Abstract : This paper precisely analyzes the dynamic shearing working mode of a CNC rack and pinion machine, and discusses various factors affecting the accuracy of dynamic shearing and their adjustment and optimization methods. The paper details anti-interference measures for encoder pulse signals, the use of PLC counters, the relationship between servo motor acceleration/deceleration time and the dynamic shearing stroke difference, and the reasons for servo motor delayed start-up. This research provides valuable insights for the design and debugging of control systems for similar machines.
Keywords : dynamic shear encoder, counting signal, anti-interference, servo motor, delayed start
1. Operating mode and control system composition of machine tools
① The unwinding step is driven by a frequency converter to unwind the steel strip.
② The forming step completes the punching and forming.
③ The moving shearing platform (hereinafter referred to as the moving platform) completes the shearing of the rack.
This article discusses the dynamic shearing process of a moving shearing platform on a rack and its influencing factors.
1.1 Dynamic shearing process of the moving shearing platform on the rack
(1) The movement of the rack
The speed of the rack movement is determined by the speed of the forming step. The forming step is driven by a frequency converter. Its speed is a fixed value.
(2) The moving platform is driven by a servo motor of the CNC system. An encoder and a punching die are mounted on the moving platform. The encoder counts the teeth of the rack. The punching die performs dynamic punching.
(3) Dynamic punching
① During normal operation, the rack is driven by the forming step to run at a specified speed. After the rack enters the moving platform, the encoder installed on the moving platform counts the number of teeth on the rack. When the counting signal reaches the "start count value", the CNC system sends a positive start signal to the servo motor.
② The moving platform moves forward following the speed of the rack. When the counting signal reaches the "tooth length count value", the speed of the moving platform is equal to the speed of the rack. The system then sends a punching signal to cut the rack.
③ The mobile platform stops moving forward and moves back to the starting point in the reverse direction. It then waits for the next punching cycle.
1.2 Composition of the mobile platform control system
(1) The main control unit of the control system is Mitsubishi FX1S-20MR. The FX1S-20MR is responsible for receiving encoder counting signals and sending mobile platform start signals, punching signals, forward stop signals, and reverse start signals.
(2) The servo motor of the moving platform is a servo axis in the CNC system. Its forward start/stop and reverse start/stop signals are sent to the CNC system by the main control unit Mitsubishi FX1S-20MR.
(3) The encoder is a domestically produced encoder, with 32 pulses per revolution. This corresponds to 1 pulse per tooth of the rack. The pulse signal is connected to the FX1S-20MR.
2. Analysis of Dynamic Punching Modes on Mobile Platforms
After completing the PLC program for the mobile platform and setting the relevant operating parameters for the servo motors, a trial cut was performed on the racks. Five racks were tested at a time, and each operating parameter was evaluated. The results showed that the mobile platform's cycle time met production requirements, but the rack lengths varied. Despite testing various parameters, satisfactory results were still not obtained. Therefore, a careful analysis of the mobile platform's punching method and the factors affecting punching accuracy is necessary to identify the main reasons affecting punching accuracy.
2.1 Analysis of Dynamic Punching Mode of Mobile Platform
The dynamic punching process of the mobile platform is shown in Figure 2:
(1) AB phase:
After the rack enters the moving platform, the encoder installed on the moving platform counts the number of teeth on the rack. When the counting signal reaches the "start count value", the CNC system sends a positive start signal for the servo motor.
The mobile platform starts to accelerate. When the difference between the rack travel and the mobile platform travel equals the following travel, i.e. point B in Figure 2, the system sends a counting completion signal (the standard shearing length has been reached).
(2) BG stage:
The moving platform continues to accelerate, and the rack also continues to move. During this phase, the moving platform's speed has not yet reached the rack's speed, and there is relative movement between them. This relative movement during this phase is known as the "shearing length error."
(3) BC stage:
Speed following phase. The goal is to make the speed of the moving platform reach the speed of the rack, so that the speeds of the two are completely equal.
