How to manage thermal effects in precision scanning systems
General precision engineering principles dictate that care be given to many important design details to ensure success in any motion control application.
General precision engineering principles dictate that care be given to many important design details to ensure success in any motion control application. For instance, best practices in machining and assembly offer the best chance to produce a high-performing system, and special attention must be paid to preserve a stable working environment. Specifically, managing and mitigating thermal effects is of paramount importance.
FIGURE 1. Galvo scanners offer very high bandwidth, which allows for faster processing times and, therefore, higher overall throughput in laser micromachining and additive manufacturing applications.
Galvanometer (galvo) scanners are finding increased use in laser micromachining and additive manufacturing applications (FIGURE 1). Their very high bandwidth allows for faster processing times compared to traditional servo stage configurations (i.e., “fixed point” laser systems), which infers much higher overall throughput. While the importance of managing system heating and the resulting distortions has long been appreciated in servo-driven stages, actuators, and systems, overcoming analogous effects in modern high-speed scanning applications has not necessarily benefited from the same level of attention. However, investigating the root sources of galvanometer system heating will yield direct insight into how designers may best manage this issue and improve overall system performance.
Thermal distortions within stage-based systems typically manifest as growth in the stages themselves. The heat source is generally the stage motor and, while often nonlinear, the effects may be measured and perhaps mitigated. For instance, clever placement of temperature sensors may allow the motion controller to directly measure a thermal change and compensate for the stage’s growth or contraction by altering the commanded motion profile. Accommodating thermal distortion in scanning systems is not quite as straightforward, as the heat sources (and therefore the mitigation strategies) are not necessarily the same. Before investigating specific heat sources and their proposed remedies, we will review the scanning system components most susceptible to heating effects.
Scanning system components
Perhaps the most obvious negative influence of heating on a scanning system is in distortion of the aluminum block/housing that mounts and aligns the galvo motors/mirrors to one another. However, as we will see, the source of heating is not always obvious and so mitigation might entail a number of design strategies. Developing temperature gradients over the mounting block serves to misalign the mirrors from each other, and develops a number of nonlinear “drift” errors that corrupt established performance and calibrations.
Motor heating is an obvious concern. The windings in the motors can sustain only a finite temperature rise before failing entirely. Furthermore, overheating can cause the mirror mounts to soften and fail, changing the system stiffness and response at a minimum, or perhaps even resulting in catastrophic failure if the mirror comes free entirely.
The position sensors within the galvo motors (often optical encoders or some other analog optical device) typically demonstrate nonlinear behavior with changing temperature as well. This performance degradation manifests as thermal- or time-based drift, and is often a primary source of overall system error.
Lastly, the reflectivity of the mirror coatings themselves exhibit a temperature dependency. Users may find that coatings optimized for a given laser wavelength at room temperature may lose efficiency at warmer temperatures, as the reflectivity peak tends to shift to higher λ values.
As you might have guessed, solutions to the above issues are as varied as the effects themselves.
The scanner mirrors are directly susceptible to overheating due to laser energy. Although the mirror coatings are engineered to be extremely efficient (typically 99.5% reflective or better at the wavelength of interest), some portion of incident laser energy is absorbed by the mirrors as heat. In addition to the overall system thermal distortion this heating may induce, too much power absorption can cause the mirror mount to fail and come loose from its motor.
Remedies for this heat source include using larger mirrors (so that a constant power absorption causes less heating) or implementing air cooling on the back side of the mirrors to carry away heat. Cooling in this manner can improve the laser power handling capability of the mirrors by upwards of a factor of 3, but could introduce a small amount of in-position dither as the air impinges on the mirrors. Custom coatings offering higher (99.8% or better) reflectivity are possible as well, but often come at the expense of substantial non-recurring charges and extended lead times.
