A light source control method and system for ultraviolet spectral flat-field calibration of high-resolution astronomical spectrometers.
By independently configuring the TEC temperature control module in the high-resolution spectrometer to adjust the LED chip temperature, and combining it with control voltage and spectral feedback, the problems of insufficient photon flux and spectral instability in the short-wave ultraviolet region of the ultraviolet flat-field calibration light source were solved, achieving high-precision ultraviolet flat-field calibration effect.
Patent Information
- Application Number
- CN202611301188.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-26
- Publication Date
- 2026-11-17
AI Technical Summary
Existing high-resolution spectrometers have insufficient photon flux in the short-wave ultraviolet region due to the ultraviolet flat-field calibration light source, and the combined spectral shape is unstable due to the change of chip temperature in the multi-channel LED spectral response, making it difficult to meet the requirements of high-precision calibration.
By independently configuring the TEC temperature control module to adjust the LED chip temperature, and combining the control voltage Vdac_i and spectral feedback, a spectral model is established to achieve fine control of the combined spectral shape, thereby improving spectral stability and accuracy.
It achieves long-term stability and accuracy of ultraviolet flat-field spectra in high-resolution astronomical spectrometers, reduces spectral response drift, and is suitable for high-precision ultraviolet flat-field calibration.
Smart Images

Figure CN123124168A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of ultraviolet flat-field calibration of high-resolution astronomical spectrometers, multi-channel ultraviolet fiber-coupled LEDs, LED chip temperature (TEC) control, multi-channel spectral modeling, and spectral synthesis. Specifically, it relates to a light source control method and system that uses the output control voltage Vdac_i of a DAC8568 for basic regulation and the operating temperature Tc_i of the LED chip for fine regulation, thereby improving the accuracy of ultraviolet flat-field spectral control by adding the operating temperature dimension. Background Technology
[0002] During laboratory calibration and status verification, high-resolution astronomical spectrometers require stable and uniform ultraviolet flat-field light to assess the instrument's response consistency and spectral shape measurement stability within the target wavelength range. An ideal ultraviolet flat-field calibration source should provide a continuous, smooth, and long-term stable spectrum within the instrument's operating wavelength range, and possess sufficient spectral energy to avoid narrow emission lines or significant spectral shape fluctuations that can be resolved by a high-resolution spectrometer.
[0003] Current high-resolution spectrometers typically use halogen tungsten lamps, xenon lamps, deuterium lamps, or broadband LEDs based on phosphor conversion for flat-field calibration. Halogen tungsten lamps and some gas radiation sources can provide relatively smooth continuous radiation in the visible to near-ultraviolet range, but their spectral energy distribution is limited by Planck's law, decreasing rapidly in the short-wavelength ultraviolet region. When the operating wavelength is below approximately 400 nm, the photon flux may be insufficient, making it difficult to obtain a sufficient signal-to-noise ratio within a reasonable integration time, thus making them unsuitable for deep ultraviolet or near-ultraviolet flat-field calibration.
[0004] Deuterium lamps provide continuous radiation in the approximately 200 nm to 400 nm wavelength range and have good output capabilities in shorter wavelength bands, making them commonly used in ultraviolet spectroscopy measurements. However, the spectrum of a deuterium lamp is not completely smooth across the entire operating wavelength range, and its output may contain atomic or molecular emission structures. In high-resolution spectrometers, these structures may be resolved and superimposed on the flat-field data, affecting the accurate calibration of the instrument response. While broadband LEDs based on phosphor conversion can provide some broadband ultraviolet output, their spectra are typically formed by the combined emission of the semiconductor chip and phosphor conversion, resulting in a relatively fixed spectral structure that may exhibit significant spectral gaps or local fluctuations. Furthermore, phosphor emission is not a strictly continuous spectrum, and some fine structures may be resolved at higher resolutions, thus limiting their application in high-resolution ultraviolet flat-field calibration.
[0005] Multi-channel ultraviolet fiber-coupled LEDs (UV-FCLEDs) feature selectable center wavelengths, fast response speeds, easy modular integration, and independent channel-by-channel control. Each LED has a specific spectral width; by moderately overlapping adjacent spectral bands of multiple LEDs with different center wavelengths, the target UV band can be covered, forming a relatively smooth combined spectral shape. The output light from each channel is then bundled via fiber and input into an integrating sphere, achieving spectral component mixing and spatial and angular distribution homogenization. Therefore, multi-channel UV-FCLEDs are suitable as combined light sources for constructing configurable UV flat-field spectral shapes. However, the output spectrum of LEDs is highly sensitive to operating temperature. Changes in the thermal state of multi-channel LEDs during long-term operation directly affect the stability of the combined spectral shape, which is one of the key factors limiting their application in high-precision calibration scenarios.
