3 Instrumentation

3.1 Optical layout

The optical layout of the video magnetograph along with the light feed is given in Fig. 1.

  

Figure 1: Optical layout of USO video magnetograph. L1 - 15 cm objective lens, HF - heat filter, FS - field stop, L2 - relay lens, PFW - pre-filter wheel, KDP - KD*P electro-optic modulator, LP - linear polarizer, FP - Fabry-Perot etalon filter, CCD - CCD video camera, EN - Wooden enclosure

The objective lens L1 $(\phi = 15$ cm, f/15 ) makes a solar image of 22 mm diameter at the focal plane where a field stop FS is placed. A heat filter HF is used to reduce the heat load on the optics and CCD camera saturation by blocking the IR radiation. A portion of the image is enlarged by a factor of 2.7 by a relay lens L2 to yield a 60 mm solar image (f/40). The pre-filter wheel PFW, KD^*P modulator, linear polarizer LP and FP etalon filter are placed in the telecentric beam following the lens L2. The fast axis of the KD^*P crystal makes an angle of 45$^\circ$ with the linear polarizer LP. Finally the image is recorded by a CCD camera. To avoid scattered light and ambient temperature variations, all the components are mounted on the optical bench, and are enclosed in a wooden box EN.

3.2 Fabry-Perot narrow band filter

A 60 mm aperture, high finesse voltage tunable LiNbO₃ FP etalon acquired from CSIRO Australia, is used as a narrow band filter to isolate a portion at the wing of the selected absorption line. The FP etalon is made by using a LiNbO₃ substrate of 0.175 mm thickness, both sides of which are polished to $\lambda/200$ and coated with high reflective SiO₂ and Ta₂O₅ films. The resultant etalon has a reflectivity of 93% over a wavelength range of 5000 to 6700 Å. The tunability of the etalon is achieved by applying high voltage across the LiNbO₃ wafer which varies its refractive index. A conductive ITO (Indium Tin Oxide) coating is deposited for the application of electric field across the crystal. The high voltage terminals made of gold wires are bonded to the ITO coating with the help of silver epoxy. The details of the FP parameters are given in Table 1. In order to avoid the drift of the FP band pass due to a change in the ambient temperature, the etalon is enclosed in a constant temperature oven. The oven temperature is maintained at 43 $^\circ$C with a stability of $\pm$ 0.05 $^\circ$C which provides a wavelength stability 5 mÅ. A bench test was performed for determining different parameters of the FP etalon, by placing it in front of the USO Littrow spectrograph coupled with 15 cm Zeiss coudé telescope. This spectrograph has a dispersion of 0.047 Å/pixel and an f-ratio of f/40 similar to the light beam used for the video magnetograph. Figure 2a show the observed FP channel spectrum recorded near CaI 6122 Å.

  

Figure 2: The observed and corrected Fabry-Perot channel spectra: a) and b) show the observed and corrected channel spectra at 6122 Å. The dotted line shows the solar spectrum, the intensity is plotted in arbitrary units

The observed profiles are corrected for the spectrograph instrumental broadening, using the CaI 6122 Å line profile obtained from KPNO digital solar spectral atlas (Debi et al. 1998). The corrected profile is shown in Fig. 2b. At H$\alpha$ 6563 Å also, the results are similar. Solar spectra were recorded by removing the FP etalon placed in front of the spectrograph slit. The measured Full-width at half maximum (FWHM), Free-Spectral range (FSR), and finesse (FSR/FWHM) of the etalon are given in the Table 1. The voltage tunability of the etalon was determined by measuring the wavelength shift of the channel spectrum as a function of voltage in the range of -3000 to + 3000 Volts in steps of 50 Volts. With in this range the wavelength shift is linear with voltage and is found to 0.45 Å per 1000 Volts.

To isolate 6122 Å (CaI) and 6563 Å (H$\alpha$) lines two narrow band interference filters of passband 2.6 Å and 3.5 Å respectively are placed before the FP, and the voltage on the FP is accordingly changed to obtain VMG in 6122 Å line and H$\alpha$ filtergrams. These two pre-filters are also enclosed in separate temperature controlled ovens and mounted on computer controlled filter wheel, in order to make near simultaneous magnetic field, chromospheric, and photospheric observations.

