Test and Improvement of a Fuze MEMS Setback Arming Device Based on the EDM Process

This paper introduces the working principle of a MEMS safety and arming (S&A) device for a fuze that is installed perpendicular to the axis of the projectile. Additionally, the application of low-speed wire electrical discharge machining (EDM) in the fabrication of the device is proposed. Microsprings are susceptible to flexural deformation and secondary deformation in the EDM process, a problem that is solved by designing the auxiliary support beam, using multiple cuts, destress annealing and optimizing the processing parameters. The difficult problem of setback slider deformation in the principle prototype test is properly solved by establishing V-shaped grooves at both ends of the setback slider. The connection mode between the microspring and the frame is changed to a clearance fit connection. The improved setback arming device can guarantee service process safety and launch reliability. The maximum overload that can be withstood in service processing is 20,000 g, and the minimum overload for safety release during launch is 12,000 g. The results show that the EDM process can greatly reduce the machining cost while improving the machining precision and machining speed, which can compensate for the defects of the current manufacturing technology.


Introduction
In recent years, with the development of intelligent ammunition technology, modern fuzes have been endowed with more functions, such as flight control, explosion point control, trajectory correction, target recognition, and positioning, in addition to state control and initiation control. However, the volume limitations of traditional fuzes restrict the expansion of these functions, and this fuze-development problem has been effectively solved by the emergence of MEMS fuzes [1][2][3][4][5]. A setback arming device is the core device in a MEMS fuze and is used to ensure service process safety and launch reliability [6][7][8][9][10]. Most setback arming devices are oriented parallel to the projectile axis when installed, meaning that the explosive train is not in the same straight line [11][12][13]. Therefore, the direction of detonation energy transmission needs to be changed, resulting in energy loss. In serious cases, the detonator may not be able to detonate the warhead charge, resulting in a "dud" phenomenon [14,15]. To create MEMS fuzes with higher performance and more space to achieve additional functions, research on MEMS setback arming devices that are placed perpendicular to the projectile axis (hereinafter referred to as "vertical-frame-based setback arming devices") has gradually become popular in recent years [16][17][18][19][20]. Figure 1 shows a vertical-frame-based setback arming device that was proposed by Koehler et al. [21], which consists of a setback cantilever beam, an arming elastic beam, an arming slider, a head latch, and a cassette latch. Generally, the arming slider is locked in the safe position by the setback cantilever beam and the arming elastic beam. When a projectile is launched, the Micromachines 2022, 13, 292 2 of 21 setback cantilever beam moves downward under the action of the setback force and breaks away from the arming slider, thus removing the restriction on the arming slider.
Koehler et al. [21], which consists of a setback cantilever beam arming slider, a head latch, and a cassette latch. Generally, t the safe position by the setback cantilever beam and the ar projectile is launched, the setback cantilever beam moves do the setback force and breaks away from the arming slider, th on the arming slider.  Figure 2 is a cantilever lock setback arming device prop stalled perpendicular to the axis of the projectile. The respons under the two acceleration environments of service processin alyzed theoretically and verified by simulation. The results s setback arming device can effectively distinguish service pr [23] improved the design on this basis and proposed a micr device, as shown in Figure 3. The scheme with a tilted micr effectively reduce the axial dimension of the device, and the verified by theoretical calculations and simulation analysis.   Figure 2 is a cantilever lock setback arming device proposed by Li [22], which is installed perpendicular to the axis of the projectile. The response characteristics of the device under the two acceleration environments of service processing and normal launch are analyzed theoretically and verified by simulation. The results show that the cantilever lock setback arming device can effectively distinguish service processing from launch. Tong [23] improved the design on this basis and proposed a microinertial pin setback arming device, as shown in Figure 3. The scheme with a tilted microinertia pin was adopted to effectively reduce the axial dimension of the device, and the feasibility of the device was verified by theoretical calculations and simulation analysis.  Figure 2 is a cantilever lock setback arming device p stalled perpendicular to the axis of the projectile. The respo under the two acceleration environments of service proces alyzed theoretically and verified by simulation. The resul setback arming device can effectively distinguish service [23] improved the design on this basis and proposed a m device, as shown in Figure 3. The scheme with a tilted m effectively reduce the axial dimension of the device, and verified by theoretical calculations and simulation analys     The structures designed by Koehler [21], Li [22] and Tong [23] all use environmental forces to release the safety. The research mainly focuses on theoretical calculations and simulation analysis, lacking effective test verification, and does not involve research on the manufacturing process. For the fabrication of MEMS setback arming devices, nonsilicon-based micromachining technology (generally a lithography electroforming micro molding (LIGA) process or an ultraviolet-LIGA (UV-LIGA) process) and silicon-based micromachining technology (generally a deep reactive-ion etching (DRIE) process or a silicon-on-insulator (SOI) process) are commonly used. The cost of these processes is high, and no cheap mass production method has been found [24,25]. To solve the above problems, a vertical-frame-based setback arming device that also uses environmental force to release the safety is proposed in this paper, and low-speed wire electrical discharge machining (EDM) is successfully applied to the fabrication of the device. The principle prototype has the advantages of high precision and low cost. A mechanical impact test and centrifugal overload test were carried out on the principle prototype, and its structure was subsequently improved according to the problems identified in the tests. The improved principle prototype can accurately identify the service processing and launch stages. The results provide technical support for promoting the application and popularization of MEMS fuzes in the field of intelligent ammunition. Figure 4 shows a MEMS safety and arming (S&A) device applied to a certain type of 105 mm grenade, and its size is 12 mm × 10 mm × 0.8 mm. When the device is installed, it is oriented perpendicular to the axis of the projectile, and it mainly includes a setback arming device and arming device. The setback arming device is composed of a setback slider, a microspring, and a frame, and the arming device is composed of a driving wheel, a pin pusher, an arming slider, a head latch, and a cassette latch. The setback arming device is mainly studied in this paper. The application background of the requirements is as follows: the peak value of the most dangerous drop overload in the service processing stage is 18,000 g, the duration is 0.1 ms, the peak value of the launch overload is 12,000 g, and the duration is 9 ms. This kind of setback arming device has the advantage of a simple structure and can be used to replace the setback slider and the zigzag slot in the traditional device with a pair of inclined setback sliders and arming sliders. In the initial state, the setback slider and the arming slider are tightly assembled together and connected to the frame with microsprings. The ANSYS/LS-DYNA finite element software simulation shows that when the slope angle of the setback slider and the arming slider is set to 87 • , the service processing and launch stages can be distinguished by controlling the movement time of the setback slider, as shown in Figure 5. The service processing action time is very short, and the inertial force induced on the setback slider is not high enough to overcome the friction between the inclined surfaces. When the inertial force disappears, the setback slider returns to its original position under the action of the microspring restoring force. The launch action time is relatively long. The setback slider moves away from the arming slider under the action of a continuous setback force. In the process of movement, the setback slider is squeezed by the arming slider, undergoes a deflection deformation, and returns to its original shape after passing through the inclined plane. At this time, the setback slider cannot return to the initial position due to the obstruction of the arming slider, and the process of safety release is shown in Figure 6.

