Coronagraphy
Principle
The coronagraph (coronOgraphe in french) is an instrument that allows to observe solar eruptions without having to use standard ultra
narrow band pass etalons. Observing the chromosphere is challenging because it is 100k times fainter than the photosphere.
-
The "easy" (this means easy to have from commercial productions) way to get rid of the most of the photospheric light is to use filter that allows only an very narrow (<1A) spectral
window of transmission, centered on H-alpha or any other emission or absorption line of interest. The problem is that the manufacturing of these filters is complex and strongly dependent on environment conditions (temperature and air pressure in particular).
Such filter (or rather assembly of filters) can be put behind (or in front of) refractors, reflectors, that does not matter as soon as the light transmitted to the filter hits the filter perpendicularly (or close to).
- The second method originates from 1931, when Bernard Lyot invented a way to observe the sun with large FWHM filters (ultra narrow band etalons didn't exist anyway)
based on a very simple optic arrangment. To deal with photosphere with "standard" filters you need to first evacuate the light coming directly from the photospher, i.e. the image of the disk.
To do that, an occulting cone in steel, or other very hard and shiny metal, is placed precisely at the focal plane where the image of the sun is formed.
This cone litterally evacuate 99.99... percent of sun total emission getting to the instrument. That is the easy part. The problem is that this is not enough, there are remaining parasitic
light not eliminated by the cone. The main source is the diffracted light from the edge of the front lens. This has nothing to do with the substrate of the lensbut only of
the presence of an edge (and that is unavoidable...). The second source are light reflection inside the front lens. This parasitic light has to be eliminated
before it reaches the sensor, and that is the key part that Bernard Lyot managed to set up: a diaphragm for the first source, and a occulting spot for the second.
Why using a coronagraph?
This is a very valid question, now that the price of H-alpha or CaK etalons have dropped considerably in the last 10 years. These etalons
coupled with the appropriate instrument and blocking filters allow to
study both the sun surface and eruptions on the edge of the disk. So why keep using these very long and potentially delicate-to-use
instruments that only allow to observe the eruptions on the edge of the sun??
The answer relates with the band width of the filtering to be used with the coronagraph.
Back in the time there were attempts with standard colored filters, but that is very dangerous since a huge fraction of the sun light can get to the eye or sensor, unprotected.
Now the spectral filtering on coronagraphs typically require an FWHM between 10A and 5A, which is 10 to 20 times larger than the
FWHM of most mica or air-etalons (cfr Lunt, Coronado, Heliostar, Daystar, Solarspectrum,...)
Such large filtering would be useless on standard instruments because of the diffusion in optics and effect of diffraction from the aperture stop,
while the coronagraph is elaborated precisely to allow these large FWHM thanks to a clever optical design.
This means two things:
- the coronagraph is able to see much fainter solar eruptions
- the coronagraph is less sensitive to Doppler-Fizeau effect affecting the eruptions with large radial speed. Let's calculate this: A 0.5A
FWHM at 6563A means a speed windows of ~0.5/6563*c=22 km/s, with c the speed of light. A 10A FWHM means ~440 km/s.
Of course mean radial speed of eruption at the edge of the sun is close to zero, but there are cases (magnetic loops for instance) where the
speed can be large, blue- or redshifting the emission line outside of narrow filter spectral acceptances.
- The visual observation in a coronagraph is simply amazing
Here is a schematic view of how it works:
- The first lens O1 creates an image i1 a the level of an occulting cone. This image is situated between the lens L1 and its focus.
- The image i2 of i1 through L1 is therefore virtual, located a bit frontward from the occulting cone.
- The lens L1 creates at image of the front lens O1 at the level of the diaphragm. This image contains the diffusion and diffraction
patterns from the front lens, which must be removed since this is one of the main contrast killer potentially ruining the vision of the
solar eruptions
- Finally the lenses L2 et L3 form a convergent system that will create an image i3 of i2. That is what is observed either visually or
through a sensor.
There is also the possibility to evacuate parasitic light from multiple reflections inside O1, but this is rarely seen on amateur coronagraph
and therefore won't be discussed here.
If the big picture is relatively simple to understand, there are some hidden subtelties required to be taken into account to have a
good-performance instrument. The key point to keep in mind is that
the coronagraph is an instrument that works excusively off-axis: the disk is occulted, so the sun is observed at a minimum angular radius of
0.25 degrees from the optical axis. This means off-axis aberrations should
be considered with care in order to not ruin the resolution. This is done mainly by aspherizing the two first lenses, in particular the second
one (called field-lens). While the ashperization (=giving a conic k-factor different from 0)
on the front lens is generally small, it is quite severe on the field-lens. This is probably the most difficult part to deal with for the optics
maker.
The Valmeca coronograph 150 f15
The french company Valmeca has produced several models of coronagraph, in particular a 90mm and big one that has a front lens of 160 mm
stopped at 150 mm. The big one is the subject of this webpage.
Needless to say that such an instrument is long, very long: more than 3-meter. But it is luckily relatively light, weighting only 12 kg
without the rings.
