A device and method for detecting three-dimensional morphology of PCB surface based on MEMS coaxial scanning
By using a MEMS coaxial scanning device and method, combined with four-channel synchronous acquisition and coherent accumulation, the problem of balancing scanning range and detection spot in close-range PCB inspection was solved, the influence of light source fluctuation and internal parasitic reflection was reduced, and stable phase ranging and detection effect were achieved.
Patent Information
- Application Number
- CN202611315586.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-27
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies struggle to balance scanning range and detection spot size in close-range PCB inspection, while also exhibiting issues such as internal parasitic reflections and instability in detection across regions with varying reflectivity.
A PCB surface three-dimensional topography detection device based on MEMS coaxial scanning is adopted. By using MEMS micromirrors and scanning focusing lens group, combined with four-channel synchronous acquisition and coherent accumulation, the influence of light source fluctuation and internal parasitic reflection is reduced. Stable phase ranging is achieved through IQ demodulation and position-related stray light correction.
It improves the balance between scanning range and detection spot size, reduces the impact of light source fluctuations and internal parasitic reflections on phase ranging, and improves detection stability and ranging repeatability in regions with different reflectivities.
Smart Images

Figure CN123111131A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of optical inspection and phase-time-of-flight ranging technology for printed circuit boards, specifically to a device and method for close-range three-dimensional topography inspection of PCB surfaces based on MEMS micromirror coaxial scanning. Background Technology
[0002] The surface of a printed circuit board (PCB) includes areas such as solder mask, copper foil, gold-plated pads, solder, silkscreen printing, and component packages. The optical reflectivity of these different areas varies considerably, and defects such as warpage, bulges, depressions, abnormal solder joint height, and foreign matter typically exhibit small lateral dimension or height variations. Two-dimensional imaging can identify some color and texture defects, but it is difficult to directly obtain stable surface height information.
[0003] Phase-time-of-flight ranging obtains distance by periodically modulating the intensity of the emitted light and measuring the modulation phase difference between the emitted light and the target's return light. Unlike long-range lidar, PCB surface detection has a shorter working distance, requires a smaller spot size, and involves denser scanning positions. Under these short-range conditions, the allowable deflection angle of MEMS micromirrors may limit the PCB scanning range; simply increasing the focal length of subsequent mirrors can increase the scanning displacement, but it easily increases the system size and the focused spot size.
[0004] Furthermore, the surfaces of the beam splitter, the front positive power lens, the rear negative power lens, and the converging lens in front of the photodetector will generate scattering and parasitic reflections. The second converging lens may collect the aforementioned parasitic light, and even if the PCB surface under test is not present, the second photodetector may still output a modulated signal with the same frequency as the laser. This parasitic light cannot be automatically removed by ordinary ambient light filtering or IQ demodulation, and will superimpose with the effective return light from the PCB in the complex phasor domain, causing a short-range phase deviation.
[0005] Under ideal linear conditions, changes in PCB reflectivity primarily affect the amplitude of the IQ signal without directly affecting the phase. However, signals in low-reflectivity regions are susceptible to noise and DC bias, while high-reflectivity regions may cause photodetectors, transimpedance amplifiers, or data acquisition units to enter nonlinear or clipping states, resulting in poor distance repeatability. Therefore, a close-range PCB surface inspection solution is needed that balances scanning range, small spot size, internal stray light correction, and detection stability in regions with different reflectivities under fixed laser drive conditions. Summary of the Invention
[0006] The purpose of this invention is to provide a PCB surface three-dimensional topography detection device and method based on MEMS coaxial scanning, so as to take into account the scanning range and detection spot in close-range detection, and reduce the influence of light source fluctuation, internal parasitic reflection and backlight intensity difference on phase ranging.
[0007] To achieve the above objectives, the present invention adopts the following technical solution: a PCB surface three-dimensional topography detection device based on MEMS coaxial scanning, comprising a modulated laser source (100), a collimating lens (110), a beam splitter (120), a first converging lens (130), a first photodetector (140), a MEMS micromirror (200), a MEMS driving unit (210), a MEMS position feedback unit (220), a positive optical power lens (310), a negative optical power lens (300), a second converging lens (500), an aperture stop (510), a second photodetector (520), an FPGA modulation and control unit (600), a synchronous acquisition unit (610), and a processing unit (620).
[0008] The FPGA modulation and control unit (600) provides a periodic modulation signal to the modulation laser source (100) so that the modulation laser source (100) outputs a modulation laser whose light intensity changes periodically according to a predetermined modulation frequency. The beam splitting element (120) splits the modulation laser collimated by the collimating lens (110) into a reference light and a detection light. The reference light is focused onto the photosensitive surface of the first photodetector (140) by the first converging lens (130).
[0009] The MEMS driving unit (210) drives the MEMS micromirror (200) to perform two-dimensional deflection; the MEMS micromirror (200), positive optical power lens (310) and negative optical power lens (300) are arranged sequentially along the direction of detection light propagation. The positive optical power lens (310) and the negative optical power lens (300) constitute a net positive optical power scanning focusing lens group. After being deflected by the MEMS micromirror (200), the detection light passes through the scanning focusing lens group and forms a scanning spot on the PCB surface (400) to be tested; the PCB backlight passes through the negative optical power lens (300), positive optical power lens (310) and MEMS micromirror (200), and is guided to the detection receiving branch by the beam splitter (120), passes through the aperture stop (510), and is focused by the second converging lens (500) onto the photosensitive surface of the second photodetector (520);
[0010] The MEMS position feedback unit (220) outputs a first scan axis position feedback signal and a second scan axis position feedback signal that characterize the actual position of the first scan axis and the actual position of the second scan axis of the MEMS micromirror (200); the synchronous acquisition unit (610) acquires the first photodetector reference signal, the second photodetector detection signal, the first scan axis position feedback signal and the second scan axis position feedback signal based on a common sampling time reference.