Process Selection
The difficulties in the fabrication of the setback arming device are mainly related to

Process Selection
The difficulties in the fabrication of the setback arming device are mainly related to two aspects. On one hand, the thickness of the microspring is 0.4 mm, while the thickness of the frame is 0.8 mm, and the microspring and the frame with different thicknesses need to be processed into a composite device. On the other hand, the contact surface between the setback slider and the arming slider is not a plane but needs to be processed into an inclined plane. The UV-LIGA process has the highest technical maturity among current nonsilicon-based micromachining technologies, while the EDM process has unique advantages in the fabrication of microparts [26][27][28][29][30]. Figure 7 shows a head latch of the MEMS S&A device fabricated by the UV-LIGA process and the EDM process, using 10 of each for process comparison, and the results are shown in Table 1. It can be seen from the table that the EDM process has the advantages of high precision, high speed, and low cost. Therefore, the EDM process is applied to manufacture the setback arming device through the technical research in this paper, and the combined electrode fixture is used to enable simultaneous processing of multiple electrodes. This method can be used to greatly reduce the machining cost while improving the machining precision and machining speed.

Processing Scheme
Microspring machining is technically difficult in the fabrication of setback arming devices by EDM; this is the main reason why the EDM process cannot be used in setback arming device manufacturing. Due to the size effect and residual stress, microsprings are prone to flexural deformation and secondary deformation in the machining process. In this paper, a high-purity hard nickel plate is selected as the processing material, and the processing machine is a CA30 immersion precision CNC low-speed wire cutting machine produced by Agie Charmilles Company (Biel, Switzerland). The processing scheme that involves orienting the auxiliary support beam in the longitudinal direction of the mi-

Processing Scheme
Microspring machining is technically difficult in the fabrication of setback arming devices by EDM; this is the main reason why the EDM process cannot be used in setback arming device manufacturing. Due to the size effect and residual stress, microsprings are prone to flexural deformation and secondary deformation in the machining process. In this paper, a high-purity hard nickel plate is selected as the processing material, and the processing machine is a CA30 immersion precision CNC low-speed wire cutting machine produced by Agie Charmilles Company (Biel, Switzerland). The processing scheme that involves orienting the auxiliary support beam in the longitudinal direction of the microspring is adopted, as shown in Figure 8a. After a trial cutting of the microspring, it is found that when the width of the support beam is set to 0.06 mm, the processed microspring undergoes a large deflection deformation and deviates to one side of the support beam, as shown in Figure 8b. On one hand, this phenomenon may be due to the long EDM discharge time that is required when the support beam with this width is removed, leading to a longer time during which the flushing pressure and discharge reaction force act on the microspring and excessive residual stress after machining. On the other hand, this phenomenon may be related to the position of the support beam. The support beam is located on one side of the microspring end, providing an uneven support force that results in a stress concentration when the support beam is removed.

