As we have seen, sinewave commutation ensures that the static torque produced by the motor does not vary based upon the shaft’s position. But the SST servo drive’s closed-loop vector torque control goes further to ensure that the torque produced under dynamic conditions is highly accurate and efficient.
 
A big step beyond simple sinewave commutation,
So let’s take a look at why a simple sinewave commutated servo amplifier is less accurate than the SST servo drive’s closed-loop vector torque control. The main difference between these two techniques is that the closed-loop vector torque control constantly measures the amount of torque (and extraneous heating) that is produced and continuously works to servo the torque to the command value and the extraneous heating to zero. In contrast, a sinewave amplifier calculates what the currents should be to produce the required torque, passes these to three separate current loops and assumes that the proper torque will be produced without ever checking. Although each individual current amp is closed-loop, the torque control is open-loop! The difference is shown in the diagrams below:

The first thing you’ll notice is that the current loops in a sinewave amplifier operate in an uncoordinated fashion. However, the phases themselves are connected together, so the currents are dependent upon each other.
Let’s explore a simple analogy that describes the dilemma this causes: Imagine you are in a house with two bathrooms. You and your spouse are both taking showers at the same time (in separate bathrooms) and you decide you want the water a little hotter. So you, like the independent current loop, open up your hot water valve a little more. You get hotter water, but because you are "connected" through the hot water pressure, your spouse’s water cools down. So your spouse adjusts to maintain the desired water temperature. This goes on for a few moments with both of you adjusting your water valves until you find an equilibrium that suits both of you. Now imagine what would happen if you were trying to constantly change the water temperature along some arbitrary profile, like the sine angles trying to follow the arbitrary position of the rotor? It would be nearly impossible for you and your spouse to do this accurately without communicating and coordinating. Add a little noise or imbalance (someone washing the dishes or flushing a toilet) and it gets even harder. So because the current sensors have some noise, and the rotor’s back-EMF is not a true sinewave, and the amplifier may run out of voltage headroom, etc., a sinewave amplifier will naturally induce torque fluctuations as the independent current loops seek an equilibrium.
The second failing of the simple sinewave-commutated amplifier is that the current loop response delay forces the currents to lead or lag (depending upon where within the four quadrants of operation the motor is operating—motoring, braking , accelerating, etc.). As the speed becomes higher, the rate of change of the sinewave demand gets faster and the phase error becomes more significant. This lead or lag in time directly affects the angle of the magnetic vector produced in the motor, moving it off the ninety degree mark, causing torque to be reduced (and inaccurate relative to the position/velocity compensator’s command) and heating increased.
The third problem is caused by limited voltage headroom as the motor approaches rated speed. This is a little harder to explain, but suffice it to say that without an active measurement and control of the magnetic vector, these amplifiers run out of headroom earlier as speed is increased, drastically affecting their Torque Response Time. (If you’re an electrical engineer, this is analogous to having reduced large signal bandwidth.) So, just when you need a rapid reversal in torque, say at the peak speed of a triangle move, your sinewave amplifier will delay its response, affecting your tracking accuracy and settling time. This reduced "large signal response" also typically limits your ability to turn up the position/velocity compensator gains, which further limits your stiffness, tracking accuracy and settling time.
 
So what is Closed-Loop Vector Torque Control?
Incontrast to the simple sinewave amplifier that uses the sinewave references as commands to uncoordinated current loops, the SST servo drive’s closed-loop vector torque controller uses the sinewave references in mathematical transforms to "de-rotate" the amplitude and angle of the electromagnetic vector from the measured currents. Then, after the torque loop calculates the voltage amplitude and angle, another transform "re-rotates" these into the three simultaneous voltages provided to the motor. See the diagram below.

O.K., that’s a big mouthful—Let’s imagine you are trying to shoot a bank robber as he flees in a getaway car while swerving, braking and accelerating. The difference just described above is like trying to shoot him from the sidewalk as he erratically speeds by, as compared to shooting him from the back seat of the getaway car. In the latter case, your reference frame is the same as his, making your objective quite easy. In the same way, the uncoordinated loops of the sinewave amplifier attempt to pump current in and out of the phases as they see the rotor move past, hopelessly trying to exactly hit the magnetic vector target which is constantly moving. (By the time they "shoot" at it, it has moved). The closed-loop vector controller, however, has as its input, the currents already "de-rotated" into the magnetic vector referred to the rotor. Therefore the torque controller sees the magnetic vector (and thus the real torque) as if the controller was sitting on top of the spinning rotor. This, and the re-rotation of the controller’s voltage outputs, essentially removes the moving target effect putting the controller "in the back seat". So the SST servo drive’s closed-loop vector torque control is extremely accurate and quick under all conditions of speed acceleration, deceleration, rate of change of torque demand, etc.
Again, let’s re-emphasize, even if the torque calculated by the position/velocity compensator is perfectly accurate, the servo control will be inferior if the actual torque at the motor shaft is not quickly and accurately produced. The SST servo not only calculates the optimum torque command, but also quickly and accurately produces it at the motor shaft.