[0006] However, the spectral response of LEDs drifts with operating conditions and chip temperature changes. The portion of the input power of an LED that is not converted into light output is primarily generated as heat. Chip temperature changes cause variations in the semiconductor bandgap and enhanced nonradiative recombination of charge carriers, typically manifesting as changes in peak light intensity, a shift in peak wavelength towards longer wavelengths, changes in the full width at half maximum (FWHM), and alterations in the overall spectral shape. The temperature response coefficients of LEDs with different center wavelengths are not entirely consistent. Therefore, chip temperature changes during long-term operation of multi-channel LEDs alter the actual contribution of each channel to the combined spectrum, causing deviations in the spectral shape set based on initial calibration results. Adjusting the driving voltage individually can change the basic output of a channel, but it is difficult to simultaneously address local spectral details. Establishing a relationship between chip temperature and spectral model parameters, and incorporating operating temperature as an additional control dimension, is beneficial for improving the control accuracy and stability of the combined spectral shape.
[0007] Therefore, there is a need for an ultraviolet spectral flat-field calibration light source that can provide a smooth, continuous, and long-term stable ultraviolet flat-field spectrum, and adjust the combined spectral shape by controlling voltage and chip temperature to meet the calibration requirements of high-resolution astronomical spectrometers. Summary of the Invention
[0008] To address the issues of resolvable emission structures in existing UV flat-field calibration light sources at high spectral resolution, and the problems of single-channel spectral drift and decreased accuracy of combined spectral shape matching caused by chip temperature variations in multi-channel LEDs, this invention proposes a UV flat-field calibration light source based on multi-channel UV fiber-coupled LEDs, chip temperature (TEC) control, and spectral feedback. This light source first achieves basic regulation through a control voltage Vdac_i, then uses a fine-tuning temperature model to determine the optimal operating temperature of each channel, using this temperature as the target temperature Tt_i for TEC adjustment, thereby improving the control accuracy and long-term stability of the UV flat-field spectral shape.
[0009] To achieve the above objectives, the present invention provides the following technical solution:
[0010] A light source control method for ultraviolet spectral flat-field calibration of a high-resolution astronomical spectrometer includes the following steps:
[0011] (1) Configure an independent TEC temperature control module for each channel in the multi-channel fiber-coupled ultraviolet LED, and set the target chip temperature Tt_i for each channel respectively; the temperature sensor is set at the LED chip to obtain the actual chip temperature Tc_i of the corresponding channel. The TEC temperature control module adjusts the TEC operating current according to the deviation between Tt_i and Tc_i to stabilize Tc_i near Tt_i.
[0012] (2) After the measured chip temperature Tc_i of each channel stabilizes, the output control voltage Vdac_i of each channel DAC module is used as the control variable. Based on the pre-established voltage basic model and temperature fine control model, the predicted spectrum S_i(λ,Vdac_i,Tc_i) of each channel under the current Vdac_i and Tc_i is generated. The predicted spectra of each channel are superimposed to obtain the system output predicted spectrum S_out(λ)=ΣS_i(λ,Vdac_i,Tc_i);
[0013] (3) Using the deviation between the predicted spectrum output by the system and the target ultraviolet flat field spectrum as the evaluation index, the initial value of Vdac_i of each channel is solved by the spectral matching algorithm and sent to each channel to obtain a basic combined spectrum close to the target spectral shape;
[0014] (4) Collect the actual output spectrum of the integrating sphere output end, compare the actual output spectrum with the target ultraviolet flat field spectrum, calculate the spectral matching error, and use it as the basis for temperature operating point calculation and matching result verification;
[0015] (5) First, perform basic regulation on each channel according to Vdac_i obtained in step (3) to obtain a combined spectrum close to the target spectrum shape; then keep Vdac_i at the corresponding basic operating point, calculate the optimal chip operating temperature of each channel based on the temperature fine regulation model and the spectral matching algorithm, and use the optimal chip operating temperature as the target temperature Tt_i of the corresponding TEC temperature control module; after Tc_i and the output spectrum are stable, compare and verify them.
[0016] Furthermore, the method for establishing the single-channel spectral response model includes:
[0017] (1) For the i-th LED channel, first stabilize Tc_i at the reference chip temperature Tc0, traverse Vdac_i within the safe operating range according to the preset step size, collect the single-channel measured spectrum under different Vdac_i, and establish the relationship between the control voltage and the basic spectral model parameters; then set multiple Tt_i operating points within the safe range under the fixed Vdac_i condition, collect the corresponding spectrum after Tc_i stabilizes, and record the spectral model parameters under each Tc_i;
[0018] (2) After performing dark background subtraction and noise reduction on the measured spectra at each working point, multiple candidate spectral models are used for fitting and comparison. The preferred model that is compatible with the measured spectral shape of the channel is selected, and the optimal values of the model parameters under each Vdac_i and Tc_i are recorded.
[0019] The relationship between the control voltage Vdac_i of the i-th channel at the reference chip temperature Tc0 and the parameter vector of the basic spectral model is established by interpolation, fitting, or lookup table: pV_i = fV_i(Vdac_i). Under the condition of fixed Vdac_i, the relationship between the chip temperature Tc_i and the parameter vector of the spectral model is established: pT_i(Vdac_i,Tc_i) = fT_i(Vdac_i,Tc_i), where pT_i(Vdac_i,Tc0) = pV_i(Vdac_i). The single-channel predicted spectrum of the i-th channel is expressed as: S_i(λ,Vdac_i,Tc_i) = G_i{λ;pT_i(Vdac_i,Tc_i)}.