Parameter/μm
Due to the small size of the inclined planes on both sides of the setback slider, if two inclined planes are processed at the same time, the requirements for preparing the formed electrode are higher. Therefore, the method for separately processing each single inclined plane is adopted in this paper. The surface to be processed is fixed at an incline during processing, and the bottom and side of the electrode wire were used for processing. The frame with an auxiliary support beam was then heated to 600 °C in a two-chamber vacuum permeable furnace for destress annealing treatment. After two hours, the device was placed in air to naturally cool [32]. After this step, the residual stress of the microspring was smaller, and there was no secondary deformation due to the high temperature during processing. Finally, the auxiliary support beam was removed using the EDM process.     To reduce the flexural deformation in the machining process, a method of increasing the pulse interval, reducing the peak current, and reducing the flushing pressure was proposed in reference [31] to reduce the reaction force during EDM discharge. After trial cutting, it is found that the deflection deformation of the microspring can be improved to a certain extent by adjusting the processing parameters, but the effect is not evident. Therefore, this paper adopts the method of optimizing the processing parameters through energetic control and changing the processing scheme to reduce the deflection deformation. The processing parameters used are shown in Table 2. Figure 9a shows the modified processing scheme, in which the support beam is set at the center of the microspring end, and the width is reduced to 0.04 mm. With this scheme, the action time of the EDM process decreases when the support beam is removed, the deflection deformation of the microspring after processing is very small and negligible, and the shape accuracy of the microspring is good without inclination, as shown in Figure 9b.     Figure 10 shows the processing scheme of the frame. The auxiliary support beam with a width of 0.04 mm divides the frame into two processing areas, A and B. Before processing, bolt holes are machined around the frame with an EDM prepiercing machine, and different machining allowances and parameters are selected to produce multiple cuts between A and B according to the cutting path planned in the scheme. This method can not only be used to improve the quality of the processing surface but also to redistribute the residual stress and correct the deflection deformation of the microspring. In this paper, a total of five cuttings were performed. After the first cutting, the surface quality of the frame was poor, and many adhesives were attached. This is because the first cutting is rough machining, the discharge energy is large, and the frame surface is seriously ablated. In addition, due to the size effect, the flushing pressure during cutting cannot be too large, the adhesive cannot be effectively removed, and the surface roughness Ra is 1.43 µm. Figure 11 is a schematic diagram of multiple cutting, the relationship of the size is: Micromachines 2022, 13, 292 8 of 2 Figure 10. Frame processing scheme.    Figure 12 shows the sample comprising the microspring and frame after processing This figure shows that the microspring has a small flexural deformation and good con sistency. The microspring and setback slider were measured after processing with stere omicroscopy. The measurement results are shown in Figure 13. The measured values o the main parameters are similar to the design values, indicating that the EDM processing scheme proposed in this paper has high precision and can meet the processing require ments. The principle prototype obtained after assembly is shown in Figure 14.  M n is the offset of the electrode wire at the n th cutting; M n+1 is the offset of the electrode wire at the n + 1 th cutting; ∆M is the offset increment after two adjacent cuts; m is the surface height difference after two adjacent cuts, which can be measured directly; R is the radius of electrode wire, brass wire with a diameter of 100 µm was selected; a n+1 is the unilateral discharge gap at the n th cutting, and the unilateral discharge gap for the five cuttings are shown in Table 3. Since the offset of the electrode wire does not correspond to the given value of the machine tool, it is necessary to correct the offset of the electrode wire one by one during multiple cutting. To ensure the machining accuracy, the offset of the electrode wire in the last cutting should be the sum of the unilateral discharge gap and the electrode wire radius. For the limitation of discharge gap, machining allowance, and machining space, the corrected electrode wire offset is shown in Table 4. The final surface roughness Ra can reach 0.49 µm after five cuttings with the offset, which is reduced by 65.7% compared with the first cutting. Due to the small size of the inclined planes on both sides of the setback slider, if two inclined planes are processed at the same time, the requirements for preparing the formed electrode are higher. Therefore, the method for separately processing each single inclined plane is adopted in this paper. The surface to be processed is fixed at an incline during processing, and the bottom and side of the electrode wire were used for processing. The frame with an auxiliary support beam was then heated to 600 • C in a two-chamber vacuum permeable furnace for destress annealing treatment. After two hours, the device was placed in air to naturally cool [32]. After this step, the residual stress of the microspring was smaller, and there was no secondary deformation due to the high temperature during processing. Finally, the auxiliary support beam was removed using the EDM process. Figure 12 shows the sample comprising the microspring and frame after processing. This figure shows that the microspring has a small flexural deformation and good consistency. The microspring and setback slider were measured after processing with stereomicroscopy. The measurement results are shown in Figure 13. The measured values of the main parameters are similar to the design values, indicating that the EDM processing scheme proposed in this paper has high precision and can meet the processing requirements. The principle prototype obtained after assembly is shown in Figure 14. Figure 11. A schematic diagram of multiple cutting. Figure 12 shows the sample comprising the microspring and frame after processing. This figure shows that the microspring has a small flexural deformation and good consistency. The microspring and setback slider were measured after processing with stereomicroscopy. The measurement results are shown in Figure 13. The measured values of the main parameters are similar to the design values, indicating that the EDM processing scheme proposed in this paper has high precision and can meet the processing requirements. The principle prototype obtained after assembly is shown in Figure 14.