[0011] The processing unit (620) performs IQ demodulation on the first photodetector reference signal and the second photodetector detection signal within the same sampling time window, determines the distance to the PCB surface to be tested based on the complex phasor relationship and phase difference of the two signals, and determines the two-dimensional scanning position corresponding to the distance based on the first scanning axis position feedback signal and the second scanning axis position feedback signal.
[0012] Furthermore, the beam splitter (120) is a wedge-shaped beam splitter or a wedge-shaped beam splitter prism with two non-parallel optical surfaces, so that the parasitic reflected light generated by the non-working surface deviates from the receiving optical path of the second photodetector; the first converging lens (130) and the second converging lens (500) are both positive power lenses, and the photosensitive surfaces of the first photodetector (140) and the second photodetector (520) are located near the image plane or beam waist of the corresponding converging lens, respectively.
[0013] The detection receiving branch has an aperture stop (510) between the beam splitter (120) and the second converging lens (500). The aperture of the aperture stop (510) is configured to allow effective backlighting from the PCB to pass through and to limit stray light deviating from the detection receiving direction from entering the second photodetector (520).
[0014] Furthermore, let the focal length of the positive power lens (310) be f. P The focal length of the negative power lens (300) is f. N The axial distance between the principal planes of the two lenses is d, where f P >0, f N <0; the paraxial focal length F of the scanning focusing lens group satisfies And F > 0.
[0015] Furthermore, the focal length of the positive power lens (310), the focal length of the negative power lens (300), the distance between the two lenses, the effective aperture of the lenses, and the position of the PCB test surface (400) are configured such that: the PCB detection range is increased relative to the lensless free propagation detection range under the same MEMS optical deflection angle and the same axial distance, and the detection spot size within the PCB detection range is less than a preset defect size threshold; the PCB test surface (400) is located within the multi-field integrated optimal imaging surface of the scanning focusing lens group or its depth of field.
[0016] Furthermore, it also includes an FPGA modulation and control unit (600) and a MEMS driving unit (210). The FPGA modulation and control unit (600) provides a periodic modulation signal to the laser driving circuit of the modulation laser source (100) and provides a two-dimensional scanning driving signal to the MEMS driving unit (210). The synchronous acquisition unit (610) includes at least a first photodetector sampling channel, a second photodetector sampling channel, a first scanning axis position feedback channel, and a second scanning axis position feedback channel. The four sampling channels adopt a common sampling time reference. When there is a fixed time difference between the sampling channels, the processing unit (620) compensates according to the pre-calibrated channel time difference.
[0017] Furthermore, the processing unit (620) determines the MEMS scanning position p based on the first scanning axis position feedback signal and the second scanning axis position feedback signal within the same sampling time window, and stores it in the normalized stray light complex phasor B(p) corresponding to the MEMS scanning position p obtained when the modulated laser source (100) is turned on and the PCB test surface (400) is not set; the processing unit (620) performs complex normalization on the second photodetector complex phasor Z2(p) obtained in actual detection and the first photodetector complex phasor Z1(p), and according to Obtain the effective luminous complex phasor Z of the PCB T (p).
[0018] Furthermore, the processing unit (620) coherently accumulates the PCB effective return complex phasors corresponding to the same MEMS scanning position p in multiple repeated scans, and determines the return phase by the in-phase component and quadrature component of the accumulated complex phasors; the modulated laser source (100) operates with a predetermined fixed modulation amplitude, and the processing unit (620) determines the number of coherent accumulations or judges whether the detection result is valid based on one or more of the amplitude of the PCB effective return complex phasors, the clipping state of the original sampling waveform, and the phase concentration of repeated measurements.
[0019] Furthermore, the detection receiving branch is also equipped with a fixed ratio second beam splitter. One output end of the second beam splitter is connected in sequence to a second converging lens (500) and a second photodetector (520) to form a first detection branch, and the other output end is connected in sequence to an additional converging lens and an additional photodetector to form a second detection branch. The first detection branch and the second detection branch have different receiving sensitivities and synchronously collect the return light from the same PCB. The processing unit (620) calibrates the complex gain and inherent phase difference of the two detection branches respectively, and selects the unsaturated detection branch that meets the signal-to-noise ratio threshold to output the ranging result.
[0020] A method for detecting the three-dimensional topography of a PCB surface based on MEMS coaxial scanning, implemented using the aforementioned MEMS coaxial scanning-based PCB surface three-dimensional topography detection device, includes:
[0021] The modulated laser source (100) is driven to output modulated laser, and a reference light and a detection light are formed by a beam splitter (120); the reference light is received by a first photodetector (140) through a first converging lens (130);
[0022] The MEMS micromirror (200) is driven to deflect according to a predetermined two-dimensional scanning trajectory, so that the detection light passes through the positive optical power lens (310) and the negative optical power lens (300) in sequence along the propagation direction, scans the PCB surface (400) to be tested and forms a detection spot; the PCB return light is made to propagate in the opposite direction along the detection optical path, pass through the negative optical power lens (300), the positive optical power lens (310) and the MEMS micromirror (200) in sequence, and is guided to the detection receiving branch by the beam splitter (120), and then focused by the second converging lens (500) to the photosensitive surface of the second photodetector (520);
[0023] The first photodetector reference signal, the second photodetector detection signal, the first scanning axis position feedback signal, and the second scanning axis position feedback signal are acquired based on a common sampling time reference. The complex phasor Z1(p) of the first photodetector reference light and the complex phasor Z2(p) of the second photodetector return light are obtained within the same sampling time window, and the MEMS scanning position p is determined based on the two position feedback signals.
