INNOVATIVE TECHNOLOGIES THAT DRIVE PERFORMANCE

Besides using precision engineering fundamentals like error budget, force and torque symmetry... we think outside the box and implement new ideas:

A stage with high dynamic performance allows its motion to precisely follow a commanded move profile that includes drastic changes of velocity over a short period of time. Besides numerous other conditions and requirements that must be met in order to achieve high dynamic performance, one fundamental requirement is: The motor must be capable of accelerating the mass! The maximum achievable acceleration of a linear motor driven stage is approximately defined by the maximum Force Fm the motor is capable to deliver at a certain velocity divided by the total moving mass.

Chart 1 shows a characteristic relationship between Fm and velocity of a linear motor: Fm decreases with increasing velocity and therefore the acceleration capability of the stage decreases with increasing velocity.

The Extended Dynamics [ED] option expands the workspace (area under the graph) horizontally (max. velocity) by approximately 45% and vertically (Fm) by round about 90% as shown in chart 2. Depending on the application ED can also be configured for extreme velocity expansion (vmax increases by more than 80% and Fm increases by 10% as shown in chart 3) or for extreme force expansion (Fm increases by 95% and vmax by approximately 10% as shown in chart 4).

Our product families XL, XM, XG, XP and DynaMax use very high quality ironless linear motors with market leading output power/volume and output power/mass ratios. These motors combined with ED enable our products for unmatched dynamic performance.

ED Advantage: Improves precision and reduces move time without increasing size or changing form factors of the stage

The graphs in the following charts vary from stage model to stage model.

One of the most important key requirements for high dynamic performance is the bandwidth of the motion system. The bandwidth is a direct measurement of how precisely a system is capable to perform a commanded move profile. A system with a higher bandwidth is capable to react faster and with less deviation between the actual and the commanded move profile sent by the controller.

A little bit of theory:

The commanded move profile (velocity as a function of time) can be described as a summary of harmonic oscillations (Fourier). Each of the so called harmonics are presented by their amplitude, phase shift angle and frequency. Ideally all amplitudes, phase shift angles and frequencies of the actual move profile are identical to the corresponding values of the commanded move profile. In this case the actual motion would be identical to the commanded motion (error would be zero).

Back in the real world:

With real world systems this is never true: The actual motion is always delayed (phase shift error) and its amplitudes do not match the original ones, because the system responds different to each of the commanded frequencies. Some amplitudes of the actual move profile are smaller (damping) and some are bigger (resonance) compared to the corresponding ones in the commanded move profile. A system does not work properly with resonances, because it vibrates or even becomes unstable. There are different well known methods such as filtering or coupling masses to avoid resonances in a given system, but everyone of these methods add additional distortion into the actual move profile and therefore reduce dynamic performance.

ARS uses the most effective of all engineering approaches: Consistent minimization of dead time and maximization of control speed:

ARS technology is essentially based on consequently minimizing the signal deadtime between motor on the one hand and feedback on the other hand. Motor and feedback system are coupled mechanically and electronically (figure 1). Assuming the mechanical coupling is very stiff ARS actively damps resonances simply because the drive is able to react fast enough. Compared to conventional drives our drives include a faster and better feedback signal tracking and is capable of reacting up to 10 times faster (100kHz).

ARS Advantage: By actively damping mechanical resonances ARS increases the system bandwidth and therefore enables higher dynamic performance

Generally Direct Drive Technology accommodates the growing demand for higher throughput without sacrificing precision while in addition reducing the number of wearing parts in the system. But as usual, nothing is perfect: One downside is that every little disturbance (vibration, friction, external forces....) is directly coupled with the motor without any damping. As a result, the complete system is much more sensitive and much harder to control compared to using screws, gears or other mechanical elements between motor and the actual moving object.

The above mentioned disturbances cause deviations from the commanded path in all different shapes and form: Damped or undamped oscillations or aperiodic. With dead times of about 250µs, conventional drives are not able to react fast enough to correct those deviations. Most of these drives are built based on a cascaded loop architecture: Current loop is the fastest and position loop is the slowest, with the velocity loop in between. Just having a fast current loop is not efficient to reduce such deviations. Some competitors implement sophisticated control algorithms to compensate some of the effects after they have built up, but in the end their reaction time is still limited by the dead time.

Our market leading drives close position and velocity loops with 100kHz and therefore are four to ten times faster than any other drive available on the market. In combination with fast feedback systems we are able to reduce the dead time to 25µs and therefore react 10 times faster on disturbances compared to conventional drives which means deviations (between commanded and actual motion) caused by disturbances are much lower.

Advantages of 100kHz Servo Loop:

Ultra quiet and smooth motion (noise suppression); Highest servo stiffness and robust control; Minimum deviations between actual and commanded move profile.

Conventional drives typically use 16kHz to 20kHz 2 Level PWM (figure 1). These signals cause noise in the current loop and consequently in the velocity loop (velocity ripple). Especially at low damped high dynamic systems this is a limitation if ultra smooth motion is required. Also in applications with high demands for position stability <<1µm the PWM signal introduces higher frequency noise into the system that cannot be eliminated. Further more it causes power dissipation in the motors which can be clearly demonstrated be displaying the currents through the motor windings and compare the results with our drives.

Our drives use an advanced 100kHz 3 Level PWM Technology (figure 2) which is especially useful if higher speeds (higher voltages) are required and the motor inductance is in the single digit mH range (typical for our products). In this case the current raises very fast and if the PWM frequency is to slow (figure 1) the actual current (red) shows high peaks (overshoot). Also the three level switching is much closer to an ideal sinusoidal signal form compared to the 2 level PWM. This reduces the current ripple in the winding by a factor of 10.