[0020] Furthermore, the spectral matching algorithm constructs a control vector xV=[Vdac_1,Vdac_2,...,Vdac_n] using the Vdac_i of each channel, rather than the abstract channel weights, as the basic variable to be controlled. After obtaining the initial value of Vdac_i, Vdac_i is kept constant, and a temperature control vector xT=[Tt_1,Tt_2,...,Tt_n] is constructed using the chip operating temperature of each channel. Based on the temperature fine-tuning model, xT is determined by the spectral matching algorithm. The evaluation index is the root mean square error.
[0021] ;
[0022] Where n is the number of wavelength grid sampling points. For the target ultraviolet flat field spectrum at wavelength Strength at that location, The simulated output spectrum is given for the candidate combinations of Vdac_i and Tc_i.
[0023] Furthermore, the TEC temperature control module includes a TEC cooling semiconductor, a temperature sensor, and a TEC temperature controller; the temperature sensor is located at the chip of the fiber-coupled LED to obtain the measured chip temperature Tc_i; the heat dissipation surface of the LED chip is attached to the cold end of the TEC cooling semiconductor through thermal grease, so that the heat generated by the chip can be transferred to the cold end of the TEC; the hot end of the TEC cooling semiconductor is attached to a heat sink, and a fan is located below the heat sink; the TEC temperature controller adjusts the TEC operating current according to the deviation between Tt_i and Tc_i.
[0024] Furthermore, the center wavelength interval between adjacent LED channels is 10 nm to 20 nm, so that the emission spectra of adjacent channels partially overlap within the target band. After multiple LED channels are superimposed, a broadband ultraviolet spectrum covering 300 nm to 430 nm is formed.
[0025] A light source system for ultraviolet spectral flat-field calibration of a high-resolution astronomical spectrometer includes:
[0026] Host computer control unit;
[0027] Multi-channel fiber-coupled ultraviolet LEDs, each channel having a different center wavelength;
[0028] Each LED channel is equipped with an independent TEC temperature control module. Each TEC temperature control module is used to detect and adjust the measured chip temperature Tc_i of the corresponding LED channel so that Tc_i is stable near the set target temperature Tt_i.
[0029] The lower-level control unit is communicatively connected to the upper-level computer.
[0030] The DAC module is connected to the lower-level control unit and is used to convert the digital control quantity output by the lower-level control unit into the analog control voltage Vdac_i of each channel. The Vdac_i of each channel is input to the driving control terminal of the corresponding LED.
[0031] Fiber optic combiner is used to combine the ultraviolet light output from each LED channel into a single beam;
[0032] An integrating sphere, positioned after the fiber combiner, is used to mix and homogenize the combined ultraviolet light.
[0033] A fiber optic spectrometer is installed at the output end of the integrating sphere to collect the actual output spectrum of the integrating sphere and feed it back to the host computer.
[0034] The host computer is configured to execute the above method.
[0035] Furthermore, the TEC temperature control module includes a TEC cooling semiconductor, a temperature sensor, and a TEC temperature controller; the temperature sensor is located at the chip of the fiber-coupled LED to obtain the measured chip temperature Tc_i; the heat dissipation surface of the LED chip is attached to the cold end of the TEC cooling semiconductor through thermal grease, so that the heat generated by the chip can be transferred to the cold end of the TEC; the hot end of the TEC cooling semiconductor is attached to a heat sink, and a fan is located below the heat sink; the TEC temperature controller adjusts the TEC operating current according to the deviation between Tt_i and Tc_i.
[0036] Furthermore, the lower-level control unit is an STM32F103ZET6 microcontroller, the DAC module is a DAC8568 chip, and the microcontroller controls the DAC8568 chip to output the analog control voltage Vdac_i of each channel through the SPI interface.
[0037] Furthermore, the inner wall of the integrating sphere is provided with a diffuse reflection layer with high reflectivity in the ultraviolet band, which is used to improve the mixing and homogenization effect of the combined ultraviolet light in the integrating sphere.
[0038] Furthermore, the fiber optic combiner is a multi-input, one-output fiber optic combiner, or is composed of multiple Y-type cascaded fibers.
[0039] This invention employs a temperature sensor located at the LED chip to obtain the measured chip temperature Tc_i, and uses Tc_i as the channel temperature feedback quantity. The thermally conductive side of each LED chip is in thermal contact with the cold end of the TEC via thermal grease. The TEC temperature control module adjusts the TEC operating current based on the deviation between Tt_i and Tc_i, thereby achieving independent adjustment of the chip temperature of each channel.
[0040] Compared with the prior art, the present invention has at least the following beneficial effects:
[0041] 1. This invention utilizes multi-channel ultraviolet fiber-coupled LEDs, which are then bundled together by optical fibers and input into an integrating sphere to form a spatially uniform ultraviolet flat-field light with a settable spectral shape, which can be used for laboratory flat-field calibration of high-resolution astronomical spectrometers.
[0042] 2. This invention uses Vdac_i as the control quantity that the DAC8568 can directly execute. Under the reference chip temperature Tc0, the relationship between the control voltage and the basic spectral parameters is established, so that the algorithm result can be directly converted into the control voltage of each channel without the need to establish the calibration relationship between Vdac_i and the internal driving current of the LED.