Testing the Setback Arming Device
According to the regulations in GJB573A-1998 "Fuse Environment and Performance Test Method", the setback arming device needs to endure mechanical impact tests and centrifugal overload tests.

Mechanical Impact Test Results
A servo vertical impact testing machine was used for the mechanical impact test, and 10 principle prototypes were used for testing. A half sinusoidal pulse with a duration of 200 µs and a peak value of 100 µs was used to simulate the setback acceleration environment during service processing. The test results show that when the peak acceleration decreased from 23,000 g to 14,000 g, the setback slider remained in the initial position, and the safety of the setback arming device was not released. The setback arming device can guarantee the safety of service processing.

Centrifugal Overload Test Results
A rotating arm centrifugal testing machine was used for the centrifugal overload test, and 10 principle prototypes were used for testing. The centrifugal force generated by rotation was used to simulate the setback force during launch. The test results show that when the centrifugal acceleration decreased from 17,000 g to 8000 g, the setback slider remained in the initial position, and the safety of the setback arming device was not released. The setback arming device cannot guarantee the reliability of the launch. This phenomenon may be due to the large bending deformation of the setback slider when the setback arming device is released, where the acceleration required for deformation is very large, so the test conditions cannot meet the deformation requirements.

Improvement of the Setback Arming Device
The results of the mechanical impact test and centrifugal overload test show that the existing setback arming device cannot effectively distinguish service processing from launch, and its structure needs to be improved.

Improvement Scheme of Setback Slider
To solve the difficult problem of the setback slider deformation in the setback arming device, a scheme including V-shaped grooves at both ends of the slider can be used for improvement. On the premise of ensuring the safety of service processing, the acceleration required for flexural deformation of the setback slider during launch can be reduced. As shown in Figure 15, the reasonable selection of the distance s from the vertical sidewall to the inclined plane, the distance l between the vertical sidewalls, the vertical sidewall depth h, and the sharp angle α of the two V-shaped grooves are the key factors for enabling the setback arming device to meet the design requirements.  The setback load in service processing is set to 18,000 g in the ANSYS/LS-DYNA software. To observe the displacement of the setback slider after the load is removed, the sim- In this paper, ANSYS/LS-DYNA finite element software is used to simulate the movement of the setback slider in the launch and service processing stages by adjusting the distance s from the vertical sidewalls of the two V-shaped grooves to the inclined plane, the distance l between the vertical sidewalls, the depth h of the vertical sidewalls and the sharp angle α. Through these simulations, the reasonable ranges for s, l, h, and α can be found. The specific steps are as follows: the distances s and l are gradually increased and the depth h and angle α are gradually decreased in the launch stage until the setback slider is able to completely release the safety. At this time, s, l, h, and α are at their minimum values. The distances s and l are gradually decreased and the depth h and angle α are gradually increased in the service processing stage until the setback slider returns to the initial position. At this time, s, l, h, and α are at their maximum values. The simulation results show that to ensure that the setback slider can be released completely in the launch and return to the initial position in service processing, the range of the distance s from the vertical sidewall of the V-shaped groove to the inclined plane should be 0.2~0.3 mm, the range of the distance l between the vertical sidewalls should be 1~2 mm, the range of the depth h of the vertical sidewall should be 0.3~0.4 mm, and the range of the sharp angle α should be 15.5 •~1 6.5 • . Therefore, the values s = 0.25 mm, l = 1.5 mm, h = 0.35 mm, and α = 16 • are preliminarily selected.

Simulation of Service Processing
The setback load in service processing is set to 18,000 g in the ANSYS/LS-DYNA software. To observe the displacement of the setback slider after the load is removed, the simulation time is set to 0.3 ms. Figure 16 shows the stress distribution nephogram of the setback arming device at 0.1 ms. It can be seen from the figure that the angle of the V-shaped groove at both ends of the setback slider decreases, but it does not break away from the arming slider. The maximum stress of the setback arming device in the figure is 495.3 Pa, which is less than the allowable stress for a nickel material, so the material does not undergo plastic deformation. Figure 17 shows the time-displacement curve of the setback slider in service processing. The displacement of the setback slider increases rapidly after the movement begins, reaching a maximum value of 0.047 mm at 0.1 ms. When the load is removed, the displacement decreases with increasing time, returning to 0 from 0.24 ms and then remaining unchanged; this indicates that the setback slider returns to the initial position under the action of the microspring restoring force. mm, the range of the distance l between the vertical sidewalls should be 1~2 mm, th of the depth h of the vertical sidewall should be 0.3~0.4 mm, and the range of th angle α should be 15.5°~16.5°. Therefore, the values s = 0.25 mm, l = 1.5 mm, h = 0 and α = 16° are preliminarily selected.