Advantages of the 100kHz 3- Level PWM technology:

Much smoother motion (you can hear the difference); higher position stability; less power losses in the motor, less leaking current

One of the most challenging problems with applications using high precision motion systems is the fact that most of the materials used in such systems (aluminium, steel, granite, glass...) change their dimensions when temperature varies. According to Thermodynamics theory an input of thermal energy into a physical body causes it to grow. The amount of expansion depends on the body's material and is characterized by its thermal expansion coefficient. In addition the amount of expansion also depends on the dimensions of the body itself: If two bodies are made of the same material and experience the same input of thermal energy body 2 expands twice as much as body 1 if body 2 is twice as long as body 1.

For example: The temperature of two objects with an identical length of 1m raises by 2K. Object one is made of steel and object two is made of aluminium. The steel object grows approximately 26µm and the object made of aluminium expands approximately 46µm. If the two objects would be 2m long instead of 1m, the steel object would grow by 52µm and the aluminium object would expand by 92µm.

HOW DO THESE NUMBERS MATCH WITH AXIS SPECIFICATION OF SINGLE DIGIT MICOMETERS OR EVEN LESS?

♦ Ideal conditions in the labs:

All accuracy and repeatability specifications from suppliers of high precision positioning axes are assuming thermally static conditions (mostly 21°C). Axes are typically measured under ideal conditions in measurement labs with very stable constant temperatures, ideal damping conditions and the measurements occur after a slow move is completed and the axis reached target position (velocity =0) (minimum thermal input by the motor). There are many more aspects not mentioned here but the fundamental line is that during measurement thermal changes are more or less eliminated in the labs.

♦ Time matters:

It takes time before material expansion caused by an environmental temperature increase occurs. That time depends a lot on the volume (mass) of the stages in the system. In many cases the process time is short compared to the reaction time of the material to respond to the thermal input. Before the next process starts the system will be referenced using internal or external reference markers. This procedure avoids adding up thermal drifts over time.

As longer a process takes the influence of thermal impact increases and as a result the achievable precision of the system decreases.

♦ Size matters:

In many cases the critical process only occurs in a relatively small workspace. Based on the above mentioned thermodynamic relationship between temperature changes and material expansion the precision might not change to much in a small workspace.

If processes require high precision over a large workspace, thermal changes cannot be ignored since they have a large negative impact in the achievable precision of the system.

♦ Smart product design:

Some of the thermal effects can be minimized already in the axis design phase by placing the critical elements inside the axis at the best strategic locations. Using best engineering practices and the knowledge of the thermal behaviour of all critical components we minimize the effects of thermal changes in precision and repeatability already by design.

For an applications that requires either a large workspace or the process time is long compared to the thermal reaction time of the materials used in the system the achievable precision is very limited due to the thermal effects. Instead of achieving precision values of a view microns thermal effects can cause errors in the range of a view hundred microns.

Fundamentally there are only two ways to approach this problem: 1. Minimize environmental thermal changes- very complex and cost intensive (complexity and cost increase exponentially with expanding the workspace) or 2. Implement a compensation technology that enables the system to correct the errors caused by environmental changes.

3LTC is a predictive intelligent compensation technology developed by NTGmotion

that enables our motion systems to correct errors caused by thermal changes before they impact precision of the system. This also includes thermal effects caused by the motor heating inside the axis due to high dynamic, high duty cycle motion is required. 3LTC requires our optional integral closed loop cooling feature and is capable of correcting a large amount of the errors caused by temperature changes.

Advantages of 3LTC:

Cost effective solution for applications with demand for high precision over long time in large workspaces, no further complex and cost intensive equipment necessary(*).

(*) 3LTC compensates errors of the motion system caused by temperature changes. It does not compensate changes in the machine frame or drifts of other parts in the machine.

A feedback system (Encoder) is an essential component used in every Servo axis. In many cases so called analog incremental Encoders are used to feed the actual linear or rotary position back to the Servo drive. The Encoder system always consists of two components: The scale (passive) and the readhead (active). The readhead includes some sort of sensor (optical, magnetic...) and some electronics to transform the sensor signal into a format the Servo drive can read.

A very prevalent standard is 1Vpp sine / cosine because this signal form allows for significant multiplication of the raw signal resolution in order to meet demands like high precision, smooth motion, small step size... Mechanical misalignment of the readhead relative to the scale causes electrical misalignment between sine and cosine and also signal distortion which lead into noise or in extreme cases into feedback errors (drive cannot read the signal). Therefore an alignment process must always be part of the readhead installation. During this alignment process the readhead must be connected to a drive or controller and it need to perform some motion relative to the corresponding scale while the sine and cosine signals are being plotted with some sort of a soft oscilloscope.

If the signals are not distorted, the angle between sine and cosine is exactly 90deg and frequencies and amplitudes are identical, the displayed figure is an ideal circle if the sine signal is being displayed horizontally and cosine signal is being displayed vertically (readhead perfectly adjusted). In real systems the mechanical alignment is not good enough and the signals need to be electronically calibrated (tuning of feedback gains) until a square or elliptical figure becomes an ideal circle.

With conventional commercial drives/ controllers this calibration is done once during the axis startup and testing procedure. Temperature changes, system transport or mechanical tolerances (especially with long travel linear axes) change the location (distances and/or angles) between readhead and scale and therefore distort the previously adjusted circle.

Our Dynamix Servo drives continuously recalibrate all five parameters (offsets, amplitudes and phase) with a frequency of 100kHz in the background during the drive is performing motion to minimize the first order elliptical errors. A comparison between figure 1 and 2 shows a drastic reduction of the elliptical errors by a factor > 8.

Advantages of ACFC:

Approved position accuracy, reduced audible noise, Allows for setting the Servo D- gain significantly higher which increases the Servo stiffness, compensation of thermal effects in the readhead mounting block and no manual recalibration required