[0043] 3. This invention adjusts the chip temperature Tc_i of each channel by using a temperature sensor and an independent TEC temperature control module located at the LED chip. This reduces chip temperature changes and thermal crosstalk between channels caused by working conditions, environmental changes, or prolonged lighting, thereby reducing spectral response drift.
[0044] 4. After the basic adjustment of Vdac_i, this invention obtains complete spectral parameters at different chip operating temperatures through a fine temperature control model, and determines the optimal operating temperature Tt_i through a spectral matching algorithm, so that the operating temperature participates in spectral matching as an additional control dimension, which is beneficial to improving the accuracy of spectral control.
[0045] 5. This invention, through the combination of basic voltage regulation, fine-tuning of chip operating temperature, and spectrometer feedback, is beneficial for compensating for device aging, environmental changes, and model errors, and improves the long-term reproducibility of ultraviolet flat-field spectral shapes. Attached Figure Description
[0046] Figure 1 This is a schematic diagram of the system structure of the present invention.
[0047] Figure 2 It is a flowchart of the control method.
[0048] Figure 3 This is a schematic diagram of a single-channel module structure. Detailed Implementation
[0049] The present invention will now be described in further detail with reference to the accompanying drawings.
[0050] The system structure of the present invention is as follows: Figure 1 As shown, it mainly includes a host computer control unit, an STM32F103ZET6 microcontroller, a DAC8568 module, a multi-channel ultraviolet fiber-coupled LED combination light source, a multi-channel TEC temperature control module, a heat sink, a fan, a fiber optic combiner, an integrating sphere, and a fiber optic spectrometer.
[0051] The host computer control unit is used to set the target ultraviolet flat-field spectrum, read the real-time spectrum fed back by the spectrometer, receive the measured chip temperature Tc_i returned by each TEC temperature control channel, and send the target chip temperature Tt_i to the TEC temperature control module. During calibration, the host computer calls the spectrometer to complete spectral data acquisition, background subtraction, and calibration data storage. During use, the host computer first calculates Vdac_i based on the basic control voltage model, then calls the temperature fine control model at the corresponding Vdac_i basic operating point, and uses differential evolution algorithm, least squares method, or other matching algorithm to determine Tt_i.
[0052] The multi-channel ultraviolet LED combined light source consists of multiple fiber-coupled ultraviolet LEDs with different center wavelengths. The peak wavelength of each LED channel is selected based on the target coverage band, the half-width at half maximum (WHM) of a single channel, the degree of spectral overlap between adjacent channels, and the requirements for smoothness of the superimposed spectral shape; preferably, the peak wavelength interval between adjacent LEDs can be set to 10 nm to 20 nm. The output optical fibers of each LED are connected to Y-type optical fibers or fiber combiners, and the combined ultraviolet light is input into an integrating sphere.
[0053] Each fiber-coupled LED is equipped with an independent TEC temperature control module, which includes a TEC temperature controller, a TEC cooling semiconductor, and a temperature sensor. The temperature sensor is located at the LED chip to obtain the measured chip temperature Tc_i. The heat dissipation surface of the LED chip is attached to the cold end of the TEC cooling semiconductor through thermal grease, allowing the heat generated by the chip to be transferred to the TEC cold end. The hot end of the TEC cooling semiconductor is attached to a heat sink, and a fan is located below the heat sink to provide forced air cooling. The TEC temperature controller adjusts the TEC operating current according to the deviation between Tt_i and Tc_i to stabilize the chip temperature within the set range.
[0054] The heat generated by the LED chip during operation is transferred to the TEC cold end via the chip's thermally conductive side. The TEC then transfers the heat absorbed at the cold end to the hot end. The TEC hot end dissipates heat to the environment through a heat sink and a fan below, maintaining the heat dissipation capacity required for chip temperature regulation.
[0055] The temperature sensor is located at the LED chip, and Tc_i serves as the measured chip temperature feedback for the corresponding channel; the system does not need to establish a correspondence between Tc_i and the internal drive current of the LED. Adjusting Tc_i through an independent TEC temperature control module can reduce the impact of chip temperature changes on peak intensity, peak wavelength, and full width at half maximum (FWHM). Multiple LED channels are each equipped with an independent TEC temperature control module, and thermal crosstalk between channels can be reduced using heat insulation pads, heat insulation grooves, or separate brackets.
[0056] The STM32F103ZET6 microcontroller serves as the lower-level control core, receiving multi-channel control commands from the upper-level computer and parsing, buffering, and verifying the communication data. Subsequently, the microcontroller controls the DAC8568 module to output corresponding analog voltage signals via interfaces such as SPI, thereby adjusting the output light intensity or power of each fiber-coupled LED channel.
[0057] The DAC8568 module converts the digital control signals output from the STM32F103ZET6 microcontroller into multiple analog control voltages. Each output of the DAC8568 corresponds to a fiber-coupled LED control channel, and the voltage of the i-th output is denoted as Vdac_i. Vdac_i is a hardware control signal that can be directly issued by the system and is input to the corresponding LED's driver control terminal to adjust the LED's internal drive current, thereby changing the light output intensity and spectral response of that channel.