(4) CD stage. A punching start signal is issued at point C. Due to the electromechanical delay of about 200ms, the rack is actually cut off at point D.
(5) DE stage.
The encoder count is reset to zero at point E. Because the vibration from the punching process can cause the encoder to malfunction and emit pulses, the reset pulse is delayed until point E to eliminate this effect. Therefore, the relative motion between the rack and the moving platform (although very small) from point B to point E is not monitored by the encoder count. (In actual experiments, the cleanest shearing length was achieved when the reset signal was emitted at point B because the rack movement was monitored throughout.)
Regardless of whether the vibration pulse is positive or negative, it is cleared at point E.
However, if vibration pulses still occur during the E-F stage, two situations may arise:
① Positive pulse ------- Short teeth appear.
② Negative pulse ------ Long teeth appear.
(6) EF stage. Counter reset---moving platform forward stop stage. In this stage, it is necessary to ensure that the reset is completed before forward stop. In this stage, there has been a phenomenon where the reset time is delayed until the forward stop point, and the normal counting pulse is reset, resulting in a "long tooth" phenomenon.
2.2 PLC program for dynamic punching on mobile platform
Based on the analysis of the dynamic punching mode of the mobile platform, the PLC program for the motion part was developed:
Step 0 is to set the "filter coefficient" to increase the frequency at which the counting port X0 receives the counting signal.
Step 6 uses X0 as the input port for receiving encoder counting signals. Counter C1 counts the number of teeth for rack length. C3 counts the number of mobile platform startup signals.
Step 14: Counter C3 reaches its position, and the mobile platform start signal (Y0) is sent.
In step 35, counter C1 reaches its maximum value, and after a delay of T201, the punch instruction (Y5) is sent.
In step 37, counter C1 reaches its position, and after a delay of T202, a counter reset instruction is sent.
2.2 Factors affecting the accuracy of shearing length
Based on the analysis of the dynamic punching mode of the mobile platform, the factors affecting the accuracy of the shearing length are summarized as follows;
(1) Encoder pulse signal
(2) Synchronous punching ----- During punching, if there is relative movement between the moving platform and the rack, the punching length cannot be guaranteed. To achieve synchronous punching, the following time must be adjusted, namely the CD stage in Figure 2 and T201 in the PLC program.
(3) Servo motor acceleration time
3. Further analysis and optimization of factors affecting punching accuracy.
3.1 Influence of Encoder Pulse Signal
Encoder pulse signal—The encoder pulse signal is fundamental for controlling the movement of the mobile platform and issuing the punching signal. If interference signals enter the counter, the punching length will be shorter. If the encoder pulse signal is missed, the punching length will be longer. Therefore, when analyzing the punching rack length, if the rack is too long or too short, the first thing to consider is that the counting pulse is abnormal.
The encoder used in the rack and pinion machine is a domestically produced encoder, specifically designed for the rack and pinion machine.
32 pulses per revolution, corresponding to 1 pulse per tooth.
The rack tooth pitch is 6mm. When the rack speed is 13000mm/min, the corresponding pulse frequency is 36Hz, while the signal frequency acceptable by the PLC's conventional interface is 25Hz. Therefore, the conventional interface cannot be used directly.
(1) Use a high-speed counter
The Mitsubishi FX1SPLC features a high-speed counter function. Initially, a single-phase high-speed counter C235 was used. However, the C235 is easily affected by interference. When the encoder signal was connected to the high-speed counter, it was observed on the PLC monitoring screen that the counter data immediately became erratic once the encoder rotated. Even when the encoder was not rotating, the counter value increased irregularly, clearly indicating interference. (Improper wiring can also exacerbate interference; in the field, the lack of terminal blocks during wiring resulted in increased interference, which was reduced after using terminal blocks.)
Using the C251 two-phase high-speed counter significantly reduced interference. However, it was still unstable. The counting was stable for some time periods, but unstable for others. Since the PLC controller, CNC servo system, and frequency converter were all housed in the same control cabinet, the CNC servo system and frequency converter clearly caused serious interference to the PLC. After numerous tests, the high-speed counter solution was abandoned.
