THK KR20 linear actuator

In photomacrography, especially in the studio/laboratory, it is essential to have some means to focus with a precision substantially higher than that provided by photographic bellows and focusing racks designed for use with a camera. To be really useful in photomacrography, a focusing rack must be able to carry the weight of a normal camera, lens and bellows without detectable sagging and vibration. It must focus smoothly without seizing and creeping, even when the camera is mounted vertically. It must allow a reliable focusing with a precision that varies, depending on magnification, from a few tens of μm to a few μm.

Since focus stacking is now commonly used in photomacrography, the focusing mechanism must be able to provide this precision repeatably, while being able to move the camera by a total travel of a few mm to about ten cm (the longer, the better). Since the length of the optics and extension tubes or bellows used in photomacrography is quite variable depending on magnification, focal length of the objective, and the use of a tube lens as well as its focal length, the amount of travel allowed by the focusing mechanism should be enough to accommodate this variable length.

In the magnification range between roughly 1x and 20x, an industrial linear stage can provide a sufficient travel length and precision. At higher magnifications, a microscope stand and focuser with coarse and fine focus controls is usually better, although work may still be possible up to about 50x with the best industrial stages.

Figure 1. Second-hand THK KR20 linear actuator with NEMA17 stepper motor.

The THK KR20 actuator shown in Figure 1 provides a total travel of 140 mm. The rail is 205 mm long and the platform moves along the rail on a double set of recirculating steel bearing balls on each side. Both rail and platform are made from solid blocks of forged and machined steel. This gives very low amounts of twist and tilt of the moving platform. THK is a Japanese company making a variety of precision mechanical equipment.

Figure 2. Platform from another THK KR20 linear actuator. The arrows indicate guides for ball bearings.

Before acquiring the actuator shown in Figure 1, I happened to have another second-hand THK KR20, albeit without motor and motor cage. It proved too difficult to find a separate motor cage, so in the end I gave up and purchased a complete actuator. Nonetheless, my first KR20 proved to be instructive when disassembled.

One word of caution: The ball bearings of both platform and nut are simply too difficult to reassemble without specialized equipment. In particular, the recirculating balls of a nut with 1 mm thread pitch are extremely small and slippery. Therefore, disassembling these parts is essentially an irreversible process. You can learn a lot through the disassembly of this and similar equipment, including that disassembling equipment that you still intend to use in the future may be a costly experience.

The screw that moves the platform has a thread pitch of 1 mm per turn and uses dozens of recirculating sub-millimeter sized steel balls between the fixed helical groove of the platform and the rotating screw. In other words, there are no metal-against-metal sliding contact surfaces between rails and platform, and between screw and nut. All surfaces between platform and screw, and platform and rail, are rolling on ball bearings. The screw likewise turns on ball bearings at both ends of the rail. The moving platform has a brass nipple for re-lubricating with grease. Rubber skirts around the platform give some protection from contamination of the ball bearings located between platform and rail.

The profile of the grooves that guide and hold the ball bearings on the sides of the platform and rail is not a simple semicircle, but slightly trapezoidal. This minimizes the contact surface between balls and grooves, reduces friction, and allows grease to flow around the balls when the latter roll within the grooves. This is also true of the profile of the helical thread on both nut and advancing screw. They are not simple trapezoidal threads and they do not directly slide against each other, unlike the threads on simpler screw-operated linear actuators.

Precision-molded plastic parts receive the balls at either end of the platform and forward them to the respective recirculating guides drilled within the platform. A single grooved plastic part shuffles the ball bearings from one end to the other of the multi-turn helical groove between platform and screw.

The THK KR20 linear actuator is sold in a variety of configurations. The most common option is a choice between 1 mm and 10 mm screw pitch. For focusing in photomacrography, 1 mm is the obvious choice, while 10 mm is probably a bit too coarse for magnifications above roughly 5x-10x. A higher screw pitch makes travel of the platform along the rail faster, less precise, and more likely to be accompanied by vibration introduced by the stepper motor.

