Радиоастрон

[Sellin]

Stalky пишет:

мастер_лукьянов пишет:

Alexandr_A пишет:

Вал пишет:
Вот и радиотелескопами так - один пиксель. Но очень чувствительный.

Может всетаки у одиночного радиотелескопа есть что-то вроде фазированной антенной решетки в точке фокуса заркала, тогда пикселей может быть сотни.

Как правило, приемник один. Но и с одним можно создать изображение, сканируя область

Вот совершенно правильное замечание...

Насколько я понимаю, одна точка получается у старых интерферометров без компьютерной обработки, ито с нюансами. Современные интерферометры синхронно оцифровывают принимаемый сигнал, затем, два полученных сигнала математически "сводятся", и в случае успеха(который зависит от точности временных меток, и пространственного положения), сдвигая фазу получают сканирование по типу ФАР, т.е. Сразу получают Линию изображения.

[Виктор Петров]

Спасибо за ответы.

Тогда еще один вопрос. Зачем такой неинтересный диапазон электромагнитного излучения выбрали, как радиоволны для космического интерферометра? В нем же только очень горячие объекты или объекты с мощным магнитным полем в основном излучают, как черные дыры и пульсары.

Почему не взяли оптический диапазон, чтобы к примеру сфотографировать у близких звезд внесолнечные планеты?

[testest]

Тут как то писали, что, если говорить условно, для создания картинки необходим интерферометр из трех телескопов. С двумя телескопами картинка вырождается в график.

Иванов Петр пишет:
Тогда еще один вопрос. Зачем такой неинтересный диапазон электромагнитного излучения выбрали, как радиоволны для космического интерферометра? В нем же только очень горячие объекты или объекты с мощным магнитным полем в основном излучают, как черные дыры и пульсары.

Почему не взяли оптический диапазон, чтобы к примеру сфотографировать у близких звезд внесолнечные планеты?

Потому что в других диапазонах интерферометрия намного сложнее а создать оптический интерферометр со сверхдлинной базой вообще нет шансов. По-моему.

[Salo]

http://forum.nasaspaceflight.com/index.php?topic=27335.msg1068881#msg1068881

websquid пишет:

About 3 months ago I made an interview with one of the project scientists for the german website raumfahrer.net. We talked about the science as well as the project history (especially since the launch, how they got the system to work). If you are interested I could post an english version here. Please some feedback: Do you want to read all this in english?

German ver sion (3 parts):
http://www.raumfahrer.net/news/astronomie/17042013190758.shtml
http://www.raumfahrer.net/news/astronomie/18042013224629.shtml
http://www.raumfahrer.net/news/astronomie/20042013013103.shtml

http://forum.nasaspaceflight.com/index.php?topic=27335.msg1069157#msg1069157

websquid пишет:

OK then I prepare it for this forum. Shouldn't last too long    

a_langwich пишет:
Is there just one spacecraft? How much would it cost I wonder to launch an identical one in a mostly opposing orbit?

Yes, there is only one satellite. A second similar satellite would cost about 200+ million $ I think, but this is only a personal estimate.

But there will be no second satellite for the RadioAstron as it looks today, the Astro Space Center is already working on the next project Spektr-M/Millimetron. This will be a combined far-infrared/millimeter-wavelength telescope located at L2. So it is in a way a bigger Herschel Telescope (the FIR-instruments will be based on Herschels) and in a way a RadioAstron successor working in smaller wavelengths (0.3mm until 1.8cm ) - this means, the longest Millimetron wavelengths is compatible with the shortest RadioAstron wavelengths. IF the Spektr-R works long enough and IF Spektr-M is in orbit fast enough - launch is currently estimated somewhere about 2020 - interferometry with two satellites would be possible.

Another project using two satellites is currently under development by China. But I only know that this project exists, no details.

[Salo]

http://forum.nasaspaceflight.com/index.php?action=post;quote=1069161;topic=27335.30;num_replies=43;sesc=c134542f9636141fd110e65650ca710e

websquid пишет:

Here it is. Sorry for the long post but you asked for it

RadioAstron - The Interview

Raumfahrer.Net: At first I would like you to tell us a bit about your work. In which field do you work and how did you come to RadioAstron?