[0024] Based on the scan position p, the stray light complex phasor B(p) pre-established when the PCB test surface (400) is not set, and the effective return light complex phasor of the PCB is calculated. Coherent accumulation is performed on multiple PCB effective backlight complex phasors corresponding to the same scanning position p; the PCB surface distance is calculated based on the corrected phase difference of the accumulated complex phasors, and the distance is correlated with the scanning position p to reconstruct the three-dimensional morphology of the PCB surface.
[0025] Furthermore, it also includes: under the condition that the modulated laser source (100) is turned on and the PCB test surface (400) is not set, the MEMS micromirror (200) traverses the predetermined scanning range; simultaneously acquiring the first photodetector calibration signal, the second photodetector calibration signal, and the two position feedback signals, to obtain the first photodetector calibration complex phasor Z corresponding to each scanning position p. 10 (p) and the second photodetector calibration complex phasor Z 20 (p), according to Establish a location-dependent two-dimensional complex number calibration table;
[0026] Data falling within the same two-dimensional scanning angle interval or the same position number p in multiple repeated scans are grouped together. The in-phase component and quadrature component are accumulated for the PCB effective backlight complex phasors of the same group. The validity of the measurement results is judged based on the amplitude of the PCB effective backlight complex phasors, the clipping state of the original sampled waveform, and the phase concentration. The PCB reference plane is fitted based on the three-dimensional coordinates of multiple effective scanning positions, and the height of each scanning position relative to the PCB reference plane is calculated. Based on at least one of the height, local height gradient, and backlight amplitude, PCB warping, bulging, dents, foreign objects, abnormal solder joint height, solder pad lifting, or component lifting are identified.
[0027] Compared with the prior art, the present invention has the following beneficial effects:
[0028] First, the scanning focusing lens group can jointly adjust the scanning angle transmission relationship and the beam convergence state, which is beneficial to balance the scanning range and the detection spot size.
[0029] Second, the transmitted light and the PCB return light share the same MEMS micromirror and scanning focusing lens group, which helps to reduce the mismatch between the transmitting and receiving fields of view.
[0030] Third, reference normalization and position-dependent stray complex number subtraction help reduce phase deviation caused by common fluctuations of the light source and internal parasitic backlight;
[0031] Fourth, the four-channel synchronous acquisition, coherent accumulation, and signal validity judgment are conducive to achieving the correspondence between distance and actual scanning position, and improving the repeatability of ranging in different backlight intensity areas. Attached Figure Description
[0032] Figure 1 This is a schematic diagram of the overall coaxial optical path structure of the detection device of the present invention.
[0033] Figure 2 This is a schematic diagram illustrating the scanning angle transfer and focusing principle of a MEMS micromirror, a front positive optical power lens, and a rear negative optical power lens.
[0034] Figure 3 This is a schematic diagram of the data connections between the FPGA, laser driver, MEMS driver, MEMS two-dimensional position feedback, first photodetector, second photodetector, four-channel synchronous acquisition unit and processing unit.
[0035] Figure 4 This is a schematic diagram of the complex phasor calibration and actual testing process for stray light without PCB.
[0036] Figure 5 This is a schematic diagram illustrating the principle of digital coherent accumulation at the same scanning position.
[0037] Figure 6 This is a schematic diagram illustrating the process for determining the validity of signals reflected by light with different reflectivities.
[0038] Figure 7 This is a simulation result image of complex background correction related to two-dimensional scanning position, in which... Figure 7 (a) is the true Z-surface. Figure 7 (b) shows the Z-reconstruction result without background removal. Figure 7 (c) is the Z-reconstruction result after removing the location-dependent complex background. Detailed Implementation
[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the following description, in conjunction with the accompanying drawings and embodiments, provides further details. These embodiments are used to explain the invention and do not limit the scope of protection of the invention.
[0040] I. Device and Optical Path
[0041] like Figure 1 As shown, the PCB surface three-dimensional topography detection device based on MEMS coaxial scanning includes a modulated laser source 100, a collimating lens 110, a beam splitter 120, a first converging lens 130, a first photodetector 140, a MEMS micromirror 200, a MEMS driving unit 210, a MEMS position feedback unit 220, a positive power lens 310, a negative power lens 300, a PCB surface to be measured 400, a second converging lens 500, an aperture stop 510, a second photodetector 520, an FPGA modulation and control unit 600, a synchronous acquisition unit 610, and a processing unit 620.
[0042] The FPGA modulation and control unit 600 drives the modulation laser source 100 to output periodically modulated light, and the collimating lens 110 forms a collimated or nearly collimated beam. The beam splitter 120 splits the modulated light into a reference beam and a detection beam; the beam splitter 120 can be a wedge-shaped beam splitter or a wedge-shaped beam splitter prism with two non-parallel optical surfaces, so that parasitic reflections from the non-working surface deviate from the receiving optical path of the second photodetector 520. The reference light is converged to the photosensitive surface of the first photodetector 140 by a first converging lens 130 with positive optical power, and the photosensitive surface is located near the corresponding image plane or the beam waist.