  

Table 1: Narrow band filter parameters

3.3 Circular polarization analyzer

A KD^*P electro-optic quarter wave plate and a linear polarizer constitute the circular polarization analyzer for measuring the longitudinal magnetic field. The modulator uses a thin (< 3 mm) Meadowlark KD^*P crystal which makes these cells suitable for using them in convergent light beam slower than f/20 (West 1989). The fast axis of the KD^*P crystal is aligned at 45$^\circ$ to the transmission axis of the linear polarizer and housed in an insulated enclosure. The KD^*P crystal is converted to $\pm \lambda/4$ retarders (at $\lambda = 6122$ Å) by applying $\pm 2100$ Volts. In order to operate the KD^*P modulator a fast switchable high voltage power supply was made, which can provide $\pm 2100$ Volts on the application of low voltage TTL pulses at the input. The TTL pulses are obtained from the centronics port of the image acquisition system and synchronized with the image frame acquisition, such that alternately captured video frames contain left or right circularly polarized images. The left and right circular Zeeman components are converted in to two mutually perpendicular linear polarizations depending on the sign of the applied voltage. One of the components is blocked by the linear polarizer allowing the selection of the image corresponding to left or right circular polarization in the emerging beam.

3.4 Detector - Cohu 4710 series CCD camera

The detector used in the video magnetograph is a Cohu make CCD monochrome camera with image sensor chip TC277 from Texas Instruments. This camera provides high resolution images with sensitivity as low as 0.25 lux, zero geometric distortion and no lag or retention of images. The CCD used is a frame transfer device with $9.2 \times 16.8 \,\mu$m pixels arranged in $699 \times 576$ array in which half of the pixel rows are masked for image storage and the other half are exposed to light. This makes a resultant image area of $6.4 \times 4.8$ mm on the CCD chip. CCIR scanning system is employed for the image read out, where a single video field (one video frame contains two video fields -odd and even) takes 1/50 seconds for scanning. During every video field the charge accumulated in the storage section is read out while the image section is exposed. The vertical blanking pulses after each video field (two vertical blanking pulses for each video frame) in the CCIR video output is detected through software and used for the synchronization of the KD^*P switching and the entire data acquisition and reduction process. In the present optical set-up an area of $\sim4 \times 3$arcmin of the solar disk is imaged by the camera, which gives a resolution of $0.3 \times 0.6$ arcsec/pixel.

3.5 Image acquisition and control system

Block diagram of the image acquisition and control system of the video magnetograph is given in Fig. 3.

  

Figure 3: The schematic diagram of the optics, image acquisition system and control electronics

An Innovision Inc., workstation, based on Motorola, MC68030, single board VMEbus computer, and integrated with Imaging Technology Series 150 image processing modules makes the complete data acquisition and control system. The image processing modules consists of one analogue-to-digital interface (ADI) unit, two frame buffers (FB) and one arithmetic and logic unit (ALU); all connected to the VMEbus of the host computer. The combination of the host computer, series 150 modules and suitable software can perform complex on-line real time digital image processing tasks such as averaging of images and subtraction. The ADI employs an 8-bit flash A/D converter at 10 MHz sampling rate, which digitizes the video input signal to 256 gray levels. The video bus transmits the digitized data (VDI) to all the other modules. The frame buffers FB0 and FB1 contains the image storage required for the real time processing of the data. Each FB consists of a single 512 by 512 by 16-bit frame store (FRAME A) and two 512 by 512 by 8-bit frame stores (B1 and B2). ALU is a pipelined image processor which provides real time image processing capabilities when used with ADI and FB. The images from FRAME A and FRAME B are processed by ALU and the result is stored in FRAME A, which is finally transferred to the host computer.

A software was developed using C-language and ITEX 150/151 image processing library (Imaging Tech, Inc) for real time processing and control of the video magnetograph (Mathew 1998). The flow chart illustrates the sequence of operation performed by the software during a single acquisition cycle (Fig. 4a).