(2) Using a regular counter
The challenge of increasing the received signal frequency when using a standard counter is to shorten the input signal filtering time. Mitsubishi PLCs have a function to shorten this filtering time by setting a digital value in the D8020. See step 0 of the PLC program in Figure 3. This method can increase the received signal frequency to 50Hz.
This satisfies the rack's operating speed requirements. However, the input signal filtering time cannot be set too small, as this reduces its anti-interference capability. The optimal values must be determined through trial and error. The optimal values are D8020 = 3-5.
The following measures were taken to mitigate interference for the encoder in the field:
① The encoder shield wire is grounded.
② Wear a metal tube alone.
3.2 Acceleration Time
The most important stage of dynamic shearing is the acceleration and following stage of the moving platform, i.e., the AG stage in Figure 2. Before discussing this stage, the relevant motion parameters must be given.
(1) Relevant motion parameters
①Rack length ------ expressed in terms of the number of teeth. For example, 200 teeth.
② Tooth pitch L-----------Unit: mm.
③ Follow the number of teeth N (distance) ----- a reserved travel distance. Within this travel distance, the moving platform accelerates to reach the rack running speed.
④ Rack and pinion speed V----mm/second
⑤ Acceleration time T-----The time it takes for the mobile platform to accelerate to the rack speed
(2) Calculation of “travel difference” and “acceleration time”
During the acceleration phase:
According to Equation 1 and Figure 2, the acceleration time T determines the travel difference during the acceleration phase. Theoretically, as long as the acceleration time is precisely adjusted, the speed of the moving platform can be made equal to the speed of the rack at the same time point when "travel difference = following distance".
During actual debugging, first determine the number of following teeth (distance) according to Equation 2; then precisely adjust the acceleration time according to Equation 3. The principle is to extend the acceleration segment as much as possible within the total travel range of the moving platform. This is because a longer acceleration time results in smoother acceleration, avoiding acceleration oscillations caused by too short an acceleration time, which would affect the stability of the synchronization speed.
3.3 Counter Reset Time
Counter Reset Time – During multiple punching processes, it was observed that the total length of the rack was frequently 1-2 teeth shorter. This shorting must be due to an abnormal pulse entering the rack. What component is generating this extra pulse? Through experimentation and comparison, it was found that during dynamic punching, the punching process generates significant mechanical vibration. Since both the encoder and the punching die are mounted on a moving platform, the punching vibration causes the encoder to jitter, sometimes emitting a single pulse signal. This pulse signal is counted in the normal count value, thus causing the rack length to be 1 tooth shorter.
To eliminate this effect, the counter's reset point must be scheduled after the punching is completed.
Extend the time period by another step, to point "E" in Figure 2. This way, even if a vibration pulse enters the counter, it will be cleared at point "E". The next counting cycle then restarts from point "E". (Theoretically, the counter should be reset at point "G", meaning it should be reset immediately upon reaching the current count value, and the next counting cycle can begin.)
In the PLC program, the counter reset time is T202. The counter reset time must be repeatedly tested to obtain the optimal value.
4. Experimental Results and Key Factors
4.1 Anti-interference measures and experimental procedures
To eliminate the influence of electromagnetic interference waves, the following measures were taken:
(1) Remove the PLC controller from the control cabinet and supply it with a separate AC220V power supply. Ground the PLC. Completely enclose the PLC in another metal cabinet. Make the PLC section completely independent. Eliminate the influence of interference.
(2) Experiment with rack running speed = 13 meters
Related parameters follow the number of teeth = 10, acceleration/deceleration mode: linear acceleration/deceleration.
Acceleration/deceleration time: 360-400ms
Cutting results: Most lengths = 1015-1022, with some being 4-10mm longer.