Kirill Sokolovsky: In fact I have to jobs. My main job is the work at the ASC LPI - here I work for the RadioAstron project. I am also a part-time software developer at the Sternberg-Astronomy-Insitute of the Moscow State University. Here I work on variable stars. This work has no connection to RadioAstron.

I was a student at the Moscow University and then I started working at ASC to get my diploma (the work was about VLBI - Very Long Baseline Interferometry). At that time they searched people for VLBI, there was the perspective that one day RadioAstron would fly.

RN: At which time did the ASC search people?

KS: That is some time ago. I began 2006 to work there.

RN: That is seven years ago. Did you work on RadioAstron at that time?

KS: I did not work on this project then. I worked with data of earth-based VLBI observations. When I got my diploma, I went to Bonn(Germany) and started working at the Max-Planck-Institute for Radio Astronomy. (MPIfR) to get my Ph.D. I continued to work on earth-based VLBI. Really incorporated in RadioAstron were I since summer 2011, when I came back to Moscow. Before the satellite started, RadioAstron was mainly work for engineers. As an astronomer, I could not really help until that time. After the launch things got much more interesting for me!

RN: After the launch it lasted some months, until scientific observations started. How did you and the other astronomers at ASC feel, when you got the first useful data?

KS: I would say, we were surprised. And very excited, of course! The turning point, when we realized, that this experiment was a success, was some time after the observation on November 14 in 2011 - the first VLBI test. The 100m-radio telescope in Effelsberg, the three russian 32m-telescopes Svetloe, Zelenchukskaya and Badary and the ukrainian 70m-telescop in Evpatoria did participate in this test.

At first, we did not have working software to do the data reduction. We realized later, that the ASC-Correlator software contained a bug, which can only be discovered when working with moving antennas (like the Spectr-R Satellite in its orbit). In principle we created two simplify correlator softwares fr om scratch, to detect the interference fringes and to find the bug in the "big" correlator.

Later James Anderson at MPIfR developed for the DiFX software correlator the ability, to work with RadioAstron data as well. DiFX is the de-facto standard for earth-based VLBI data processing. But when we made the first tests, this software was not available.

Also there is to say, that there was a smooth transition fr om technical tests to scientific observations. The system was in a constant debugging process until summer 2012. Only then we were convinced, that the whole systems works correct and that we can now collect good scientific data.

RN: You mentioned the participation of the MPIfR. What do you think about this and how important is this cooperation?

KS: The cooperation with MPIfR was crucial for the project. At first, they granted us observation time in Effelsberg for RadioAstron tests. The Effelsberg telescope proved itself as one of the most useful telescopes during the tests and later the Early Science Program (ESP, running fr om february 2012 until june 2013). It is very sensitive, fully steerable and is often used for VLBI. That is why the people there watch every detail necessary for interferometric observations and calibration. And it is located on the same continent as the tracking station in Pushchino near Moscow. Therefore there are long common timeframes for observations with the satellite and the ground telescope.

In addition the MPIfR (especially its director Anton Zensus) supported data analysis and data logistics. The main point was the implementation of RadioAstron-support in DiFX. This was a big deal. It became the most important software for data analysis, at least in our AGN-group (active galactic nuclei) group.

RN: Is DiFX only used in Bonn or also at the ASC?

KS: Most of the "DiFXing" for the RadioAstron AGN survey is done in Moscow. Bonn correlates the imaging experiments, which were performed together with the European VLBI Network (EVN). These produce large amounts of data (around 20 TB per observation) and the MPIfR correlator can handle this easily, which is not really true for us in Moscow.

We do not have DiFX on our supercomputer at ASC (it uses the ASC-made software) and the machines which run DiFX do not have enough power to correlate a complete imaging experiment.

At the moment there is work on a detailed comparison between the results of both correlators. But to make this reasonable, all datasets have to be processed with both softwares.