Simulation of Service Processing
The setback load in service processing is set to 18,000 g in the ANSYS/LS-DYN ware. To observe the displacement of the setback slider after the load is removed, ulation time is set to 0.3 ms. Figure 16 shows the stress distribution nephogram setback arming device at 0.1 ms. It can be seen from the figure that the angle o shaped groove at both ends of the setback slider decreases, but it does not brea from the arming slider. The maximum stress of the setback arming device in the 495.3 Pa, which is less than the allowable stress for a nickel material, so the mater not undergo plastic deformation. Figure 17 shows the time-displacement curve of back slider in service processing. The displacement of the setback slider increases after the movement begins, reaching a maximum value of 0.047 mm at 0.1 ms. W load is removed, the displacement decreases with increasing time, returning to 0 fr ms and then remaining unchanged; this indicates that the setback slider return initial position under the action of the microspring restoring force.

Simulation of the Launch
The setback load at launch is set to 12,000 g, and the simulation time is set to The setback slider moves away from the arming slider under the action of the lo proaches the arming slider under the action of the microspring restoring force, a lides with the obstructing inclined plane of the arming slider. After the collision, back slider moves away from the arming slider again under the action of the inerti and this process is repeated until the inertial force energy is completely exhausted. the angle of the V-shaped groove at both ends of the setback slider becomes larger gets stuck on the arming slider. Figure 18 shows the stress distribution nephogram setback arming device at 10 ms. The maximum stress value in the figure is 157 Pa is less than the allowable stress of a nickel material. Figure 19 shows the time-d ment curve of the setback slider at launch. The figure shows that the displacemen setback slider increases slowly at the beginning of the movement process, the slider starts to move away from the arming slider after 2.75 ms, and the displa increases rapidly with time, reaching the maximum value of 0.963 mm at 4.4 ms tionally, the displacement of the setback slider produces an oscillation curve un actions of the inertial force and microspring restoring force. After 9.9 ms, the displa is stable at 0.452 mm and remains unchanged, indicating that the setback arming is completely released.

Simulation of the Launch
The setback load at launch is set to 12,000 g, and the simulation time is set to 10 ms. The setback slider moves away from the arming slider under the action of the load, approaches the arming slider under the action of the microspring restoring force, and collides with the obstructing inclined plane of the arming slider. After the collision, the setback slider moves away from the arming slider again under the action of the inertial force, and this process is repeated until the inertial force energy is completely exhausted. Finally, the angle of the V-shaped groove at both ends of the setback slider becomes larger, and it gets stuck on the arming slider. Figure 18 shows the stress distribution nephogram of the setback arming device at 10 ms. The maximum stress value in the figure is 157 Pa, which is less than the allowable stress of a nickel material. Figure 19 shows the time-displacement curve of the setback slider at launch. The figure shows that the displacement of the setback slider increases slowly at the beginning of the movement process, the setback slider starts to move away from the arming slider after 2.75 ms, and the displacement increases rapidly with time, reaching the maximum value of 0.963 mm at 4.4 ms. Additionally, the displacement of the setback slider produces an oscillation curve under the actions of the inertial force and microspring restoring force. After 9.9 ms, the displacement is stable at 0.452 mm and remains unchanged, indicating that the setback arming device is completely released.

Simulation of the Launch
The setback load at launch is set to 12,000 g, and the simulation time is set t The setback slider moves away from the arming slider under the action of the lo proaches the arming slider under the action of the microspring restoring force, lides with the obstructing inclined plane of the arming slider. After the collision, back slider moves away from the arming slider again under the action of the inerti and this process is repeated until the inertial force energy is completely exhausted. the angle of the V-shaped groove at both ends of the setback slider becomes large gets stuck on the arming slider. Figure 18 shows the stress distribution nephogra setback arming device at 10 ms. The maximum stress value in the figure is 157 Pa is less than the allowable stress of a nickel material. Figure 19 shows the time-d ment curve of the setback slider at launch. The figure shows that the displacemen setback slider increases slowly at the beginning of the movement process, the slider starts to move away from the arming slider after 2.75 ms, and the displa increases rapidly with time, reaching the maximum value of 0.963 mm at 4.4 m tionally, the displacement of the setback slider produces an oscillation curve un actions of the inertial force and microspring restoring force. After 9.9 ms, the displa is stable at 0.452 mm and remains unchanged, indicating that the setback arming is completely released.   The above simulation results show that the improved design of the se device solves the difficult problem of setback slider deformation, achieves t lay arming, and can reliably distinguish service processing from launch.

Processing the Improved Setback Slider
According to the improvement scheme of the setback slider, two V-sh with a vertical sidewall depth of 0.35 mm, a vertical sidewall distance of 0. sharp angle of 16° need to be processed on a setback slider with a thickne Due to the special shape of the V-shaped groove and the strict requireme sional accuracy and position accuracy, it is difficult to use wire EDM (WED is generally used instead. EDM has unique advantages for processing precis such as narrow slits and sharp grooves [33,34]. This method uses forming process workpieces, so it is necessary to prepare special electrodes for shaped grooves. To ensure the dimensional accuracy and positional accur shaped groove and improve the processing efficiency, the electrode shapes to be two triangles that correspond to the V-shaped groove. The distance b two electrodes should be equal to the distance between the vertical sidew shaped groove. In addition, to prevent the discharge phenomenon on the co face between the electrode and the workpiece, the height of the electrode propriately increased. The processing scheme of the V-shaped groove is sho 20. In the figure, the distance between the two electrodes is l′ = l = 1.5 mm, height is h′ = 0.9 mm, and the sharp angle is ' α = α = 16°. The above simulation results show that the improved design of the setback arming device solves the difficult problem of setback slider deformation, achieves the goal of delay arming, and can reliably distinguish service processing from launch.