  

Figure 4: Flow chart showing: a) the operations involved in a single cycle of image acquisition, b) VMG operation for a single video magnetogram

The centronics port data bits D0 to D5, performs various control operations such as opening and closing the telescope shutter, changing the pre-filter, moving the relay lens for focusing the image for H$\alpha$ and CaI, changing the polarizer wheel, and switching the KD^*P high voltage supply. A single acquisition cycle consists of several operations to obtain selected number of magnetograms, CaI 6122 Å and H$\alpha$ images. To obtain the photospheric CaI 6122 Å images, the filter is kept tuned at the same position where the Stokes V signal is taken. The flow chart in Fig. 4b shows various steps involved to obtain a single video magnetogram. For making magnetograms, the Series150 modules are set up for continuous acquisition and adding up of the images. The ALU is programmed for adding the incoming video data on to the pervious image already present in the FRAME A. The alternate incoming video images are routed to FRAME A of FB0 and FB1 respectively. This is synchronized with the KD^*P high voltage switching such that the alternate frames are taken in the left and right circular polarizations. This process can be repeated for a pre-selected number of frames (maximum of 256 frames) to increase the signal to noise (S/N) ratio. To integrate 256 frames, the systems takes 12 seconds. The whole sequence of one observation, takes about 1 minute which consists of a video magnetogram averaged over 256 frames, CaI photospheric image and H$\alpha$ filtergram.

3.6 Calibration and comparison

The calibration of the video magnetograms were made by using the profile line slope method (Varsik 1995), which is based on the weak field approximation (Jefferies & Mickey 1991). The Stokes parameter V can be expressed as,

$$B_{\vert\vert}= k_{\rm cal}\,V$$ (2)

$$k_{\rm cal}=\left(- { {\rm d}I \over { {\rm d} \lambda}}\, \mu \, \lambda^2_0 \, g \right)^{-1}$$ (3)

where $\mu = e\,/\,4\,\pi$ mc²; e and m are the charge and mass of the electron, $\lambda_0$ the central wavelength of observation in Å, g the Landé factor of transition for absorption line, B_|| is the longitudinal magnetic field in Gauss and $k_{\rm cal}$is the calibration factor. I the Stokes intensity obtained in the absence of magnetic field and V is the intensity obtained by subtracting a single pair of video frames taken in opposite circular polarization. dI/d$\lambda$ is the measured line slope by tuning the filter from - 40 mÅ to -170 mÅ from the line center in the steps of 25 mÅ. The measured slope dI/d$\lambda$ is found to be 0.047, which gives the value of $k_{\rm cal} = 2059$ using Eq. (3). This constant is used in Eq. (2) for conversion of measured Stokes V intensity into the magnetic field values in Gauss. The measured longitudinal magnetic field is the observed average field over the spatial resolution element and limited by the solar seeing. Further as the spectral resolution of the FP etalon is small, the accuracy of calibration by this method is rather low.

Figures 5a and b show USO and SOHO/MDI video magnetograms obtained around the same time on 09 April 1997.

  

Figure 5: The magnetograms obtained for comparison on 09 April 1997 by: a) USO video magnetograph at 09:32 UT, b) SOHO/MDI at 09:41 UT

The SOHO/MDI magnetogram is a part of the calibrated full-disk data (Scherrer et al. 1995). It may be noticed that all the features observed in both the video magnetograms match well. After accurate registration of the two VMGs and degrading the spatial resolution of USO VMG to match the SOHO/MDI, a scatter diagram has been plotted and shown in Fig. 6.

  

Figure 6: Scatter plot made between calibrated USO and SOHO magnetograms

  

Figure 7: A typical: a) video magnetogram and b) CaI 6122 Å (-140 mÅ) image

  

Figure 8: Composite picture of magnetogram, CaI 6122 Å (-140 mÅ) and H$\alpha$ images showing one cycle multi-wavelength observation. The upper panel shows the images of active region NOAA#7843 taken on 19 February 1995 and the lower panel shows the same active region on 20 February 1995

This plot demonstrate the validity of the calibration procedure, except for a small deviation of the slope of the scatter line from unity. This deviation may be attributed to the use of different spectral lines and difference in spectral and spatial resolution of the two instruments.