It is about amateur's realization which, at first
(1990), consisted in making a prototype to test the possibility
to build a SHG. The presence of a microcomputer at home
as well as a linear CCD, and especially the recovery of a grating,
allowed this starting up. Thus the budget should be
minimum and the "D.I.Y.
used. The constraints were severe because the first version of
the SHG should be in "piggy-back" on
a classic 115/900 telescope supported by the most classic and frail equatorial
mounting on a wooden tripod. Useless to specify that the weight and the dimensions
of the spectro should be minimum too !
Where starting ?
The purpose is to make the smallest possible spectroheliograph
giving at the same tme a sufficient spectral resolution to make appear chromospheric
structures in Ha or K lines on the sun disc (about 0,05 nm), and a sufficient
spatial resolution so that these same structures are clear and detailed.
It is in fact the size and the number of photoelements of the CCD which
will guide, or determine, the continuation of the realization
Dispersion and spectral resolution
If 1 pixel of the CCD is about 10µm wide (perpendicular
direction to the axis of the bar), it is necessary that the dispersion of
the spectrum is such as this 10µm intercepts a spectral band of 0,05 nm
(We will say that the dispersion is 5 nm / mm). This dispersion is a function
of 3 elements: The number of lines by mm of the grating, the order of the
grating, and the focal length of the camera lens (the one that forms the
image of the spectrum). Dispersion increases if these parameters increase.
Dispersion is not all, it is also necessary to take
account spectral resolution. Nothing serves for having a strong
dispersion if the resolution is not good. The problem is the same
one as a telescope. Let us suppose an objective of 1,2 m focal
length and 100 mm in diameter, good quality. This objective is
able to give at focus plane an image of the Moon of 10mm in diameter.
If we reduce to 20 mm the aperture of the objective by means of
a diaphragm, the image of the Moon measures always 10 mm in diameter
but the resolution of details is 5 times less good.
For a spectral
resolution in connection with the dispersion, it is needed a wide
Matériel et réalisation
The elements which I had were the following ones: a linear
CCD Thomson TH7803 made up of 1728 photodiodes of 10x13µm in the step of 10 µm,
a plane reflection grating 1180 grooves per mm and 62 mm aside.
to the spectral resolution, R=73000, one reaches in theory 0,01 nm in the first
order so, no problem. The angle of diffraction for Ha line is 22 ° 48 ' and
to cover 13 µm with a spectral band of 0,05 nm, a 200 mm focal length camera
is sufficient - in theory !. It is there about a minimum value. But if one
wants to focus all the light from the grating, it is necessary to find an objective
of 62 x square-root( 2 ) - is 87 mm - in diameter with a F/D = 2,2. Let us
say at once that in practice, it is widely better to largely raise the focal
length, on one hand to minimize aberration of sphericity of the objective,
and on the other hand to minimize aberrations of the assembly monochromator
of which we shall speak again, and finally to minimize budgetary expenditure.
Spatial resolution :
Let us see now what should be the spatial resolution, i.e. the possibility of separating fine details in the image. The size of a pixel, in the axis of the bar CCD, is here 10 µm, and there are about 1700 pixels what allows to record a complete image of the Sun 17 mm in diameter. The spatial resolution in the plan of the image is connected to the angular resolution by the relation:
x = f . tg(α)
f = focal length of the objective of the telescope
α = angle of the incidental beam on the optical axis
x = Distance to the optical axis of the corresponding image point
Obtaining of a 17 mm solar image requires an objective of almost 2 m focal length. 1 pixel (10 µm) represents then a little more than 1 arc-second. An objective of 12 cm in diameter can - always in theory - allow this resolution but let us not forget that we observe the Sun and that the turbulence is generally limiting. One will thus gain nothing in resolution by using a larger objective.
Digital image :
It is now necessary to consider the digital aspect of the image. By using
to the maximum linear CCD (1700 pixels) and to synthetize a circular image,
it will be necessary "to cut the Sun" in 17000/13 = 1307 successive sections.
This represents 1700x1300 = 2,2 million pixels and, even digitized in 256
levels of grey (8 bits), it represents more than 2 MB of memory. In 1990,
microcomputers capable of treating this kind of image were still extremely
expensive. I had then an ATARI 1040 without hard disk and allowing to display
320*200 pixels images with only 16 colours. The first limitation was the
support of recording: 720 KB floppy disk; the second was the impossibility
to display "big" images.
I thus decided to limit the size of the images
to 640x400 pixels by making the reading of 640 lines of 800 pixels and
by taking only one pixel out of two in each line.
These elements led me to use the focal image of
8 mm supplied by the telescope 115/900 and to acquire 2 spherical mirrors
76/700 to make the monochromateur. Ratios F/D are close and grating is
practically totally enlightened.
The figure opposite shows an example
of image in Hα light obtained with
this preliminary assembly. Filaments, plages, prominences are visible.
Spectral resolution is that expected but spatial resolution is not very
good because all the adjustements are manual and very delicate to realize
without making vibrate the instrument.
After this first experiment and an unproductive period
of a few years, I consider again the construction of a SHG because
certain favorable circumstances were gathered. I built an equatorial
mounting motorized in right ascension and in declination for my 115/900
telescope. Intended originally for the photography of the deep sky
with a 500 mm telephoto lens in parallel with 115/900, the stability,
robustness and convenience are without comparison with the first mounting.
The recovery of a new grating - although far from the new state - blazed
500 nm and 1440 grooves per mm, the diversion of an other linear CCD from
a hand-scanner and the purchase of a microcomputer revived my interest for
Here is the description of this instrument that is however only an example and each can modify and improve it.
Optical components :
- 1 telescope D=115 mm, F=900 mm, helical focuser, forming the primary image
of the Sun
- 1 divergent meniscus F=-300 mm "to take out" the image plane of the telescope
and obtain 1170 mm resultant focal length. (usage similar to Barlow's lens)
- 2 spherical mirrors D=76 mm, F=700 mm used as collimator and camera (monochromateur)
- plane reflexion grating, 62 x 62 mm, 1440 tr/mm, blazed 500 nm
- 3 small plane mirrors
Particular mechanical elements :
- 1 metal slit (which can be also considered as optical component)
- 1 stepper motor for the rotation of the grating
- 1 micro-motor + reducer for the orientation of the CCD
- 1 system to focus the CCD
The electronic elements :
- 1 linear CCD made up of 2048 pixels of 14x14 µm, step 14 µm
- 1 clock-generator circuit for driving the CCD
- 1 amplifier and digitizer circuit for the video signal
- 1 interface with the computer
- hand controler for grating and CCD motors
The microcomputing :
- 1 microcomputer (Atari 1040 until 2000, and PC 233 MHz from 2001 for the acquisition)
- A software for CCD control and image acquisition
The optical layout
As a drawing is better that 10000 words, here is the plan of the arrangement of the optical components constituting the SHG. Telescope is in the bottom.
Global aspect of the spectroheliograph