The author of that app. note needed to learn more before publishing, I'd say. I'd also been casually wondering, expecting that a Hall or such sensor and IC would significantly increase the cost of the motor.
This reminds me of AC synch. timing motors which start randomly in either direction, following a chaotic startup. A resilient brake (probably wrap spring?) blocks further wrong-way motion, and kicks the rotor in the wanted direction.
It took a long time to learn why the single-coil stepping motors always step forward in quantized-analog watches with hands, and also such clocks. Apparently, they all feed the coil with a pulse having polarity opposite that of the previous pulse.
They're likely all only two-pole, rotor and stator, with perm. mag. rotors. Asymmetrical pole pieces are the key, with air gaps decreasing progressively (sometimes? See below) along each pole face. Between pulses, the magnetic axis of the rotor aligns with the shortest/shorter air gap. That's offset by perhaps 30° from what it would be with a uniform air gap. A pulse must initially repel the rotor, but the offset ensures that it starts rotating forward.
The inventor's name is known, but apparently not widely; I've lost track. It's in the field of horology.
I have a little low-cost fan with an outside rotor and a four-pole wound stator. Rotor is simply a ring, magnetized presumably with four poles, alt. N & S. I removed the little plastic retaining ring and pulled out the rotor. I'd left it out, collecting dust, alas, hoping to learn more, also wondering how it worked without a flux sensor.
Has a porous bronze bearing, nice.
Well, surely enough, the poles each extended 75° or so, and doggone if they weren't asymmetrical! Think of each as a T, with sagging ends of the (curved) top writing stroke. One half of that stroke has a smaller outer radius! Of course, that means a bigger air gap.
Yes, but ... It wants DC. Well, there's a round PC board hiding under the stator. There's a little transistor down there, just possibly in an L-C oscillator circuit! Room to hide a little capacitor... It must be an AC synch. motor. (^_^)
In the future, I should use some mag. viewing film to find out the rotor mag. pattern, and disconnect the stator leads, feed it with DC after marking the rotor, and try stepping by alt-polarity pulses. (^^)(^^)
Early stepping motors, perhaps developed by Sperry Gyroscope for remote compass repeaters (displays), had salient-pole (spoked) rotors with maybe six teeth, no windings. Have forgotten whether rotor was perm. magnetized.
Stator was maybe four-pole, fed over two circuits by staggered pulses from a rotary switch in the gyrocompass. I think only one "phase" of the stator was energized at a time, statically, but iirc the other "phase" was energized also at the beginning of a step, partly moving the rotor. Tutorial from a many-decades-old Navy doc. explains it in painstaking detail, with fine explanatory illustrations. Alas, have forgotten some critical details.
Of course, there was a reduction gear train between the stepper and the big indicating disc.
When first powered up, each repeater had to be manually set, "synchronized" to the current course as indicated inside the gyrocompass.
Some time later, most likely, C.P. Steinmetz (or Ernst F.W. Alexanderson?, both with G.E. Research) invented the Selsyn™ (self-synchronizing, more than likely), known in the Navy as a synchro. The Navy adored them, using them to send data to powered gun mounts and turrets, among other purposes. Wondrous as they are, I'll forgo explanation, save to say they're internally like motors, usually with a two-pole rotor and a stator much like a three-phase motor's. They don't spin fast continuously, (or must not...). Not sure there's a good, first-rate explanation.
An amazingly-ingenious stepping motor made by Haydon Switch and Instrument was a modest variation of the simplest two-pole/shaded pole squirrel-cage induction motor found in countless lowest-cost AC fans and blowers, along with other applications.
Its rotor was, iirc, simply a two-pole magnet, diametrally (sp?) magnetized. Stator, including shading coils, was the same, except that a permanent bar magnet with quite-good coercivity (? resistance to being demagnetized) was placed between the coil and the stator poles. Perhaps there was an air gap inside the coil.
Power off, the combo. of the perm. magnets ensured that the rotor had only one stable position. Applying DC of the proper polarity and sufficient voltage reversed the stator flux polarity at the rotor.
Shading coils (those brazed loops) delayed flux buildup at their halves of the poles, ensuring consistent direction of rotation. During the pulse, assuming it had sufficient duration, the rotor settled half a turn from home position. When the pulse ended, shading coils ensured a return to home, in the same direction.
One unidirectional pulse to one stator coil per revolution was amazing enough, but modifying an existing design only moderately was remarkable, as well.
Perhaps the fact that most applications needed smaller step angles limited sales; real pity if so.
IIrc, Barber-Colman made a related variant of the ubiquitous two-pole/shading-coil motor mentioned above. It was reversible! It had 2+2 shading coils, wire wound and multiturn. Diagonally-opposite coils in series (quite likely), same for the other two, were connected to three screw terminals. Short-circuiting one pair determined start direction.
Fans of exotic motors must see the old electromechanical traffic-light controllers.