Another option is the type of motor (servo vs. stepper, or no motor). In general, stepper motors are easier to control with general-purpose microprocessor boards and simple motor controllers.

Yet another option is whether the rail has one or two moving platforms. Two platforms reduce the amount of twist and sag of equipment attached to the platform(s), but also reduce the maximum amount of actuator travel along the rail. Using one or more additional rails placed some distance from the actuator and multiple sliders on the additional rail(s) provides an even better reduction of twisting, sagging and play, but requires all rails to be perfectly aligned parallel to each other.

The KR20 is usually equipped with a NEMA17 stepper motor with 200 steps per revolution (1.8° rotation per step). I replaced this motor with one providing 400 steps per revolution (0.9° per step). This gives a theoretical travel of 2.5μm per full step of the motor. Microstepping can give a finer travel resolution, but the platform should not reverse in direction during travel, if backlash is to be avoided. In addition, with simple stepper drivers the size of individual microsteps varies non-linearly according to where the microstep is taken with respect to the boundaries of full steps, and changing the loading of the platform changes the exact location of both full steps and microsteps along the rail.

Further options are the length of the rail (between roughly 10 cm and over 1 m), the width of the rail (1 inch in the KR20), the stepper motor size (usually, NEMA 17 for the KR20), and whether the screw is protected from direct hand contact by an aluminium cover (the SKR20 series, which, however, is not dust- and dirt-proof). Second-hand KR20 actuators salvaged from equipment may or may not include a flexible coupler between motor and screw, a cage for the motor and its coupling, and a damper (roughly working as a flywheel) to reduce the vibration of the stepper motor. One or two limit switches (usually optical) may also be present at one or both ends of the rail.

The KR20 rail can be attached to a support with a number of M3 screws (six in the specimen shown in Figure 1). The moving platform has threaded sockets for four M3 screws and two M2.5 screws. Only the M3 sockets are generally used.

There are a number of "tricks of the trade" that make it easier to work with linear actuators, stepper motors and their drivers. For example, a simple but useful "trick" to identify which pair of wires is connected to the same winding (or "phase") of an unknown stepper motor is shorting a pair of wires while turning by hand the motor axle. When you short the two wires of the same "phase", the force necessary to turn the axle suddenly increases.

Two-phase stepper motors have two independent windings (usually with four connecting wires, although sometimes three if the two windings use a common earth). These motors are the simplest to interface to a driver and to control in software. Multi-phase stepper motors can have a smoother rotation, but require more complex drivers and are usually reserved for higher torque applications.

Plenty of useful information can be obtained on forums that discuss the construction and maintenance of 3D printers and CNC mills (which also use linear actuators).

KR20 equipped for focus stacking

MJKZZ is a small company based in China and specializing on motorized precision rails and controllers for these rails. The company also has a German website and web shop that ships merchandise directly from Germany (which, for EU residents, if of course far more practical than ordering directly from China). The German web shop seems to be staffed by only one person, but my experience with him has always been good, with quick replies to e-mail inquiries and fast shipment of ordered items. Occasionally, some items may need to be back-ordered from the Chinese supplier, but it seems that the German web shop tries to keep a small local stock of most items. They also have a US-based reseller, with a web site concentrating on the MJKZZ water-drop photography machinery. They do mention also focus stacking, so if you are looking for the latter equipment and are based in the US, you could try to contact them.

MJKZZ sells different types of motorized rails, including THK KR20 rails salvaged from industrial equipment, cleaned up and re-greased, and equipped with a number of added parts machined from aluminium alloy and black anodized. The quality and finish of these added parts, apparently manufactured in China, is good, on par with equipment sold by the best brands of photographic equipment based in western countries.

If, like me, you already have a KR20 rail, you can purchase most or all of of the necessary parts from MJKZZ and assemble your own stacking system. I strongly recommend that you align the camera platform to the body of the rail by using machine squares and the like. Just screwing the parts together is likely to leave them slightly misaligned, which may show up along the edges of your stacked images. However, unless you already have a KR20 lying around, it is probably both cheaper and easier to purchase a complete kit from MJKZZ, which comes with the added bonus of the rail and associated mechanical parts being already aligned, serviced and tested.