[0043] The detection light is deflected by the MEMS micromirror 200 and then passes sequentially through the positive optical power lens 310 and the negative optical power lens 300 to form a scanning spot on the PCB surface 400 to be tested. The positive optical power lens 310 pre-converges the scanning beam, and the negative optical power lens 300 and the positive optical power lens 310 together adjust the principal ray angle and the image-side convergence state.
[0044] The PCB backlight propagates in the reverse direction along the detection optical path, passing sequentially through the negative optical power lens 300, the positive optical power lens 310, and the MEMS micromirror 200, before entering the detection receiving branch via the beam splitter 120. An aperture stop 510 is positioned between the beam splitter 120 and the second converging lens 500 to allow effective PCB backlight to pass through and to limit stray light deviating from the receiving direction. The second converging lens 500, with positive optical power, converges the backlight to the photosensitive surface of the second photodetector 520, which is located near the corresponding image plane or the beam waist.
[0045] II. Scanning Focusing Lens Group Parameter Configuration
[0046] like Figure 2 As shown, a positive optical power lens 310 and a negative optical power lens 300 are sequentially arranged after the MEMS micromirror 200 along the direction of light propagation. If the mechanical deflection angle of the MEMS is α, then the optical deflection angle θ of the reflected light is approximately 2α under paraxial conditions. The two lenses are spaced apart and form a net positive optical power scanning focusing lens group.
[0047] Let the focal length of the positive optical power lens 310 be f. P The focal length of a negative power lens of 300 is f. N The distance between the principal planes of the two lenses is d, where f P >0, f N If <0, then the paraxial combined focal length is:
[0048] (1)
[0049] Select f P f N The combined focal length F is greater than zero. In equation (1), d is the distance between the principal planes of the two lenses. For a lens with actual thickness, d is not equal to the air gap between the adjacent surfaces of the two lenses.
[0050] The scanning range should be determined by the complete ABCD ray transfer matrix from the MEMS micromirror 200 to the PCB surface under test 400, or by non-paraxial ray tracing. The transfer coefficient of the incident principal ray slope to the lateral displacement of the PCB surface is denoted as B. When representing the slope of the principal ray emitted from the MEMS, the height y of the principal ray on the PCB surface approximately satisfies The full scan width of a single axis can be expressed as B is not equivalent to the focal length F of the lens group combination.
[0051] The PCB surface under test, 400, is positioned within the optimal imaging plane of the multi-field view or its depth of field. The focal length, spacing, effective aperture, and PCB axial position of the lens group are jointly determined based on the full field-of-view scanning width, edge light transmission, and RMS spot size. This ensures that the detection range under the same MEMS optical deflection angle and axial distance is greater than the free propagation range without lenses, and that the detection spot is smaller than a preset defect size threshold. The PCB does not need to be strictly located on the paraxial focal plane defined by F.
[0052] III. Electrical Control, Synchronous Data Acquisition and...
[0053] like Figure 3 As shown, the FPGA modulation and control unit 600 provides a periodic modulation signal to the laser driving circuit of the modulation laser source 100, and drives the MEMS micromirror 200 to perform two-dimensional scanning via the MEMS driving unit 210. The MEMS position feedback unit 220 outputs a first scanning axis position feedback signal and a second scanning axis position feedback signal.
[0054] The synchronous acquisition unit 610 includes at least a first photodetector sampling channel, a second photodetector sampling channel, and two scanning axis position feedback channels. All four channels use a common sampling time reference. When a fixed channel time difference exists, the processing unit 620 performs interpolation or delay compensation based on a pre-calibrated value. Communication between the FPGA modulation and control unit 600 and the computer can transmit scanning parameters or scanning sequence numbers, but these are not used as the precise sampling time reference for the four signals.
[0055] The synchronous acquisition unit 610 samples the receiving channels corresponding to the first photodetector 140 and the second photodetector 520 respectively, obtaining two discrete sampling sequences x1[n] and x2[n]. Where x k [n] represents the sampled value obtained by the k-th receiving channel at the n-th sampling time. The sampled value is a digital quantity obtained after amplification, signal conditioning and analog-to-digital conversion of the output signal of the corresponding photodetector.
[0056] Processing unit 620 performs IQ demodulation on the two sampled sequences respectively. Specifically, the sampled sequences are respectively coupled with the sampling domain demodulation frequency f. d The corresponding in-phase quadrature reference signals are multiplied and summed within the demodulated data block to obtain the in-phase component I. k and orthogonal components Q k :
[0057] (2)
[0058] (3)
[0059] The complex demodulation quantity is composed of in-phase and quadrature components. ,Right now:
[0060] (4)
[0061] Where k is 1 or 2, corresponding to the first photodetector 140 and the second photodetector 520 respectively; n is the sampling point number; f s f is the sampling rate; N is the number of sampling points in a demodulated data block; d To receive the frequency corresponding to the modulated component in the sampling sequence, during direct sampling, f d Equal to the laser modulation frequency; when performing frequency domain transformation before sampling, f d It equals the frequency after frequency domain transformation. The processing unit 620 obtains Z1, Z2, and the actual scan angle determined by position feedback within the same sampling time window.