Processing the Improved Setback Slider
According to the improvement scheme of the setback slider, two V-shaped grooves with a vertical sidewall depth of 0.35 mm, a vertical sidewall distance of 0.15 mm, and a sharp angle of 16 • need to be processed on a setback slider with a thickness of 0.4 mm. Due to the special shape of the V-shaped groove and the strict requirements of dimensional accuracy and position accuracy, it is difficult to use wire EDM (WEDM), and EDM is generally used instead. EDM has unique advantages for processing precision structures such as narrow slits and sharp grooves [33,34]. This method uses forming electrodes to process workpieces, so it is necessary to prepare special electrodes for processing Vshaped grooves. To ensure the dimensional accuracy and positional accuracy of the Vshaped groove and improve the processing efficiency, the electrode shapes are designed to be two triangles that correspond to the V-shaped groove. The distance between these two electrodes should be equal to the distance between the vertical sidewalls of the Vshaped groove. In addition, to prevent the discharge phenomenon on the connection surface between the electrode and the workpiece, the height of the electrode should be appropriately increased. The processing scheme of the V-shaped groove is shown in Figure 20. In the figure, the distance between the two electrodes is l = l = 1.5 mm, the electrode height is h = 0.9 mm, and the sharp angle is α = α = 16 • . Figure 19. Time-displacement curve of the setback slider at launch.
The above simulation results show that the improved design of the se device solves the difficult problem of setback slider deformation, achieves t lay arming, and can reliably distinguish service processing from launch.

Processing the Improved Setback Slider
According to the improvement scheme of the setback slider, two V-sh with a vertical sidewall depth of 0.35 mm, a vertical sidewall distance of 0. sharp angle of 16° need to be processed on a setback slider with a thickne Due to the special shape of the V-shaped groove and the strict requireme sional accuracy and position accuracy, it is difficult to use wire EDM (WED is generally used instead. EDM has unique advantages for processing precis such as narrow slits and sharp grooves [33,34]. This method uses forming process workpieces, so it is necessary to prepare special electrodes for p shaped grooves. To ensure the dimensional accuracy and positional accur shaped groove and improve the processing efficiency, the electrode shapes to be two triangles that correspond to the V-shaped groove. The distance b two electrodes should be equal to the distance between the vertical sidew shaped groove. In addition, to prevent the discharge phenomenon on the co face between the electrode and the workpiece, the height of the electrode s propriately increased. The processing scheme of the V-shaped groove is sho 20. In the figure, the distance between the two electrodes is l′ = l = 1.5 mm, height is h′ = 0.9 mm, and the sharp angle is ' α = α = 16°.

Electrode Processing
The processing accuracy of the electrode directly affects the processin the V-shaped groove. To reduce the influence of electrode loss on the process

Electrode Processing
The processing accuracy of the electrode directly affects the processing accuracy of the V-shaped groove. To reduce the influence of electrode loss on the processing accuracy, the V-shaped groove forming electrode was fabricated through the WEDM process. A CA30 immersion precision CNC low-speed wire cutting machine tool was used for processing. The electrode material was a copper-tungsten alloy. The copper-tungsten alloy electrode has the characteristics of high-temperature resistance, arc ablation resistance, high electrical conductivity, and high thermal conductivity. The electrode loss during processing was extremely low, which was conducive to improving the quality of the processed surface. The electrode wire was a brass wire with a diameter of 0.2 mm, the dielectric was deionized water, and the resistivity was 1 × 10 5 Ω·cm. During electrode processing, the machining tool uses a lower feed speed and keeps the speed constant for multiple cuttings. The main parameters are shown in Table 5. Figure 21 shows an actual picture of the processed electrode; the picture shows that the electrode has a good processing quality and high shape accuracy. The processed electrodes were measured via stereoscopic microscopy. The measured results are shown in Table 6. The distance l , height h , and angle α between the left and right electrodes are similar to the design values, meeting the processing requirements. scopic microscopy. The measured results are shown in Table 6. The distance l′, h and angle α between the left and right electrodes are similar to the design value ing the processing requirements.

V-Shaped Groove Formation
The V-shaped groove was fabricated through the EDM process, and the ma tool used was an EA12AM CNC ultraprecision mirror spark machine prod Mitsubishi Electric Automation Co., Ltd. (Tokyo, Japan) Figure 22 shows the ac ture of the V-shaped groove after processing; the picture shows that the vertica shape accuracy of the V-shaped groove sidewall are increased, and there are fewe ments on the processed surface. The formed V-shaped grooves were measured v oscopic microscopy, and the measured results are shown in Table 7. The distance h, and angle α between the vertical sidewalls of the left and right V-shaped gro close to the design values, meeting the processing requirements.