Figure 3. THK KR20 equipped with MJKZZ parts and home-made limit switches.

Figure 3 shows a rail assembled with parts separately purchased from MJKZZ, with the addition of a home-made limit-switch system. Although the MJKZZ controllers have an input for limit switches, their KR20 modified rail does not come with such switches.

One peculiarity of the MJKZZ KR20 rail is that the Arca-compatible camera clamp, like the underlying square platform (MJKZZ part UP) is equipped with holes for two locating pins that prevent the clamp from twisting around. This does not remove the need to carefully align the two parts, because there is a little play between pins and their holes. All works well if you orient the clamp with the locking handle parallel to the screw of the actuator. The base of the clamp does not completely overlap the square plate, but this is not a problem in practice.

If you try to mount the clamp oriented with its locking handle perpendicular to the actuator screw, on the other hand, the clamp is no longer centered over the actuator, but offset to one side by several mm. The latter orientation of the clamp is far more useful than the alternative for which the square plate is apparently designed, because most photomacrography lenses and microscope tubes are better equipped with a long Arca-compatible plate parallel to the optical path. Among other things, this makes it easy to coarsely position the optics close to the subject, by sliding the long plate within the slightly untightened clamp. However, the offset placement of the clamp in this orientation simply makes no sense, and screwing the clamp to the center threaded hole of the plate should be the most obvious - and correct - solution. Albeit, this does not allow the use of the locking pins. The square plate should have been designed with this in mind, allowing the locking pins to be used in any of the four orthogonal orientations of the clamp. This is an obvious design miss that would be simple to correct by relocating the pin holes in the plate (and/or in the base of the clamp).

The orientation of the plate in the above figure is only optimal when attaching a camera body to the rail, via a camera plate. My solution to this problem is using a right-angle clamp-and-plate adapter between the rail and the lens foot (inserted in the clamp of the rail in the above figure).

Figure 4. THK KR20 from MJKZZ, with additional third-party long Arca-compatible plate and plastic protection for the actuator screw.

Most MJKZZ KR20 kits come with the rail mounted on a section of black 20x80 channel aluminum profile. In both my KR20 rails, I attached the rail instead to long Arca-compatible plates, because this is what I use for virtually all my equipment. Figure 3 shows the MJKZZ Arca-compatible plate pre-drilled with holes suitable for this purpose. Figure 4 shows a KR20 supplied by MKJZZ as part of a kit, where I replaced the base of the rail with a longer and thinner Arca-compatible plate I drilled myself. The longer plate slightly increases the amount of travel of the rail when mounted on a support via an Arca clamp, and therefore enhances the versatility of the equipment. An even longer plate could be used, extending under the motor cage.

Note also that I added a relatively thin and flexible plastic sheet between the top of the rail and the bottom of the square plate, passing through the hollow portion of the C-shaped aluminum part joining the rail and the square plate. This part carries no MKJZZ part identification, but the German web site identifies it as Carriage Plate (for THK KR20 and HIWIN KK40).

Some KR20 models are equipped with a thin aluminum plate in the same position as my plastic sheet. This aluminum or plastic part protects the grease-covered screw of the actuator from accidental contact with one's hands while handling the rail or preparing for a stacking session, and is therefore a useful thing to have.

Sometimes, surplus KR20s are equipped with other types of motor and additional equipment. One example is a synchronous motor with the axle extending from both ends of the motor casing. One end is coupled to the screw of the actuator, the other carries a dampener shaped like a wheel, which dampens some of the vibrations caused by the motor stepping. Another example, also equipped with double axle, carries an optical encoder used to inform the motor controller of the actual number of steps performed. In applications where the motor must rotate very fast, or is subjected to loads of variable intensity, the motor may stall, which causes it to skip one or more steps. In other cases, the motor axle may be forced to rotate under an external load, effectively causing the motor to perform "extra", unintended steps. Feedback from the encoder allows the controller to re-issue the signals for the missing steps or to back up the extra steps, while keeping a correct count of the performed steps.