[0062] IV. Position-Related Scattered Beam Determination
[0063] like Figure 4 As shown, when the modulated laser source 100 is turned on and the PCB test surface 400 is not set, the MEMS micromirror 200 traverses the predetermined scanning range. The synchronous acquisition unit 610 acquires two calibration photoelectric signals and two position feedback signals, and the processing unit 620 obtains the first photodetector calibration complex phasor Z corresponding to each scanning position p. 10 (p) and the second photodetector calibration complex phasor Z 20 (p), and establish the normalized stray light complex phasor:
[0064] (5)
[0065] B(p) is the position-dependent stray light complex correction value corresponding to the scan position p, which includes both the amplitude and phase of the stray light.
[0066] In actual testing, the PCB test surface 400 is set within the scanning range of the MEMS micromirror 200 to obtain the complex phasor Z1(p) of the first photodetector and the complex phasor Z2(p) of the second photodetector. The effective backlight complex phasor Z of the PCB is then calculated in the complex domain. T (p):
[0067] (6)
[0068] Reference normalization is used to reduce the influence of laser modulation amplitude and common phase changes; when the operating temperature or modulation frequency changes, stray beam calibration can be performed again and the position-dependent stray complex correction B(p) can be updated.
[0069] V. Coherent Accumulation and Phase Ranging
[0070] like Figure 5As shown, the processing unit 620, based on two-way position feedback, groups data falling within the same two-dimensional scanning angle interval or the same position number p in repeated scans into the same group. After reference normalization and position-related stray light correction, the effective PCB return complex phasor Z corresponding to the scanning position p is obtained from each measurement. T,m (p), where the subscript T indicates the effective backlight of the PCB target and the subscript m indicates the repeat measurement sequence number.
[0071] Processing unit 620 performs coherent accumulation on the effective return complex phasor at scan position p:
[0072] (7)
[0073] Among them, M p w represents the number of coherent accumulations corresponding to scan position p. m The weight for the m-th measurement; the accumulated backlight phase is:
[0074] (8)
[0075] Where Re represents the real part of the complex number, and Im represents the imaginary part of the complex number.
[0076] The processing unit 620 determines the phase branch according to the preset working distance range, and the PCB surface distance measurement value d(p) corresponding to the scanning position p is calculated according to the following formula:
[0077] (9)
[0078] Where c is the speed of light in vacuum; n is the refractive index of the propagation medium; f m The modulation frequency is used to modulate the laser source 100.
[0079] VI. Signal Validity and Dual Sensitivity Reception
[0080] like Figure 6 As shown, the processing unit 620 first checks the original sampled waveform of the second photodetector 520. When the sampled value reaches the upper or lower limit of the acquisition channel range, the waveform shows a continuous flat top, or the original signal amplitude exceeds the pre-calibrated linear operating range of the analog receiving link, it is determined that the corresponding scanning position has clipped or exceeded the linear operating range, and the scanning position is marked as saturated or invalid.
[0081] For signals that have not undergone clipping and are within the linear operating range, processing unit 620 performs IQ demodulation, reference normalization, and position-dependent stray complex correction, and coherently accumulates the corrected effective return complex phasors at the same scan position. Processing unit 620 calculates the amplitude A of the corrected complex phasors. T :
[0082] (10)
[0083] The modulated laser source 100 operates with a predetermined fixed modulation amplitude, and the processing unit 620 processes the amplitude A. T The clipping state of the original sampled waveform and the phase concentration R of repeated measurements determine the number of accumulations or the validity of the result. The phase concentration at scan position p can be expressed as:
[0084] (11)
[0085] Where, φ m (p) represents the backlight phase of the m-th repeated measurement at scan position p, M p The number of repeated measurements corresponds to the scanning position p. The phase concentration R(p) ranges from 0 to 1; the closer R(p) is to 1, the more concentrated the phase of the repeated measurements, and the more stable the measurement results.
[0086] The modulated laser source 100 operates with a predetermined fixed modulation amplitude. The processing unit 620 determines the amplitude A of the effective return complex phasor of the PCB. T (p) The validity of the coherent accumulation count or detection result is determined by at least one of the following: the clipping state of the original sampled waveform and the phase concentration R(p) of repeated measurements. When the original sampled waveform is clipped or exceeds the pre-calibrated linear working range, the corresponding scan position is marked as saturated or invalid; when the return signal is weak and no clipping has occurred, or the phase concentration has not reached the preset threshold, and the current coherent accumulation count is less than the maximum coherent accumulation count M. max When the coherent accumulation count at the same scanning position is increased, the distance and three-dimensional point of the corresponding scanning position are output when the preset validity condition is met. If the preset validity condition is not met after the maximum coherent accumulation count is reached, the corresponding scanning position is marked as a weak signal or invalid.
[0087] As an optional implementation, a fixed-ratio second beam splitter can be set in the detection receiving branch, with its two output ends connected to a second converging lens 500, a second photodetector 520, an additional converging lens, and an additional photodetector, respectively, forming two detection branches with different receiving sensitivities. The two branches synchronously acquire the same PCB backlight, and the processing unit 620 calibrates their complex gain and inherent phase difference, and selects the branch that is not saturated and meets the signal-to-noise ratio threshold to output the ranging result.
[0088] VII. Three-dimensional topography reconstruction
[0089] The processing unit 620 correlates the effective distance with the MEMS two-dimensional position feedback within the same sampling time window, and obtains three-dimensional points on the PCB surface based on the feedback voltage-scanning angle relationship and the internal and external parameters of the system; after fitting the PCB reference surface, it calculates the relative height of each point, and can identify warping, bulging, dents, foreign objects, abnormal solder joint height, raised pads, or raised components by combining local height gradients or backlight amplitude.