[0058] Since changes in Vdac_i will cause changes in the internal driving current, power dissipation, and chip temperature of the LED, the system adjusts Vdac_i within the experimentally calibrated stable operating range. After each change in Vdac_i, the system waits for Tc_i and the output spectrum to stabilize before performing spectral acquisition or subsequent control.
[0059] The integrating sphere is used to mix and homogenize the light emitted by multi-channel LEDs. Light from different LED channels undergoes multiple diffuse reflections on the inner wall of the integrating sphere, forming a spatially uniform, fully mixed ultraviolet flat-field light. A spectrometer is connected to a fixed measurement position at the output of the integrating sphere. During calibration, it acquires single-channel spectra; during use, it acquires the system output spectra and feeds them back to the host computer.
[0060] To establish the relationship between control voltage, chip operating temperature, and LED spectral response, the Tc_i of each LED channel was first stabilized at the reference chip temperature Tc0. The Vdac_i within the safe operating range was then traversed, and the measured single-channel spectra after dark background subtraction and noise reduction were collected to establish the voltage basic model data. Then, under the condition of fixed Vdac_i, multiple Tt_i operating points within the safe range were set. After Tc_i stabilized, the measured spectra were collected, and the spectral model parameters at each operating temperature were recorded.
[0061] System workflow as follows Figure 2 The process is as follows: First, set the initial target chip temperature Tt_i for each channel and wait for the measured chip temperature Tc_i to stabilize. Then, at the reference chip temperature Tc0, call the basic control voltage model and use a spectral matching algorithm to solve for the initial value of Vdac_i for each channel and form a basic combined spectral shape. Next, keeping Vdac_i fixed for each channel, call the temperature fine-tuning model to generate predicted spectra at different chip operating temperatures, and use differential evolution algorithm, least squares method, or other matching algorithms to calculate the optimal operating temperature for each channel, setting it as Tt_i. The spectrometer collects the actual output spectrum and calculates the root mean square error. After Tc_i and the spectrum stabilize, repeat the fixed sequence of "basic Vdac_i control—fine-tuning of operating temperature—spectral verification" until the preset matching requirements are met.
[0062] To enable those skilled in the art to understand and implement the present invention, the following is combined with Figure 1 , Figure 2 , Figure 3 The assembly and chip temperature control debugging of the multi-channel ultraviolet flat field light source, the basic model of control voltage, the fine temperature control model, and the target spectrum matching and feedback verification in the order of "first basic control voltage control, then fine temperature control" are explained.
[0063] Example 1: Assembly and Chip Temperature Control Adjustment of Multi-channel Ultraviolet Flat Field Light Source
[0064] The system in this embodiment includes a host computer, an STM32F103ZET6 microcontroller, a DAC8568 module, multi-channel ultraviolet fiber-coupled LEDs, a multi-channel TEC temperature control module, a heat sink, a fan, a Y-type fiber or a fiber combiner, an integrating sphere, and a spectrometer. The center wavelengths of the multiple LEDs are selected based on the target ultraviolet band, the half-width at half maximum (FWHM) of a single channel, and the overlap requirements of adjacent spectral bands, ensuring that there are no obvious spectral gaps after the adjacent channels are superimposed. The output fibers of each LED are connected to a Y-type fiber or a fiber combiner, and the combined ultraviolet light is input to the integrating sphere; the spectrometer is connected to a fixed measurement position at the output end of the integrating sphere.
[0065] Its single-channel module structure is as follows Figure 3 As shown, a temperature sensor is installed at each fiber-coupled LED chip to measure the actual chip temperature Tc_i of that channel. The heat dissipation surface of the LED chip is attached to the cold end of the TEC cooling semiconductor via thermal grease, allowing the heat generated by the chip to be transferred to the TEC cold end. The hot end of the TEC cooling semiconductor is attached to a heat sink, and a fan is positioned below the heat sink to dissipate heat. Tc_i serves as the chip temperature feedback value, used for chip temperature control and spectral model invocation.
[0066] After assembly, first check the thermal grease application, the bonding between the chip's thermally conductive side and the TEC's cold end, the fiber optic connection, the heatsink installation, and the fan rotation status. Then, start the fan and TEC temperature control module, setting the target chip temperature Tt_i for each channel, and continuously collecting Tc_i, TEC current, and TEC hot-end temperature. Once Tc_i enters the allowable deviation range and remains stable within the specified stabilization time, change Vdac_i for each channel one by one, checking if the chip temperature remains stable near Tt_i. Then, simultaneously turn on multiple channels to check inter-channel thermal crosstalk, TEC hot-end temperature rise, and air cooling capacity; if necessary, reduce the thermal impact of adjacent channels using thermal pads, thermal grooves, or separate brackets. If hot-end overheating, TEC overcurrent, or fan failure occurs, stop the corresponding TEC and LED channel.
[0067] The host computer sends control commands to the STM32F103ZET6 via serial port, and the microcontroller controls the DAC8568 via the SPI interface. The i-th output voltage Vdac_i of the DAC8568 is input to the corresponding LED driver control terminal to adjust the internal drive current of the LED, thereby changing the basic light output and spectral response of that channel; in this embodiment, the correspondence between Vdac_i and the internal drive current of the LED is not measured. After temperature stabilization and optical path checks are completed, Vdac_i is set to be within the calibrated operating range, so that each LED channel enters the predetermined operating state.