V-Shaped Groove Formation
The V-shaped groove was fabricated through the EDM process, and the machining tool used was an EA12AM CNC ultraprecision mirror spark machine produced by Mitsubishi Electric Automation Co., Ltd. (Tokyo, Japan) Figure 22 shows the actual picture of the V-shaped groove after processing; the picture shows that the verticality and shape accuracy of the V-shaped groove sidewall are increased, and there are fewer attachments on the processed surface. The formed V-shaped grooves were measured via stereoscopic microscopy, and the measured results are shown in Table 7. The distance l, depth h, and angle α between the vertical sidewalls of the left and right V-shaped grooves are close to the design values, meeting the processing requirements.

Improvement Scheme for the Microsprings
In the existing structure, the microspring and the frame are connected as a which produces some inconvenience in the processing stage. Additionally, the fra not be reused in the test process, increasing the test cost. Considering that the installed perpendicular to the axis of the projectile, there is ample space in the d of the frame plane. Therefore, the microspring and the frame can be separated in sign, and a T-shaped mass block is left at the bottom of the microspring. Moreo bottom width of the frame is increased, a T-shaped groove corresponding to th the mass block is maintained, and the two components are connected through c fitting. Figure 23 shows an improved design scheme in which the frame can be pr directly, greatly simplifying the fabrication process.

Processing the Improved Setback Arming Device
The processing accuracy of the microspring and V-shaped groove is the ke for ensuring the safety and launch reliability of the setback arming device. In a there is a clear positional relationship between the microspring, the V-shaped groo the setback slider, so the processing error and positioning error must be strictly co during processing. Additionally, it is very important to reasonably arrange the the processing procedure. The microspring was fabricated through a combination tiple cutting and destress annealing steps. Although the flexural deformation of crospring was significantly improved, deformation still existed. If the "microsprin the V-shaped groove" processing procedure is followed, the positional accuracy o

Improvement Scheme for the Microsprings
In the existing structure, the microspring and the frame are connected as a whole, which produces some inconvenience in the processing stage. Additionally, the frame cannot be reused in the test process, increasing the test cost. Considering that the frame is installed perpendicular to the axis of the projectile, there is ample space in the direction of the frame plane. Therefore, the microspring and the frame can be separated in the design, and a T-shaped mass block is left at the bottom of the microspring. Moreover, the bottom width of the frame is increased, a T-shaped groove corresponding to the size of the mass block is maintained, and the two components are connected through clearance fitting. Figure 23 shows an improved design scheme in which the frame can be processed directly, greatly simplifying the fabrication process.
Micromachines 2022, 13,292 frame" procedure was adopted during processing, and the processed sample is s Figure 24.

Processing the Improved Setback Arming Device
The processing accuracy of the microspring and V-shaped groove is the key factor for ensuring the safety and launch reliability of the setback arming device. In addition, there is a clear positional relationship between the microspring, the V-shaped groove, and the setback slider, so the processing error and positioning error must be strictly controlled during processing. Additionally, it is very important to reasonably arrange the steps in the processing procedure. The microspring was fabricated through a combination of multiple cutting and destress annealing steps. Although the flexural deformation of the microspring was significantly improved, deformation still existed. If the "microspring before the V-shaped groove" processing procedure is followed, the positional accuracy of the V-shaped groove cannot be guaranteed. Therefore, the "V-shaped groove-microspringframe" procedure was adopted during processing, and the processed sample is shown in Figure 24.
frame" procedure was adopted during processing, and the processed sample is s Figure 24.   Figure 25 shows the improved principle prototype. Compared with the structure, the bottom width of the frame is increased by 1.1 mm. 10 improved p prototypes are taken as one group, and a total of 100 groups of mechanical imp and centrifugal overload tests are carried out under the same test conditions as Section 4 of this paper.   Figure 25 shows the improved principle prototype. Compared with the original structure, the bottom width of the frame is increased by 1.1 mm. 10 improved principle prototypes are taken as one group, and a total of 100 groups of mechanical impact tests and centrifugal overload tests are carried out under the same test conditions as those in Section 4 of this paper.   Figure 25 shows the improved principle prototype. Compared wit structure, the bottom width of the frame is increased by 1.1 mm. 10 impro prototypes are taken as one group, and a total of 100 groups of mechanica and centrifugal overload tests are carried out under the same test conditio Section 4 of this paper. The results of the mechanical impact test of group 1 are shown in Tab peak acceleration decreases from 23,000 g to 22,000 g, the angle of the V-sha

Mechanical Impact Test Results of Group 1
The results of the mechanical impact test of group 1 are shown in Table 8. When the peak acceleration decreases from 23,000 g to 22,000 g, the angle of the V-shaped groove at both ends of the setback slider increases after it completely moves away from the arming slider. The setback slider then gets stuck on the arming slider and cannot return to the initial position. When the peak acceleration decreases to 21,000 g, the angle of the V-shaped groove at both ends of the setback slider decreases, and some inclined planes break away from the arming slider and cannot return to the initial position. When the peak acceleration decreases from 20,000 g to 14,000 g, the setback slider remains in the initial position, and the angles of the V-shaped grooves at both ends do not change. Figure 26 includes the scanning electron microscopy (SEM) images of the setback arming device when the peak acceleration is 21,000 g and 20,000 g.