I bought this instrument in second-hand in 2016, the tube had been destroyed, the occulting cones (different diameters for different sun
apparrent size across the year) showing also alteration (scratchs, bumps,...).
I contacted Valmeca to get a new tube, just in time to do some tests in high altitude at the observatory of St-Véran.
The rear part consists in
- L1 and its holder, whose position latitude is about 5 cm.
- A second part that contains the diaphragm, L2, the filter and L3. The distance between the diaph and L2 can be adjusted between ~10 and
30 mm
I quickly realize that there was something strange: it was not possible to focus simultaneously the image of O1 on the diaphragm and to adjust
the position of the L2 and L3 so that the front focus point would coincide with i2.
Because of many good and bad reasons the instrument remained unused until now, when I decided to tackle this problem of focus. I did measure
the focal length of every optics part, and in particular of L1, L2 and L3.
- F(O1): ~2350 mm.
- F(L1): ~300 mm.
- F(L2)=F(L3)~300 mm.
- distance D between L1 and the occulting cone (front surface): 48mm
It turns out that this configuration is not optimal at all: in these conditions the distance between the cone and L2 should be F(L2)+D~240 mm,
and at the same time the distance between L1 and the diaphragm should be close to 340 mm. This is not possible since
the diaphragm is located between L1 and L2. I therefore had to change L2 and L3 to FL=400 mm doublets in order to approach an acceptable
configuration.
Assuming a conic constant of -0.7 for O1 and -30 (!) for L1, the off-axis performance are excellent, with a diffraction limited area that goes
way beyond sun radius.
I cannot confirm these are the conic constants of O1 and L1 in my coronagraph, but I know this particular coronagraph
was corrected by Dany Cardoen many years ago before this instrument was disamssembled by a previous owner, resulting
in a much better resolution than initially.
Filter
A key element of the coronagraph is the filtering of incoming light.
This filtering is required to enhance contrast between solar eruptions and the sky backgroound, but also to some extent protect the eye or
the sensor in case the telescope is not totally occulted by the sun (bad tracking, wind,...).
As said above there is no need for an etalon, typical FWHM are between 5 and 10A, i.e. 10 times larger than typical etalon's FWHM. That
means also these filter are generally cheaper, less sentitive to environment conditions, etc...
Depending on the observation, the filters can have their bandpass centered on signatures of H-alpha, sodium, ...
In this case I'm focusing -at least for now- on H-alpha, that is by far the easiest case scneario thanks to the high brighness of the
eruptions at this wavelength.
For the first test I use an Alluxa 656.3 filter with 10A FWHM.
As we can see on the graph above, tha maximum transmission is close to 100%. The slopes of the transmission curve are steep, and the 10 percent transmission width is only 14A.
There are several other on-the-shelf commercial solutions for H-alpha filtering with FWHM between 5A and 10A. As of today (Sep 2025), the most
interesting ones are:
- Antlia 5A with diameter 12mm, transmission~ 85% Price: $400. This one needs to be completed by an extra UV/IR filter such as Baader
20nm bandpass (~90 eur) and perhaps a BelOptik UV-IR/KG3 combo for visual observation.
- Alluxa 656.3 5A diameter 25mm $3800. Needs to be completed by a BelOptik UV-IR/KG3 combo for visual observation.
- Edmund Optics 5A diameter 25mm, transmission>45%, Price: $1750
- Edmund Optics 2A diameter 25mm, transmission>45%, Price: $1500
First tests with new configuration (Sep 2025)
The length of the Valmeca 150 f15 coronagraph requires a very sturdy mount to ensure a minimal stability against wind of manipulation at the
level of the focuser. I first tried with a takahashi EM-400: it is too weak! This has nothing to do with the maximum acceptable load of the
mount (38kg), but rather to the torque resistance. Here below is the coro on the takahashi EM400. The rings (160mm inner diameter) and dovetail were made by
AOK Swiss.
Using an old AP1200-DA (yes, pre-QMD era!), the tube is way more stable, although my impression is that mounts of the category of AP1600,
GM3000 would be more appropriate. Well in fact, keeping in mind that no goto or fancy electronics is needed, many very old-and-cheap mounts capable
of holding 100 or 150 kg would work: VMA300, Alt7-ADN, SECIA, Mathis, Byers,....
Sep 15th 2025
The first "light" was done in difficult condition, with a lot of wind although and excellent sky transparency (both are often correlated).
The scope is mounted on the AP1200-DA
The focusing consists first at placing the occulting cone exactly a the focal plane of the front lens in H-alpha
(because of the chromatism of the front lens the focal length strongly depends on the wavelength).On this model of coronagraph the procedure must be done iteratively
and as precisely as possible (up to 1 or 2 mm of precision).
The next step is way easier: place the Lyot diaphragm where the image of the front lens through L1 is built. That takes 5 seconds, you just need to
get that image sharp.
Here below is the very first image from the coro with its new transfer doublets. The camera is a Player One Uranus-M, with an exposure of 6ms and a gain close to zero :
Back to homepage