[0068] Example 2: Establishment of Basic Voltage Model and Fine Temperature Control Model
[0069] This embodiment illustrates a method for single-channel spectral calibration and model building using Vdac_i and measured chip temperature Tc_i as variables, including a voltage-based model at a reference chip temperature and a fine-tuning model for the relationship between chip temperature and spectral model parameters under a fixed control voltage.
[0070] Before calibration, turn off all LED channels and complete the dark background acquisition, wavelength calibration, and integration time setting of the spectrometer. Then, only turn on the i-th LED to be calibrated, and keep the other LED channels off; the channel to be calibrated is connected to the integrating sphere via a Y-type fiber or fiber combiner, and the spectrometer acquires the spectrum at the output of the integrating sphere.
[0071] When calibrating the i-th LED single channel, first stabilize Tc_i at the reference chip temperature Tc0, and then iterate through Vdac_i within the safe operating range of this channel according to a preset step size. After each change in Vdac_i, wait for Tc_i and the LED output to meet the stability conditions before acquiring the corresponding spectrum to form the voltage basic model data. Subsequently, select each Vdac_i operating point or a preset representative Vdac_i operating point, keep Vdac_i unchanged, and sequentially set multiple target temperatures Tt_i within the safe range. For each Tt_i operating point, wait for Tc_i to enter the allowable deviation range and stabilize before acquiring the corresponding spectrum and recording the parameter data.
[0072] For each Vdac_i and Tc_i operating point, spectral data is continuously acquired a preset number of times. The raw spectra are then subjected to dark background subtraction, outlier removal, repeated measurement averaging, and wavelength grid unification to form measured spectral data for modeling. Each data record includes the channel number, Vdac_i, Tc_i, Tt_i, ambient temperature, spectrometer acquisition parameters, and the corresponding measured spectrum.
[0073] For each set of processed measured spectra, a Gaussian model, a Gaussian-Lorentz model, a modified Gaussian model, or other candidate models were used for fitting. Initial values of the parameters for each candidate model were obtained based on the measured spectral shape, and then the parameters were adjusted using least squares, nonlinear optimization, or other parameter optimization methods to reduce the error between the fitted spectrum and the measured spectrum.
[0074] For each LED channel, the fitting error, peak wavelength error, full width at half maximum (FWHM) error, residual distribution, and parameter stability of different candidate models are compared, and the preferred spectral model G_i that best matches the measured spectral shape of that channel is selected. The optimal values of the model parameters under each Vdac_i and Tc_i are recorded to form a calibration data table; under the condition Tc_i=Tc0, pV_i=fV_i(Vdac_i) is established, and under the condition of fixed Vdac_i, pT_i(Vdac_i,Tc_i)=fT_i(Vdac_i,Tc_i) is established, where pT_i(Vdac_i,Tc0)=pV_i(Vdac_i).
[0075] The predicted spectrum of the i-th LED single channel is represented as S_i(λ,Vdac_i,Tc_i)=G_i{λ;pT_i(Vdac_i,Tc_i)}. For a given Vdac_i and the current Tc_i, the corresponding complete spectral parameters can be obtained and the predicted spectrum can be generated by the relationship between temperature and spectral model parameters.
[0076] After establishing the calibration data table, verification spectra were collected at stable operating points Tt_i and Vdac_i, which were not directly used to establish parameter relationships. The measured spectra were compared with the predicted spectra calculated from S_i(λ,Vdac_i,Tc_i). When the prediction error met the preset requirements, the channel model was used for subsequent multi-channel spectral matching. If the error did not meet the requirements, calibration data for the corresponding operating points were supplemented, or the interpolation and fitting methods of the parameter functions were adjusted.
[0077] After comprehensive calibration, the system output spectrum is expressed as Sout(λ)=ΣS_i(λ,Vdac_i,Tc_i). Each single-channel model was obtained under the same hardware and optical path conditions as actual operation, therefore no additional fixed efficiency coefficient is introduced when multiple channels are superimposed.
[0078] Example 3: Spectral matching based on a control voltage-based regulation and temperature fine-tuning model
[0079] Upon system startup, the fan and TEC temperature control module are activated first to check the TEC hot-end temperature, TEC current, and fan status. Then, the initial target chip temperature Tt_i for each channel is set. After each Tc_i meets the stability condition, the single-channel spectral response data is retrieved, and the initial Vdac_i for each channel is set according to the target UV flat-field spectrum.
[0080] Before spectral matching, the target spectrum, single-channel calibration spectrum, and spectrometer-measured spectrum are unified to the same wavelength grid. First, the Tc_i of each channel is stabilized at the reference chip temperature Tc0. The differential evolution algorithm or other spectral matching algorithm uses xV=[Vdac_1,...,Vdac_n] as the control vector, and replaces the abstract channel weight k_i with the control voltage Vdac_i, which can be directly executed by the DAC8568. For each set of candidate control voltages, pV_i(Vdac_i) is called to generate the basic predicted spectrum for each channel and superimposed to obtain the corresponding multi-channel predicted output spectrum.
