Are motors usually controlled in current or voltage or it depends?

[bonzer]

Hello everyone, I'm new to the motor world so sorry if this question sounds a bit stupid but please help me. I studied the control loops of a motor speed controlled by current and by voltage and what I see is that the transfer function changes. When we use a current control it looks better because I already have a pole at 0 so a -20dB/dec for low frequency without regulator. If I use a PI controller I get -40dB/dec and it's even better at reducing regulation error.  Instead in the case of voltage I have constant gain in low frequency without regulator so I have a type 0 system.  But for the current control I need it's own loop that is internal to the outer speed control loop while voltage control doesn't have a feedback so the whole system has only one loop. Therefore current control might be more complex.

Maybe it depends on the motor, maybe I should take them both into account? I don't need a full explanation I'm just looking for any answer to have a little idea. Feel free to ask me for details if it's not clear enough but I feel like my question is general.

[richard.cs]

Generally it depends what the motor is mechanically attached to. The motor is just a transducer that produces a torque proportional to current and a back-emf proportional to voltage, the latter effect means that when fed from a voltage source the motor approximates speed proportional to voltage. The control approach is generally selected based on what works for the attached load.

To model the control loop you need to include the electrical equivalent of whatever is connected mechanically. For example a flywheel or anything else with lots or rotational inertia acts to keep the motor at constant speed (voltage) and requires large torques (current) to change it so it's equivalent to a capacitor connected across the electrical terminals of the motor. Mechanical loads generally appear as resistors but may be non-linear.

[T3sl4co1l]

^ That.

Once you know the current/torque gain and EMF/RPM gain, you can model the motor and drivetrain as an equivalent RLC network.  Note that the motor itself has DCR in series.  The load manifests in parallel, along with core loss* and windage loss.

*Core loss varies with RPM, due to skin effect.  This is probably unlikely to matter on large, slow motors, but may be noticeable on small, high speed motors.  Or odd builds with thick laminations, or solid cores.  Depends on geometry.

This applies most directly to PMDC machines.  For electromagnetic types (series, parallel or mixed field), connect the field accordingly, and set EMF proportional to field strength.

If the motor is AC, this also still applies, but you must drive frequency proportional to EMF (because the EMF actually arises from frequency * flux), and AC-specific effects show up (phase angle, the necessity for a rotating field or initial velocity to begin motion).

And if it's not a synchronous (PMAC) machine, you again need to add the term for rotor field.  Which, for an induction (non-synchronous) machine, it's induced by the stator field; the rotor is a shorted turn on a transformer, and it's also a shorted turn wrapped around a huge chunk of steel, so it has a quite long time constant (100s ms, actually!).  That means, once a current has been induced into it, it tends to continue to flow.  So, once rotating, it acts much like a PMAC machine.  But then rotor current would decay, there needs to be some reason for the rotor current to be maintained.  As it happens, the rotor current itself rotates around the rotor -- and this happens at exactly the slip frequency.  Which if you didn't know, is why induction motors are always listed a bit below synchronous speed, say 1700-1750 RPM for a 4 pole machine at 60Hz.  The difference, the 50-100 RPM slip, gives the rotor field rotation rate.

The direct consequence of the flux limitation (voltage proportional to frequency), is very weak torque at low RPM for an induction machine.

PMAC machines have constant rotor field, so can deliver full torque from a stop, just as PMDC machines can.  A rotating field is required, so these are always* multiphase.  (It should be no surprise why they are in such high demand for automobiles!)

*There are single phase "BLDC" (actually PMAC + driver) motors; the reason they start is a lucky initial kick.  AFAIK, they can't be used for general applications with heavy loads; they're great for loads that have very little startup torque, like fans and blowers.

Tim

[bonzer]

Thanks a lot for your answers. Even though I'm new to motors I got you both because I studied DC motors these days and I feel in touch with these concepts. Now I understand.