[0090] Example
[0091] Example 1:
[0092] To verify the effect of the scanning focusing lens assembly on the scanning range, the following near-field PCB scanning optical simulation embodiment is provided. This embodiment targets a single-axis full-scan width of approximately 120 mm, providing an optical simulation of a near-field PCB scanning focusing lens assembly. The modulated laser wavelength is 638 nm, and the maximum optical deflection angle of the MEMS micromirror 200 is ±14°; both the positive power lens 310 and the negative power lens 300 are made of N-BK7 glass, with an effective aperture of 25.4 mm.
[0093] The axial distance between the MEMS micromirror 200 and the front surface of the positive power lens 310 is 5.0 mm. The positive power lens 310 is a plano-convex lens with a nominal focal length of +60 mm, a convex surface radius of curvature of 30.00 mm, and a center thickness of 4.50 mm; the negative power lens 300 is a plano-concave lens with a nominal focal length of -50 mm, a concave surface radius of curvature of −25.00 mm, and a center thickness of 3.50 mm. The air gap between adjacent surfaces of the two lenses is 19.0 mm. The axial distance between the rear surface of the negative power lens 300 and the PCB surface under test 400 is 148.0 mm, and the mechanical distance between the MEMS micromirror 200 and the PCB surface under test 400 is 180.0 mm.
[0094] Based on the refractive index of N-BK7 at 638nm (approximately 1.5149), the paraxial matrix is calculated. The distance between the principal planes of the two lenses is approximately 21.97mm, the paraxial equivalent focal length of the scanning focusing lens group is approximately 230.73mm, and the paraxial angle-displacement transfer coefficient Br from the MEMS micromirror 200 to the PCB test surface 400 is approximately 240.16mm. Considering the non-paraxial effect of ±14° edge field of view, when performing Zemax full-field ray tracing using the actual curvature, thickness, material, and spacing shown in the figure, the center positions of the two edge field-of-view spots on the PCB test surface 400 are approximately -61.275mm and +61.275mm, respectively. The single-axis full-scan width calculated based on the spot center position is approximately 122.55mm; the full ray envelope, including the outer edge of the edge spot, is approximately -61.814mm to +61.814mm, corresponding to a total illuminated width of approximately 123.63mm. The dot plot results show that the RMS spot radius of both edge fields of view at ±14° is approximately 311.593 μm, and the geometric spot radius is approximately 561.263 μm. For comparison, while maintaining the MEMS optical deflection angle at ±14° and the mechanical distance L = 180.0 mm between the MEMS micromirror 200 and the PCB surface under test 400, and removing the positive and negative power lenses 310 to allow the beam to propagate freely, the single-axis full-scan width is 89.76 mm. When the scan width is defined by the spot center position, setting the aforementioned scanning focusing lens group increases the scan width by approximately 32.79 mm, a relative increase of approximately 36.5%. The above scan width, spot size, and comparative data are all results from optical simulation.
[0095] Before detection, B(p) is established, and then φ is obtained using one or more planar calibration plates with known distances. c (p). In actual testing, stray light subtraction, coherent accumulation and distance calculation are completed according to equations (4) to (7), and three-dimensional points on the PCB surface are generated by combining two-axis position feedback and system calibration relationship.
[0096] Example 2:
[0097] To illustrate the effect of position-dependent complex background correction on stray light ranging bias, a simulation example of two-dimensional scanning position-dependent complex background correction is provided. This example uses MATLAB to generate four-channel synchronously acquired data: AIN1 is the detection signal containing effective PCB backlight and stray light; AIN2 is the reference light signal; and AIN3 and AIN4 are the position feedback signals for the first and second scanning axes of the MEMS, respectively. The sampling rate is 100 kS / s, the synthesized intermediate frequency is 5 kHz, the phase-time-of-flight modulation frequency is 20 MHz, and the nominal axial distance is 100 mm. The first scanning axis frequency is 500 Hz with an optical peak-to-peak scanning angle of 10°, the second scanning axis frame rate is 5 Hz with an optical peak-to-peak scanning angle of 8°, and the feedback delays for the two axes are 640 μs and 660 μs, respectively.
[0098] The real Z-surface used for comparison is composed of a superposition of tilt terms, local convexity terms, and periodic undulation terms, where the units for x, y, and Z are all millimeters:
[0099]
[0100] The detection channel is superimposed with static stray light and dynamic background related to the two-axis scanning angle and scanning direction. During the calibration phase, the modulated laser source 100 is turned on and the PCB test surface 400 is not set, allowing the MEMS micromirror 200 to continuously perform 12 frames of full-field scanning along a predetermined trajectory. The scanning position p is determined based on the two-axis position feedback after delay compensation. The calibration data is grouped into 32×32 two-dimensional scanning position intervals and the fast axis scanning direction. The median of the real and imaginary parts of the complex phasors in each group is taken and smoothed to obtain the first photodetector calibration complex phasor Z. 10 (p) and the second photodetector calibration complex phasor Z 20 (p), and establish the position-related stray light complex calibrator B(p) according to Equation (3).
[0101] In actual testing, the PCB surface under test is set to 400, and the AIN1 detection signal, AIN2 reference signal, and two-axis position feedback signal are collected simultaneously. IQ demodulation and reference normalization are performed on AIN1 and AIN2. The scanning position p is determined according to the two-axis position feedback after delay compensation, and B(p) of the corresponding position is subtracted according to equation (4). Then, the distance and Z coordinate are calculated according to the corrected phase. In the simulation, the phase integer branch is selected using the known nominal distance. In the actual device, the known working distance constraint or multi-frequency measurement can be used to eliminate phase ambiguity.