So far there were three imaging experiments together with the EVN. The first was in March 2012 and is completely processed. The processing of the other two is still under way (the observations were in october 2012 and march 2013).

RN: The first image was published. Do you plan to publish the produced images?

KS: For sure yes!

But at the moment we do not have correlated datasets for the other two experiments and when we have these, it can still last one or two months until the final images are created. That is how things go with VLBI. The data processing is much more complicated and "delicate" as in other fields of astronomy, which I did work on.

But I should add that most of the observations are not done in imaging mode. The aim of such observations is just the detection of interferometric fringes. The amplitude of the fringes can be measured for a range of baselines and this can be compared with a simple model of the source. For example we can assume a gaussian model of the source and measure its size.

This was the way earth-based VLBI was done in the time when there were only few telescopes which could do such observations. The problem with image reduction is that you need to cover the so called uv-plane as good as possible. You need many telescopes with different baselines.

When RadioAstron is far from earth we have the situation that all ground telescopes are in fact at the same place - so no imaging is possible. Instead we almost have an interferometer with only two elements. This means we only can create images, when the satellite is not more than 3-5 earth diameters away. At this baseline we do not have the highest resolution, but we can do imaging. At longer baselines we can only compare the measured visible amplitude with a simple model (for example a gaussian). In this case we only can determine the size and brightness of the source. In addition we can determine roughly the shape (is the object stretched or not?). The more data points we get, the more complex models can be used to describe the object.

RN: You mentioned size and brightness, which can be measured with RadioAstron. Often there is talked about "brightness temperature". What does this mean and how can you measure it?

KS: Brightness temperature is a certain way of radio astronomy to talk about surface brightness, something that is determined in optical astronomy as magnitude per area. In the easiest case brightness temperature is the brightness of the object divided by its size. In addition there is a factor to convert this value in Kelvin.

There is some freedom in the way you determine the "size" of a source when the real size can not be determined. Instead you have a measured interferometric visibility which you can compare with a model. So a spherical model had a slightly different size than a gaussian model for the same source. But in general the difference is for all possible models only a small factor. 10^14K can always be distinguished from 10^11K without problems, no matter which model is used.

RN: What do this different models mean?

KS: The gaussian model is simply a radio source, which brightness profile follows a gaussian curve. There are two variants of this model: circular and elliptical. The gaussian model was simply chosen because it is simple to calculate with it: The Fourier transform of a gaussian curve is another gaussian curve. And the visibility measured by the interferometer is the Fourier transform of the brightness distribution of the source.

A spherical radio source has a slightly different brightness profile, it follows the surface of a sphere. In both cases this is no physical model of the source, but only a "toy model" which allows us to characterize size and flux density of a source. A source wh ere we see almost no details, but wh ere we already can see, that it is no point source - but a slightly increased point source (a typical situation for interferometry).

RN: How does the ability to measure brightness temperature depend on the baseline?

KS: You need many observations with different baselines. When you consider the low sensitivity of the space radio telescope, a simple model (a point with a certain size) seems to fit for all the baselines, even for the extreme angular resolution RadioAstron can achieve.

From a physical point of view, this bright spot (when observing AGN) is the brightest part of the relativistic jet, wh ere the jet becomes transparent for its own synchrotron radiation. This area is called the VLBI-core, the brightest spot in a many lightyears long jet. But there are concurring ideas, what a VLBI-core really is. In principle there could be other options, not only these transparency-thing.

RN: What are these alternate models?

KS: A reasonable alternative is that a VLBI-core is a collimation shock in the jet. In this model the core would be a certain physical region, not a place were the jet becomes transparent for the observed frequency. There are some more models, but these do not seem to work for radio-loud AGN.

A way to test the "synchrotron transparency" is to determine the size of a VLBI core at different wavelengths. The transparency model predicts that the VLBI core should be bigger for smaller frequencies. The reason is that the jet has an opening angle, you could speak of the jet as a cone. For smaller frequencies the jet becomes transparent in a larger distance to its base.

