it is driving

A portion of the spike comes from the additional current that the motor absorbs at 0 rpm. Modern high efficiency motors can take 10X rated current for 10s of ms. And more current means more torque, even with the lower torque curve.

As I understand it, the back-EMF limits the current, which limits the torque. So if you can't push the amps, you don't maintain the torque. The idea is that if you take a motor wound for 100V and drive it at 4X the speed, you need 400V instead. I have not been able to test that so far. I guess I'm not creative enough

As for the control getting more difficult, I don't know. Getting more rough ... that's not what I see in commercial controllers. Is it possible the feedback is getting more coarse as you have fewer samples per motor rotation? Above perhaps 20 Hz output, I've never seen a motor run rough for electrical reasons ... bearing issues, alignment issues, load issues - sure. But not controller issues.

I thought that it did. The impedance of the motor winding is related directly to frequency?

On the induction motor, I've seen graphs in presentations on VFDs (controller sales guys) below rated speed called Constant Torque and above rated speed called Constant Horsepower, as the torque degrades linearly in an ideal motor, minus the wind resistance, bearing friction, etc. The graphs look linear to me ..

[CrazyRick]

E*clipse, when an induction motor is driven by its correct V/Hz ratio, the torque will remain constant and horsepower increases the faster you go, when the Hz moves beyond the maximum voltage, the motor is in what's refered to as "constant horsepower" mode because the RPM will increase but the torque decreases since the V/Hz ratio is not met.

It would be very easy to run a 240V 60Hz motor on 480V 120Hz have the same torque and double the horsepower!


Rick

[e*clipse]

Um, yes, that's what I posted above. Look at graph 2 and 3; that's basically what you are describing. Paul has been very successful within the base speed range of performance. Beyond the base speed (which has a LOT to do with the input voltage, which Paul is wisely keeping low for now) things get non-linear. To make full use of a motor's capability, it's good to make full use of the "constant power" part of the performance curve.

The problem is that we are not starting with a complete VFD. We are making one. Ive done many motor control applications where one starts with a servo motor and a controller, which will have some performance curve like graph 2. No understanding of the motor or controller is necessary in this case to make an application (like a CNC machine) work. However, since we are actually making the controller, an understanding of what is physically happening underneath the nice control may be helpful in actually making a nice control. All of the CNC's I've owned have had seperate controllers: 1 for the overall machine and 1 for each servo. We are making the motor controller, not the overall control.

- E*clipse

Originally Posted by CrazyRick

E*clipse, when an induction motor is driven by its correct V/Hz ratio, the torque will remain constant and horsepower increases the faster you go, when the Hz moves beyond the maximum voltage, the motor is in what's refered to as "constant horsepower" mode because the RPM will increase but the torque decreases since the V/Hz ratio is not met.

It would be very easy to run a 240V 60Hz motor on 480V 120Hz have the same torque and double the horsepower!


Rick

[MPaulHolmes]

I've found a few more bugs. won't be long now I hope. I had a scare when it said there was a desat event during debugging. it turned out to be a debugging array index that went out of bounds, and modified a couple variables. thank goodness the hardware protection is independent of the micro.

I made a rotor time constant vs rpm after 1 second curve. you get a nice clear smooth curve with a global maximum where the time constant is. I had to recompute it because there were a couple bugs in the rotor flux angle function.

[thingstodo]

Originally Posted by Cyruscosmo

I have measured the torque it takes to slip a clutch before on the bench. What you do is secure the flywheel you will be using to the bench with clamps or bolts or whatever. Then install the clutch disc and pressure plate and torque to spec. Then you get a splined shaft from a junk transmission and cut it off long enough for it to stick out past the pressure plate and weld a one inch nut to the cut off end.

Apply a torque wrench to the nut, pull the handle and watch the gauge.

Easy as that.

That makes much sense. But my clutch is in the car now, and is coupled up and ready to go.

Perhaps I can remove the parts and do something like you describe after I get through my test. I thought the test was crude when I first read about it, but it's grown on me.

Not sure how I'd put the nut on the shaft, or find the junk transmission. But it sounds a LOT more accurate than what I will be doing.

[thingstodo]

Originally Posted by e*clipse

Here's the Prius MG2 Buried Permanent Magnet Synchronous motor:


I'm pretty sure that curves similar to the ones above are what the VFD salesmen showed in their presentations.

I went back through the slide deck (or the parts that I kept) and your graphs are better than the ones I have.

I remembered wrong and was too arrogant to go and verify my info. Sorry

Running out of voltage, and using the available voltage and current to best advantage, just makes sense.

I think I get it

[MPaulHolmes]

My main contactor evidently has an on resistance between 7 and 300 ohms. also, either the motor now has a problem or something needs to be fixed on the controller. I don't think its the controller, but I'll check. what is a common failure mode for a 3 phase motor that has had too much current? this is happening at a terrible time. I think I got rid of just about all the bugs.

[thingstodo]

Originally Posted by MPaulHolmes

My main contactor evidently has an on resistance between 7 and 300 ohms. also, either the motor now has a problem or something needs to be fixed on the controller. I don't think its the controller, but I'll check. what is a common failure mode for a 3 phase motor that has had too much current? this is happening at a terrible time. I think I got rid of just about all the bugs.

7 ohms is too much, 300 ohms is WAY WAY TOO HIGH.

Too much current heats up the coils, the insulation degrades. Depending on the insulation class, this is 60C, 80C 105C or 125C above ambient temperature (it is a relative temperature). This link seems pretty good

http://toshont.com/ag/mtrldesign/AG0...%20Rise%29.pdf

The insulation failure allows current to flow to ground (about 2/3 of the time) or between phases (the other 1/3). There are good tests for phase to ground - a megger will put high voltage DC on the phase and check for current flow to ground. For phase to phase faults, there is really no easy way to check.

By the time you take the motor apart, you can usually smell a problem (failed electrical insulation has a distinctive smell). A small scorch mark is all you are likely to see.

This is a *BAD* way to verify your hardware over-current circuit!

I would replace the contactor first. If the contactor is over 7 ohms, your controller may be starved for voltage and amps