Centrifugal Overload Test Results of Group 1
The results of the centrifugal overload test of group 1 are shown in Table 9. When the centrifugal acceleration decreases from 17,000 g to 12,000 g, the setback slider can move to the fully released position. When the centrifugal acceleration decreases to 11,000 g, the angles of the V-shaped grooves at both ends of the setback slider decrease, and some inclined planes break away from the arming slider, but the safety is not completely released. When the centrifugal acceleration decreases from 10,000 g to 8000 g, the setback slider always remains in the initial position and cannot be released. Figure 27 shows the SEM images of the setback arming device when the centrifugal acceleration is 12,000 g and 11,000 g. Table 9. Centrifugal overload test results of group 1.

Sample Number Centrifugal Acceleration/g Setback Slider Position
#1 17,000 The angle of the V-shaped groove increases and it gets stuck on the arming slider. #2 16,000 The angle of the V-shaped groove increases and it gets stuck on the arming slider. #3 15,000 The angle of the V-shaped groove increases and it gets stuck on the arming slider. #4 14,000 The angle of the V-shaped groove increases and it gets stuck on the arming slider. #5 13,000 The angle of the V-shaped groove increases and it gets stuck on the arming slider. #6 12,000 The angle of the V-shaped groove increases and it gets stuck on the arming slider.

#7 11,000
The angle of the V-shaped groove decreases and some inclined planes break away from the arming slider.  Figure 28 shows the maximum overload that the improved setback arming devic can be withstood in the mechanical impact test, and Figure 29 shows the minimum over load that the improved setback arming device for safety release in the centrifugal overload test. Table 10 shows the statistical results. It can be seen from the table that the test result of the two groups fall within the tolerance range, which satisfies the "3σ criterion". Taking into account the influence of machining errors and test conditions, the statistical result show that the improved setback arming device can guarantee the safety of service pro cessing and the reliability of launch. The maximum overload that can be withstood in service processing is 20,000 g, and the minimum overload for safety release during launch is 12,000 g.   Figure 28 shows the maximum overload that the improved setback arming device can be withstood in the mechanical impact test, and Figure 29 shows the minimum overload that the improved setback arming device for safety release in the centrifugal overload test. Table 10 shows the statistical results. It can be seen from the table that the test results of the two groups fall within the tolerance range, which satisfies the "3σ criterion". Taking into account the influence of machining errors and test conditions, the statistical results show that the improved setback arming device can guarantee the safety of service processing and the reliability of launch. The maximum overload that can be withstood in service processing is 20,000 g, and the minimum overload for safety release during launch is 12,000 g.  Figure 28 shows the maximum overload that the improved setback arming device can be withstood in the mechanical impact test, and Figure 29 shows the minimum overload that the improved setback arming device for safety release in the centrifugal overload test. Table 10 shows the statistical results. It can be seen from the table that the test results of the two groups fall within the tolerance range, which satisfies the "3σ criterion". Taking into account the influence of machining errors and test conditions, the statistical results show that the improved setback arming device can guarantee the safety of service processing and the reliability of launch. The maximum overload that can be withstood in service processing is 20,000 g, and the minimum overload for safety release during launch is 12,000 g.

Conclusions
In this paper, the EDM process is successfully applied to the fabrication of a verticalframe-based setback arming device, and a mechanical impact test and centrifugal overload test are carried out on the principle prototype. The structure of the principle prototype is then improved according to the problems indicated by the test, and the conclusions are as follows: 1.
The setback arming device fabricated by the EDM process can greatly reduce the machining cost while improving the machining precision and machining speed, which can compensate for the defects of the current manufacturing technology. The problems of flexural deformation and secondary deformation in the microspring processing can be effectively solved by designing the auxiliary support beam, using multiple cuts, destress annealing, and optimizing the processing parameters. When the width of the auxiliary support beam is 0.04 mm and the position is set at the center of the microspring end, the influence on the microspring during processing is minimized. At this point, the numbers of gears that are selected for the machining tool are pulse interval SB = 18, peak current IP = 3, flushing pressure LQ = 5, reference voltage VG = 75, electrode wire tension WT = 4, and wire speed WS = 9.

2.
The difficult deformation problem can be effectively solved by establishing V-shaped grooves at both ends of the setback slider. The distance between the vertical sidewall of the V-shaped groove and the inclined plane is 0.25 mm, and the distance between the vertical sidewalls is 1.5 mm. The vertical sidewall has a depth of 0.35 mm and a sharp angle of 16 • . 3.
The setback arming device connected by a clearance fit between the microspring and the frame can not only be used to simplify the fabrication process and minimize testing costs but also guarantees the service processing safety and launch reliability. The maximum overload that can be withstood in service processing is 20,000 g, and the minimum overload for safety release during launch is 12,000 g.
Author Contributions: Y.Q. designed and performed the tests; Y.Q. and X.Z. helped to analyze the data and write the paper; Y.S. guided and revised the paper; Y.H. conceived the framework of this research. All authors have read and agreed to the published version of the manuscript.
Funding: This research was funded by Analysis and Testing Technology of Thermal Characteristics of Thin Film Materials, grant number 2020YFB2008702.