When the KR20 is used for focus stacking, I found that the number of performed steps is very accurate, and neither a dampening flywheel nor an optical encoder are necessary. Also, limit switches may not be strictly necessary because the rail carries internal hard rubber dampeners at either end. However, hitting the rubber dampeners with the platform of the rail causes strong vibrations of the platform, and I prefer to avoid this by using limit switches, especially since the MJKZZ controllers have an input for these switches.

Other linear actuators

Before the THK KR20, I used several other sliding linear rails and actuators, albeit models meant primarily for manual rotation of the micrometer screw. For example, the iShoot Macro Focusing Rail 150 proved disappointing, while the Newport UMR 8.25A is extremely well made. The Manfrotto 454 is a very basic focusing rail, with a manufacturing quality and tolerances more similar to something made decades ago in the USSR rather than in present-day Western Europe.

I also used, and was not satisfied with, a Velmex Unislide linear rail. I had heard good things about the Velmex rail, but I discovered that it has multiple things I do not like: thin nylon sliding inserts between platform and rail, no way to adjust the tolerances between platform and rail once these inserts wear down, advancing screw running within a slightly elastic, self-lubricating plastic nut, and fiddly adjustment of the anti-backlash mechanism of the nut. Comparing the THK KR20 with a Velmex Unislide is like comparing a classical Mercedes with a Soviet-era Lada.

Relatively cheap, China-made industrial linear stages operated by screw micrometers have been flooding the market for several years. Surprisingly, I found that they are usually quite adequate as focusers and subject positioners in photomacrography, and far better than virtually all the focusing racks and subject positioners targeted to photographers that I have tried.

Further examples of focusing racks I used in the past can be seen here.

There are other brands and models of industrial motor-operated linear actuators on the market. I have little experience with any of them, and cannot say much about alternatives to the THK KR20, except that larger models, e.g. in the KR33 and KR45 series, may also be suitable. They are stiffer and can carry a higher load, but are also heavier, wider and significantly more expensive. Some models are equipped with two platforms moving along the same rail, one of which may be an idler platform (i.e. not operated by the screw). The function of an added idler platform is to reduce the already very low amount of twisting and play of a single platform and increasing the load bearing capabilities of the latter. Sometimes an idler rail, independent of the screw-operated rail, is used with an idler platform and for a similar purpose.

Extremely precise (on the nanometer scale) motorized linear actuators are occasionally available on the second-hand market. They are so delicate that they should never be operated by hand, never lifted by holding their actuator part, and preferably not touched at all. Considering the rough treatment typical of surplus industrial equipment, I would use extreme caution when purchasing any of these precision actuators in used condition.

Piezoelectric linear actuators have been the subject of interest by photographers on occasion, and are probably more tolerant of rough treatment, albeit their total travel is quite small. They are typically operated by an electronic controller. Sometimes a piezoelectric actuator is equipped with a piezoelectric sensor that can provide feedback to the controller to largely correct the non-linearities of the piezoelectric elements and their mechanical flexing frames.

My opinion of different models of focusing rails may seem excessively severe. However, when compared with the best industrial actuators, microscope focusing blocks and microscope subject positioners, the precision of focusing racks designed as consumer-level items for photographic use is typically orders of magnitude poorer.

External resources

Since the THK KR20 is primarily meant to be operated by a motor, the following resources can be useful:

  • (requires login) introductory-level discussion of choices when choosing a stepper motor driver and operating mode.
  • thread on using the THK KR20 with 400 steps/revolution stepper motor, and links to other threads discussing microstepping drivers.


The THK KR20 linear activator (and larger models) equipped with 1 mm-pitch screws are at present one of the best choices as a motorized, computer-controlled focuser for photomacrography within a reasonable magnification range, higher than feasible for consumer-level focusers designed for macrophotography use, but typically lower than provided by the stands and focusing blocks of compound microscopes.