[0081] The root mean square error (RMSE) between the predicted output spectrum and the target UV flat-field spectrum is used as the error evaluation index and optimization target of the matching algorithm. The RMSE is expressed as:
[0082] ;
[0083] Where n is the number of wavelength grid sampling points. For the target ultraviolet flat field spectrum at wavelength Strength at that location, The simulated output spectrum is given for candidate combinations of Vdac_i and Tc_i. Differential evolutionary algorithms or other matching algorithms are used within the stable operating range of each Vdac_i and under TEC safety constraints to reduce M. RMSE Iteratively update the control vector x for the target to obtain the initial combination of Vdac_i.
[0084] The obtained Vdac_i is sent to the corresponding LED channel via an STM32F103ZET6 and a DAC8568, and the spectrometer acquires the actual output spectrum of the integrating sphere. The host computer keeps Vdac_i constant and calls the temperature fine-tuning model to generate predicted spectra at different chip operating temperatures. A temperature control vector xT=[Tt_1,...,Tt_n] is constructed using the target temperature Tt_i of each channel. Differential evolution algorithm, least squares method, or other matching algorithms are used to determine xT with the goal of reducing the root mean square error between the predicted output spectrum and the target spectrum. The determined Tt_i is sent to the corresponding TEC temperature control module. After Tc_i and the output spectrum stabilize, the matching results are measured again and verified.
[0085] Vdac_i is updated via STM32F103ZET6 and DAC8568, and Tt_i is sent to the corresponding TEC temperature control channel. After each change to Tt_i, it is necessary to wait for Tc_i and the output spectrum to stabilize before calculating the root mean square error and starting the next round of "basic control voltage adjustment - fine control of operating temperature - spectral verification" to avoid using unstable spectra during temperature changes for matching judgment.
[0086] The matching ends when the root mean square error meets the preset requirements, the improvement in root mean square error through continuous iterations is lower than the preset value, or the maximum number of iterations is reached. The system saves the target spectrum, the measured output spectrum, Vdac_i, Tc_i, Tt_i of each channel, the ambient temperature, the spectrometer acquisition parameters, and the protection status for subsequent repeated calls, periodic verification, or recalibration after device replacement.
[0087] The above embodiments are merely typical implementations of the present invention. The fiber optic combiner can also be replaced by a multi-input Y-type fiber; the number of LED channels, center wavelength, packaging location of the temperature sensor at the chip, TEC temperature control structure, candidate spectral model, parameter function form, and optimization algorithm can be adjusted according to the actual needs of the target ultraviolet band and instrument. Under the technical concept of not changing the channel-level chip temperature control, using Vdac_i for basic spectral regulation, using a fine-tuning temperature control model for operating temperature matching, and using feedback verification based on measured spectra, the above equivalent substitutions are all optional implementations of the present invention.
Claims
1. A light source control method for ultraviolet spectral flat-field calibration of a high-resolution astronomical spectrometer, characterized in that, Includes the following steps: (1) Configure an independent TEC temperature control module for each channel in the multi-channel fiber-coupled ultraviolet LED, and set the target chip temperature Tt_i for each channel respectively; the temperature sensor is set at the LED chip to obtain the actual chip temperature Tc_i of the corresponding channel. The TEC temperature control module adjusts the TEC operating current according to the deviation between Tt_i and Tc_i to stabilize Tc_i near Tt_i. (2) After the measured chip temperature Tc_i of each channel stabilizes, the output control voltage Vdac_i of each channel DAC module is used as the control variable. Based on the pre-established voltage basic model and temperature fine control model, the predicted spectrum S_i(λ,Vdac_i,Tc_i) of each channel under the current Vdac_i and Tc_i is generated. The predicted spectra of each channel are superimposed to obtain the system output predicted spectrum S_out(λ)=ΣS_i(λ,Vdac_i,Tc_i); (3) Using the deviation between the predicted spectrum output by the system and the target ultraviolet flat field spectrum as the evaluation index, the initial value of Vdac_i of each channel is solved by the spectral matching algorithm and sent to each channel to obtain a basic combined spectrum close to the target spectral shape; (4) Collect the actual output spectrum of the integrating sphere output end, compare the actual output spectrum with the target ultraviolet flat field spectrum, calculate the spectral matching error, and use it as the basis for temperature operating point calculation and matching result verification; (5) First, perform basic regulation on each channel according to Vdac_i obtained in step (3) to obtain a combined spectrum close to the target spectrum shape; then keep Vdac_i at the corresponding basic operating point, calculate the optimal chip operating temperature of each channel based on the temperature fine regulation model, and use the optimal chip operating temperature as the target temperature Tt_i of the corresponding TEC temperature control module; after Tc_i and the output spectrum are stable, compare and verify them.