[0102] like Figure 7 As shown, the median ratio of the amplitude of the background complex phasor to the effective return complex phasor is approximately 1.994%. Without background removal, the Z-axis deviation is approximately 7.423 mm and the root mean square error is approximately 10.818 mm; after correction, the Z-axis deviation is approximately 0.058 mm and the root mean square error is approximately 1.017 mm. These results are simulation results of synthetic data under given scanning trajectory, noise, stray light model, and filter bandwidth conditions, and do not represent the actual measured accuracy of the prototype.
[0103] The technical features in the above embodiments can be combined without contradiction. Equivalent changes or substitutions to the beam splitting method, scanning trajectory, number of lenses, demodulation method, calibration dimension, or three-dimensional reconstruction method do not depart from the concept of the present invention.
Claims
1. A PCB surface three-dimensional topography inspection device based on MEMS coaxial scanning, characterized in that, It includes a modulated laser source (100), a collimating lens (110), a beam splitter (120), a first converging lens (130), a first photodetector (140), a MEMS micromirror (200), a MEMS driving unit (210), a MEMS position feedback unit (220), a positive optical power lens (310), a negative optical power lens (300), a second converging lens (500), an aperture stop (510), a second photodetector (520), an FPGA modulation and control unit (600), a synchronous acquisition unit (610), and a processing unit (620). The FPGA modulation and control unit (600) provides a periodic modulation signal to the modulation laser source (100) so that the modulation laser source (100) outputs a modulation laser whose light intensity changes periodically according to a predetermined modulation frequency. The beam splitting element (120) splits the modulation laser collimated by the collimating lens (110) into a reference light and a detection light. The reference light is focused onto the photosensitive surface of the first photodetector (140) by the first converging lens (130). The MEMS driving unit (210) drives the MEMS micromirror (200) to perform two-dimensional deflection; the MEMS micromirror (200), positive optical power lens (310) and negative optical power lens (300) are arranged sequentially along the direction of detection light propagation. The positive optical power lens (310) and the negative optical power lens (300) constitute a net positive optical power scanning focusing lens group. After the detection light is deflected by the MEMS micromirror (200), it passes through the scanning focusing lens group and forms a scanning spot on the PCB surface to be tested (400). The backlight from the PCB passes through a negative power lens (300), a positive power lens (310), and a MEMS micromirror (200), and is guided by a beam splitter (120) to the detection and receiving branch. It then passes through an aperture stop (510) and is focused by a second converging lens (500) onto the photosensitive surface of the second photodetector (520). The MEMS position feedback unit (220) outputs a first scan axis position feedback signal and a second scan axis position feedback signal that characterize the actual position of the first scan axis and the actual position of the second scan axis of the MEMS micromirror (200); the synchronous acquisition unit (610) acquires the first photodetector reference signal, the second photodetector detection signal, the first scan axis position feedback signal and the second scan axis position feedback signal based on a common sampling time reference. The processing unit (620) performs IQ demodulation on the first photodetector reference signal and the second photodetector detection signal within the same sampling time window, determines the distance to the PCB surface to be tested based on the complex phasor relationship and phase difference of the two signals, and determines the two-dimensional scanning position corresponding to the distance based on the first scanning axis position feedback signal and the second scanning axis position feedback signal.
2. The PCB surface three-dimensional topography inspection device based on MEMS coaxial scanning according to claim 1, characterized in that, The beam splitter (120) is a wedge-shaped beam splitter or wedge-shaped beam splitter prism with two non-parallel optical surfaces, so that the parasitic reflected light generated by the non-working surface deviates from the receiving optical path of the second photodetector; the first converging lens (130) and the second converging lens (500) are both positive power lenses, and the photosensitive surfaces of the first photodetector (140) and the second photodetector (520) are located near the image plane or beam waist of the corresponding converging lens, respectively. The detection receiving branch has an aperture stop (510) between the beam splitter (120) and the second converging lens (500). The aperture of the aperture stop (510) is configured to allow effective backlighting from the PCB to pass through and to limit stray light deviating from the detection receiving direction from entering the second photodetector (520).
3. The PCB surface three-dimensional topography inspection device based on MEMS coaxial scanning according to claim 1, characterized in that, Let the focal length of the positive power lens (310) be f. P The focal length of the negative power lens (300) is f. N The axial distance between the principal planes of the two lenses is d, where f P >0, f N <0; the paraxial focal length F of the scanning focusing lens group satisfies And F >
0.
4. The PCB surface three-dimensional topography inspection device based on MEMS coaxial scanning according to claim 3, characterized in that, The focal length of the positive power lens (310), the focal length of the negative power lens (300), the distance between the two lenses, the effective aperture of the lens, and the position of the PCB test surface (400) are configured such that the PCB detection range is increased relative to the lensless free propagation detection range under the same MEMS optical deflection angle and the same axial distance, and the detection spot size within the PCB detection range is smaller than a preset defect size threshold; the PCB test surface (400) is located within the multi-field integrated optimal imaging surface of the scanning focusing lens group or its depth of field.