Adjusting the servo motor's acceleration and deceleration time has an effect, but even at its best, the length is still 4-10mm too long. The length varies even with the same set of parameters. Even when several sets of cuts are of the same length, the length is still too long.
(3) Experiment with rack running speed = 8 meters
Related parameters follow the number of teeth = 7, acceleration/deceleration mode: linear acceleration/deceleration.
Acceleration/deceleration time: 300-360ms
Cutting results: Most lengths were 1015-1022 mm, which was 4-10 mm too long. There were also 1-2 pieces in one group that were 5 mm too short.
4.2 Analysis of Experimental Results
Experiments were conducted on the dynamic punching of the mobile platform using various parameters, but the punching effect remained poor. Under the same set of parameters, the length of the punching rack varied. The experimental results are shown in Table 1.
Table 1 Dynamic punching test record
rack speed = 13000 mm/min | Follower number of teeth = 10 | ||
Standard length = 1012mm | The delay time is approximately 125ms. Delayed trip 27mm (S4) | Calculate length error | |
Acceleration time #2004 (ms) | punching length | Acceleration phase travel difference S3=109*T0 | |
500 | 1035 | 54.5 + 27 = 81 | 81-63=18 |
400 | 1022 | 43.6 + 27 = 70.6 | 70.6 - 63 = 8 |
360 | 1017 | 39.24 + 27 = 66.24 | 66-63=3 |
350 | 1016–1015 | 38.15 + 27 = 65.15 | 65.13-63=2 |
340 | 1016 | 37 + 27 = 64 | 64-63=1 |
Based on the above data, the delay is approximately 29mm. | |||
The experimental data in Table 1 show that adjusting the acceleration time is effective; as the acceleration time gradually decreases, the length of the shearing rack gradually approaches the standard length, but it cannot reach the standard value. Furthermore, the lengths of a set of racks are inconsistent. Among the factors affecting punching accuracy, the effects of interference and missed pulses (which reduce operating speed) have been eliminated. Moreover, the acceleration time, synchronization time, and zeroing time have been repeatedly adjusted and are under control. However, the punching length data is still so scattered. Therefore, there must be an "uncontrolled factor" or "random factor" at play.
5. Identify key factors
5.1 Impact of Delay Time
Upon further analysis of the "dynamic punching mode of the mobile platform" and careful observation of the actual punching process, it was found that there is a delay in the startup of the mobile platform—that is, there is a delay of about 120ms from the time the PLC sends the start signal to the actual startup of the mobile platform.
The control system of the rack and pinion machine consists of a PLC and a NC. The signal transmission process and time between the PLC and the NC are as follows:
(1) The PLC is responsible for receiving the counting signal, and after processing it, it sends out the start signal of the mobile platform. The "PLC scan cycle + output delay" is about 20ms.
(2) The start signal is sent to the CNC system and processed, which takes about 60ms.
(3) The CNC controller sends the servo axis start signal to the servo driver via the bus. This process takes approximately 40ms.
Therefore, the total latency is approximately 100-120ms. This time is determined by the system hardware performance and is beyond our control.
During this delay period, (when the rack is running at a speed of 13000 mm/min) the rack has moved about 29 mm.
In the dynamic punching mode shown in Figure 2, the 0-A stage is the delay stage.
The formula for calculating the travel difference must be revised to:
In the control system of a rack and pinion machine, the delay time is not a stable value, which greatly affects the cutting accuracy of the rack and pinion.
5.2 Rectification Measures and Effects
To mitigate the impact of delays, the following measures were taken:
(1) Replace the mobile platform drive system with one directly controlled by the PLC. This reduces the number of intermediate signal transmission links.
(2) Reduce the rack running speed.
After the above processing, the dynamic shearing accuracy of the mobile platform is guaranteed.
6. Conclusion:
Dynamic punching differs from static punching. In static punching, a 100ms delay has no impact on punching accuracy, while in dynamic punching, the delay becomes the main factor affecting shearing accuracy. Maintaining synchronous operation between the moving platform and the rack is also fundamental to dynamic shearing.