Laser power incident on the galvo mounting block itself is also a major source of heat in scanning systems. Most free-space delivered laser beams have a Gaussian profile to their power distribution, with a defined beam diameter linked to their “1/e2” value. This value (equal to about 13.6% of the Gaussian peak) infers that about 4.4% of the beam energy is not contained within the nominal diameter (and indeed—the tails of the Gaussian distribution roll off infinitely). For example, if a beam with a stated diameter of 14 mm shines on a galvo housing with a 14 mm aperture, 4.4% of the beam energy will actually be spent warming the aluminum block. Furthermore, beam misalignment may place even more laser energy incident on the housing instead of shining it through the block’s input aperture.
Obviously, perfecting the incoming beam alignment will mitigate some of this heating. Using a laser with an Airy disk (or “top hat”), uniform power distribution can remove most of this effect, and implementing a chilling circuit is also a good solution—especially if the water circuit is properly designed to carry heat away from the beam input aperture. In fact, using water cooling can not only remove heat from the system, but can also be used to establish a stable working temperature, even if no heating concerns exist within the galvo itself. If outside influences (for instance, neighboring power supplies or other lab equipment) eject unmanageable amounts of heat, a water cooling circuit in the galvanometer can grant some level of thermal immunity to the scanning system, even in the presence of other environmentally disruptive devices (FIGURE 2).
FIGURE 2. An example of a water cooling circuit for a scanning system is shown.
Implementing water cooling can have still more beneficial effects. In some very aggressive applications, such as via hole drilling, the motors are commanded to make high-acceleration moves at extreme duty cycles. This causes very large current draw into the motor windings, and the resulting I2R losses are a source of heat in the system. This might be managed by clever use of motion profile optimizations, such as “sky-writing,” corner rounding, or velocity profiling. These tools, available in some advanced motion controllers, can lower the root-mean-squared acceleration values in the motion profile, thus reducing the current commanded in the motor windings.
Alternatively, the same cooling circuit that services the laser input aperture might also pass fluid through a water jacket on the galvo motors, carrying away that heat before thermal runaway occurs. A properly designed water cooling circuit is extremely effective at managing system thermal stability (FIGURE 3).
FIGURE 3. A properly designed water cooling circuit is extremely effective at managing system thermal stability (a) when compared to a system without one (b).
Although not many applications have motion profiles demanding enough to draw large current values that warrant motor cooling, water cooling seems to be fairly common in high dynamic scanning systems. The reason for this is perhaps not very intuitive. In addition to heating caused by resistive losses in the galvo motor windings, a substantial amount of heat (indeed, often more than from the motor losses) is ejected from the power amplifiers housed within the galvo mounting block. Modern switched-mode power supplies are efficient but not ideal, and often can lose upwards of 10–15% of their throughput power as heat.
As inferred by FIGURE 3, power losses far greater than what should be experienced by the galvo motors are needed to cause substantial temperature rise in a cooled mounting block. An interesting conclusion to be drawn from this is to consider scanner designs that remove the power electronics from the mounting block entirely. The resulting environment may no longer even need water cooling, and at the very least will be far more stable.
The proliferation of higher-precision laser scanning systems mandates a critical look at their environment controls and considerations. A sober assessment of heat sources yields a short list of possible remedies to sources of temperature problems. The end result of these exercises will be more accurate galvo motion and more precise parts being produced by the scanning systems.
Just as damaging is the influence of slowly changing (“drifting”) temperatures. Even a relatively small heat source can have a dramatic effect on system stability, so all efforts should be considered to maintain the most stable environment possible.
While this discussion has attempted to address a few of the most common thermal concerns, much more work is needed to completely address these types of concerns. For instance, perhaps a system may be characterized by its thermal repeatability (how reproducible are the thermal-induced errors) and in essence calibrated for these effects. Or, what lessons should we attempt to learn that may guide us in designing a mounting block that is most resistant to thermal distortion? Lastly, how could design choice of the galvo feedback devices influence overall system thermal immunity? Some technologies are likely to be better than others. Can the scanner controller’s digital control algorithm be modified according to the changing system temperature to overcome thermal degradation? Future investigation and analysis will undoubtedly improve accuracy and performance of these dynamic systems.