5. The PCB surface three-dimensional topography inspection device based on MEMS coaxial scanning according to claim 1, characterized in that, It also includes an FPGA modulation and control unit (600) and a MEMS driving unit (210). The FPGA modulation and control unit (600) provides a periodic modulation signal to the laser driving circuit of the modulation laser source (100) and provides a two-dimensional scanning driving signal to the MEMS driving unit (210). The synchronous acquisition unit (610) includes at least a first photodetector sampling channel, a second photodetector sampling channel, a first scanning axis position feedback channel and a second scanning axis position feedback channel. The four sampling channels adopt a common sampling time reference. When there is a fixed time difference between the sampling channels, the processing unit (620) compensates according to the pre-calibrated channel time difference.
6. The PCB surface three-dimensional topography inspection device based on MEMS coaxial scanning according to claim 1, characterized in that, The processing unit (620) determines the MEMS scanning position p based on the first scanning axis position feedback signal and the second scanning axis position feedback signal within the same sampling time window, and stores it in the normalized stray light complex phasor B(p) corresponding to the MEMS scanning position p obtained when the modulated laser source (100) is turned on and the PCB test surface (400) is not set; the processing unit (620) performs complex normalization on the second photodetector complex phasor Z2(p) obtained in actual detection and the first photodetector complex phasor Z1(p), and according to Obtain the effective luminous complex phasor Z of the PCB T (p).
7. The PCB surface three-dimensional topography inspection device based on MEMS coaxial scanning according to claim 6, characterized in that, The processing unit (620) coherently accumulates the PCB effective return complex phasors corresponding to the same MEMS scanning position p in multiple repeated scans, and determines the return phase by the in-phase component and quadrature component of the accumulated complex phasors; the modulated laser source (100) operates with a predetermined fixed modulation amplitude, and the processing unit (620) determines the number of coherent accumulations or judges whether the detection result is valid based on one or more of the amplitude of the PCB effective return complex phasors, the clipping state of the original sampling waveform, and the phase concentration of repeated measurements.
8. The PCB surface three-dimensional topography inspection device based on MEMS coaxial scanning according to claim 1, characterized in that, The detection receiving branch is also equipped with a fixed ratio second beam splitter. One output end of the second beam splitter is connected in sequence to a second converging lens (500) and a second photodetector (520) to form a first detection branch. The other output end is connected in sequence to an additional converging lens and an additional photodetector to form a second detection branch. The first detection branch and the second detection branch have different receiving sensitivities and synchronously collect the return light from the same PCB. The processing unit (620) calibrates the complex gain and inherent phase difference of the two detection branches respectively, and selects the unsaturated detection branch that meets the signal-to-noise ratio threshold to output the ranging result.
9. A method for detecting the three-dimensional topography of a PCB surface based on MEMS coaxial scanning, characterized in that, The method of implementing the PCB surface three-dimensional topography inspection device based on MEMS coaxial scanning as described in any one of claims 1 to 8 includes: The modulated laser source (100) is driven to output modulated laser, and a reference light and a detection light are formed by a beam splitter (120); the reference light is received by a first photodetector (140) through a first converging lens (130); The MEMS micromirror (200) is driven to deflect according to a predetermined two-dimensional scanning trajectory, so that the detection light passes through the positive optical power lens (310) and the negative optical power lens (300) in sequence along the propagation direction, scans the PCB surface (400) to be tested and forms a detection spot; the PCB return light is made to propagate in the opposite direction along the detection optical path, pass through the negative optical power lens (300), the positive optical power lens (310) and the MEMS micromirror (200) in sequence, and is guided to the detection receiving branch by the beam splitter (120), and then focused by the second converging lens (500) to the photosensitive surface of the second photodetector (520); The first photodetector reference signal, the second photodetector detection signal, the first scanning axis position feedback signal, and the second scanning axis position feedback signal are acquired based on a common sampling time reference. The complex phasor Z1(p) of the first photodetector reference light and the complex phasor Z2(p) of the second photodetector return light are obtained within the same sampling time window, and the MEMS scanning position p is determined based on the two position feedback signals. Based on the scan position p, the stray light complex phasor B(p) pre-established when the PCB test surface (400) is not set, and the effective return light complex phasor of the PCB is calculated. Coherent accumulation is performed on multiple PCB effective backlight complex phasors corresponding to the same scanning position p; the PCB surface distance is calculated based on the corrected phase difference of the accumulated complex phasors, and the distance is correlated with the scanning position p to reconstruct the three-dimensional morphology of the PCB surface.
10. The method for detecting three-dimensional morphology of PCB surface based on MEMS coaxial scanning according to claim 9, characterized in that, Also includes: With the modulated laser source (100) turned on and the PCB test surface (400) not set, the MEMS micromirror (200) traverses the predetermined scanning range. Simultaneously acquire the calibration signals of the first photodetector, the second photodetector, and two position feedback signals to obtain the complex phasor Z of the first photodetector calibration corresponding to each scanning position p. 10 (p) and the second photodetector calibration complex phasor Z 20 (p), according to Establish a location-dependent two-dimensional complex number calibration table; Data falling within the same two-dimensional scanning angle interval or the same position number p in multiple repeated scans are grouped together. The in-phase component and quadrature component are accumulated for the PCB effective backlight complex phasors of the same group. The validity of the measurement results is judged based on the amplitude of the PCB effective backlight complex phasors, the clipping state of the original sampled waveform, and the phase concentration. The PCB reference plane is fitted based on the three-dimensional coordinates of multiple effective scanning positions, and the height of each scanning position relative to the PCB reference plane is calculated. Based on at least one of the height, local height gradient, and backlight amplitude, PCB warping, bulging, dents, foreign objects, abnormal solder joint height, solder pad lifting, or component lifting are identified.