2. The method according to claim 1, characterized in that, The method for establishing the single-channel spectral response model includes: (1) For the i-th LED channel, first stabilize Tc_i at the reference chip temperature Tc0, traverse Vdac_i within the safe operating range according to the preset step size, collect the single-channel measured spectrum under different Vdac_i, and establish the relationship between the control voltage and the basic spectral model parameters; then set multiple Tt_i operating points within the safe range under the fixed Vdac_i condition, collect the corresponding spectrum after Tc_i stabilizes, and record the complete spectral model parameters under each Tc_i; (2) After performing dark background subtraction and noise reduction on the measured spectra at each working point, multiple candidate spectral models are used for fitting and comparison. The preferred model that is compatible with the measured spectral shape of the channel is selected, and the optimal values of the model parameters under each Vdac_i and Tc_i are recorded. (3) Establish the relationship between the control voltage Vdac_i of the i-th channel and the parameter vector of the basic spectral model at the reference chip temperature Tc0 by interpolation, fitting or lookup table: pV_i=fV_i(Vdac_i); under the condition of fixed Vdac_i, establish the relationship between the chip temperature Tc_i and the parameter vector of the spectral model: pT_i(Vdac_i,Tc_i)=fT_i(Vdac_i,Tc_i), where pT_i(Vdac_i,Tc0)=pV_i(Vdac_i). The single-channel predicted spectrum of the i-th channel is expressed as: S_i(λ,Vdac_i,Tc_i)=G_i{λ;pT_i(Vdac_i,Tc_i)}.
3. The method according to claim 1, characterized in that: The spectral matching algorithm constructs a control vector xV=[Vdac_1,Vdac_2,...,Vdac_n] using the Vdac_i values of each channel, rather than the abstract channel weights, as the basic variable to be controlled. After obtaining the initial value of Vdac_i, Vdac_i is kept constant, and a temperature control vector xT=[Tt_1,Tt_2,...,Tt_n] is constructed using the chip operating temperature of each channel. Based on the temperature fine-tuning model, xT is determined by the spectral matching algorithm. The evaluation index is the root mean square error. ; Where n is the number of wavelength grid sampling points. For the target ultraviolet flat field spectrum at wavelength Strength at that location, The simulated output spectrum is given for the candidate combinations of Vdac_i and Tc_i.
4. The method according to claim 1, characterized in that: The TEC temperature control module includes a TEC cooling semiconductor, a temperature sensor, and a TEC temperature controller. The temperature sensor is located at the chip of the fiber-coupled LED to obtain the measured chip temperature Tc_i. The heat dissipation surface of the LED chip is attached to the cold end of the TEC cooling semiconductor through thermal grease, so that the heat generated by the chip can be transferred to the cold end of the TEC. The hot end of the TEC cooling semiconductor is attached to a heat sink, and a fan is located below the heat sink. The TEC temperature controller adjusts the TEC operating current according to the deviation between Tt_i and Tc_i.
5. The method according to claim 1, characterized in that: The center wavelength interval between adjacent LED channels is 10 nm to 20 nm, so that the emission spectra of adjacent channels partially overlap within the target band. After multiple LED channels are superimposed, a broadband ultraviolet spectrum covering 300 nm to 430 nm is formed.
6. A light source system for ultraviolet spectral flat-field calibration of a high-resolution astronomical spectrometer, characterized in that, include: Host computer control unit; Multi-channel fiber-coupled ultraviolet LEDs, each channel having a different center wavelength; Each LED channel is equipped with an independent TEC temperature control module. Each TEC temperature control module is used to detect and adjust the measured chip temperature Tc_i of the corresponding LED channel so that Tc_i is stable near the set target temperature Tt_i. The lower-level control unit is communicatively connected to the upper-level computer. The DAC module is connected to the lower-level control unit and is used to convert the digital control quantity output by the lower-level control unit into the analog control voltage Vdac_i of each channel. The Vdac_i of each channel is input to the driving control terminal of the corresponding LED. Fiber optic combiner is used to combine the ultraviolet light output from each LED channel into a single beam; An integrating sphere, positioned after the fiber combiner, is used to mix and homogenize the combined ultraviolet light. A fiber optic spectrometer is installed at the output end of the integrating sphere to collect the actual output spectrum of the integrating sphere and feed it back to the host computer. The host computer is configured to perform the method according to any one of claims 1 to 5.
7. The system according to claim 6, characterized in that: The TEC temperature control module includes a TEC refrigeration semiconductor, a temperature sensor, and a TEC temperature controller. The temperature sensor is located at the chip of the fiber-coupled LED. The thermally conductive substrate or heat sink side of the LED chip is in thermal contact with the cold end of the TEC refrigeration semiconductor via thermally conductive silicone grease. The hot end of the TEC refrigeration semiconductor is attached to a heat sink. A fan is located below the heat sink. Thermal insulation pads, thermal insulation grooves, or separate brackets are provided between adjacent LED channels to reduce thermal crosstalk between channels.
8. The system according to claim 6, characterized in that: The lower-level control unit is an STM32F103ZET6 microcontroller, and the DAC module is a DAC8568 chip. The microcontroller controls the DAC8568 chip to output the analog control voltage Vdac_i of each channel through the SPI interface.
9. The system according to claim 6, characterized in that: The inner wall of the integrating sphere is provided with a diffuse reflection layer with high reflectivity in the ultraviolet band, which is used to improve the mixing and homogenization effect of the combined ultraviolet light in the integrating sphere.
10. The system according to claim 6, characterized in that: The fiber optic combiner is a multi-input, one-output fiber optic combiner, or it is composed of multiple Y-type cascaded fibers.
