PDA

View Full Version : Canon EF lenses - the micro USM drive


Doug Kerr
December 23rd, 2010, 09:02 PM
The Canon "USM" focus drive system comprises two very different mechanism families, the "ring USM" drive and the "micro USM motor" drive.

The micro USM motor is a small-diameter cylindrical motor (about 0.43" in diameter), which drives the lens focusing cam by way of gearing. There are two versions. The original micro USM motor is about 1.05" long, while the newer version, the micro USM II, is about 0.53" long.

In this note I explain the operating principle of these motors. The figure I will use does not at all illustrate many of the physical features of either version. The reader interested in such may wish to look at the nice illustrations in the Canon EF Lens Work III book.

Figure 1 shows the mechanism in a highly-schematic way.

http://dougkerr.net/illustrations/usm_micro_01.gif
Figure 1. Principles of the Canon Micro USM drive

View A shows the motor in a de-energized state. The stator sits on a piezoelectric element array which we can think of as sitting on the base of the motor. (That is actually not how it is arranged, but it is the easiest way to make the rest of the description credible). The rotor is connected by a "slip joint" to the output gear; a spring inside the rotor presses it axially against the top of the stator.

The piezoelectric element array has four "petals", each of which can be made to get thicker or thinner (by a very small amount) by the application of voltage to electrodes plated on the element.

These are energized in a "two-phase" fashion. The result is that the stator moves in a way reminiscent of a person twisting his head to loosen a stiff neck while looking straight ahead. The stator does not rotate, but its axis swings around the axis of the motor proper. (The technical terms for movement of this type is nutation.)

In view B, we see the motor in operation. We see it at the phase of its nutation in which the stator is "leaning to the left" from our point of view. Assuming a certain direction of nutation, it will next lean away from us, then (as seen in View C) lean to the right, then lean toward us, and so forth. The movement illustrated is greatly exaggerated for clarity.

Since the stator is now continually leaning, and the rotor rides on its high edge, the rotor has had to move up a little, against its spring. (Of course, the actual movement is very tiny.)

The little rim of the top of the stator contacts the lower face of the rotor in a single spot. Note that the radius from the axis of the stator to the rim, R, is greater than the radius from the axis of the rotor to the point of contact, r.

The point of contact travels equal actual distances along the contact track on the lower face of the rotor and along the contact ring of the stator. But the circumference of the latter is greater than the former.

Thus, when the contact point has moved a full circle's worth around the stator, this is more than a full circle's worth around the track on the rotor, and the rotor is compelled to move a small amount to accommodate that. This is how the rotor is moved.

Since the tilt of the stator is really very small, the radii are very nearly the same, and so the excess distance per nutation cycle is very small. Thus, the rotor turns at a relatively slow rotational velocity compared to the rate of nutation (and I don't know what that is). Accordingly, there is usually not any need for significant reduction gearing in the drive chain to the focusing cam.

Very clever.

Best regards,

Doug

Doug Kerr
December 28th, 2010, 08:00 AM
This replaces an earlier version of this note in order to provide an improved drawing and the update the text to match:

**********

The Canon "USM" focus drive system comprises two very different mechanism families, the "ring USM" drive and the "micro USM motor" drive.

The micro USM motor is a small-diameter cylindrical motor (about 0.43" in diameter), which drives the lens focusing cam by way of gearing. There are two versions. The original micro USM motor is about 1.05" long, while the newer version, the micro USM II, is about 0.53" long.

In this note I explain the operating principle of these motors. The figure I will use does not at all illustrate many of the physical features of either version. The reader interested in such may wish to look at the nice illustrations in the Canon EF Lens Work III book.

Figure 1 shows the mechanism in a highly-schematic way. The movements shown are greatly exaggerated.

http://dougkerr.net/illustrations/usm_micro_02.gif
Figure 1. Principle of the Canon Micro USM drive

View A shows the motor in a de-energized state. The stator sits on a piezoelectric element array which in turn sits on a "counterpoise" (my term) - sort of an inertial anchor. The entire sandwich is suspended (essentially at mid-thickness) on a "spine" running through the motor (not shown). The rotor is connected by a "slip joint" to the output gear; a spring inside the rotor presses it axially against the top of the stator.

The piezoelectric element array has four "petals", each of which can be made to get thicker or thinner (by a very small amount) by the application of voltage to electrodes plated on the element.

These are energized in a "two-phase" fashion. The result is that the stator moves in a way reminiscent of a person twisting his head to loosen a stiff neck while looking straight ahead. The stator does not rotate, but its axis swings around the axis of the motor proper. (The technical term for movement of this type is nutation.) The counterpoise makes a complementary movement so there will be no net inertial reaction that would tend to make the whole motor "sway". (Canon actually refers to both the stator and the counterpoise as "oscillators".)

In view B, we see the motor in operation. We see it at the phase of its nutation in which the stator is "leaning to the left" from our point of view. Assuming a certain direction of nutation, it will next lean away from us (I didn't draw that), then (as seen in View C) lean to the right, then lean toward us (I didn't draw that either), and so forth. The movement illustrated is greatly exaggerated for clarity - the edge of the stator only actually raised by a few thousandths of a millimeter.

Since the stator is now continually leaning, and the rotor rides on its high edge, the rotor has had to move up a little, against its spring. (Of course, the actual movement is very tiny.)

The little rim of the top of the stator contacts the lower face of the rotor in a single spot. Note that the radius from the axis of the stator to the rim, R, is greater than the radius from the axis of the rotor to the point of contact, r.

The point of contact travels equal actual distances along the contact track on the lower face of the rotor and along the contact ring of the stator. But the circumference of the latter is greater than the former.

Thus, when the contact point has moved a full circle's worth around the stator, this is more than a full circle's worth around the track on the rotor, and the rotor is compelled to turn a small amount to accommodate that. This is how the rotor is moved.

Since the tilt of the stator is really very small, the radii are very nearly the same, and so the excess distance per nutation cycle is very small. Thus, the rotor turns at a relatively slow rotational velocity compared to the rate of nutation (and I don't know what that is, but perhaps 30,000 nutations per second)). Accordingly, there is usually not any need for significant reduction gearing in the drive chain to the focusing cam.

Very clever.

Best regards,

Doug