There were ground based observations on this question, but with RadioAstron we can measure this effect with much higher accuracy. The collimation shock theory in contrast does not predict any difference in size and position at different wavelengths.

RN: Are there already hints which model fits better?

KS: At the moment - no. We are still under way to collect more data.

RN: Do you believe that you can unveil the real nature of VLBI cores?

KS: Yes. We only have one real size measurement - the core of the blazer 0716+714 from the imaging experiment with the EVN. We need more time to determine this for other sources based on the fringe survey. For some sources (for example for BL Lac) we should have enough data, but we need more time to process it. I am very sure that we will solve this question with RadioAstron finally.

RN: Another thing is the interstellar matter (ISM). It was expected that this would scatter radio waves but it seems like this does not happen in the expected way?

KS: Yes! The interstellar scattering is a big thing for RadioAstron. I am not an expiret for this, but as I hear from my colleagues the difference between expected and observed scattering is very huge. The scattering is there but very smaller than expected. It seems in general that the scattering is not as homogenous as expected. This is a reason why it is in total smaller than thought before.

In fact was the expectation of strong scattering a large point of objection before RadioAstron started. There were many very smart people who thought that RadioAstron could not see anything beyond 1-2 earth diameters baseline. With the current orbit we reach up to 25 earth diameters. This means the resolution is convenient for everything one can dream of. The question is - what are the smallest details which we can see despite the interstellar scattering? We already can say that we see much more than some experts expected.

RN: When did you realize this behavior of the ISM?

KS: We are the first who measured this directly with an interferometer with baselines larger than earth! For AGN there were early hints, based on the intra-day variability of sources (fast changes of the flux density caused by scattering) that there size is small. But these estimates are even more model-dependent than the size measurement with VLBI.

RN: So the scattering itself changes in hours?

KS: Yes! And this is something which we want to study in the next years in detail. A large point of the AGN survey program for the next year is the combination of RadioAstron observations with earth-based measurements of daily flux variation to study the scattering.

RN: This sounds a bit like this behavior is comparable with our atmosphere in visible light - like stars blink. Is this comparison legit?

KS: Yes, this analogy seems to be fully correct. And this immediately leads to the idea, wether it is possible to create something like an optical speckle interferometer for radio waves?! I can not answer this question. Maybe it is possible, but I do not know enough about this topic to say it for sure.

RN: So we have to wait for more information. You mentioned the following years. At the moment the Early Science Program is running, which will end soon. In july the Key Science Program starts. The ESP had a strong focus on AGN. Will this be the same in KSP or will other topics get a bigger fraction of observation time?

KS: Emanating from the accepted KSP proposals the most observation time will still be used for AGN. Maser and pulsars will also be observed, but I think their fraction will roughly stay the same. I have an accepted project to study transient radio sources like supernovae or tidal disruption events, but this is an experimental project which was never tried before with space-VLBI. So it is not clear whether something will come out of it. But I think it is worth a try.

RN: Which countries made the most KSP proposals? Russia is for sure leading, but in which other countries is the largest interest in RadioAstron?

KS: Add first I would like to add something to the AGN observations: There will be more imaging experiments for sure. Not just three in one year. Concerning the origin of the proposals I think that most of them, if not all, come from international groups consisting of scientists from many different countries as co-authors. I know Germany, the USA, Spain, Australia, Poland and Japan and for sure I am forgetting some. Sadly there is not yet official information about the accepted proposals. The RadioAstron Program Evaluation Committee made its decision but some of the proposals depend also on the evaluation from the project committees of single ground telescopes or networks like EVN or VLBA, so there is no final list of KSP projects. It can happen that some of the projects accepted for RadioAstron are rejected by ground telescopes so they can not be carried out.

RN: Can you estimate how many imaging experiments will be performed in the first KSP period (july 2013 - june 2014)?

KS: The imaging experiments depend the most of the availability of ground telescopes. A chance for these experiments is at each perigee passage of Spektr-R, roughly every 8,5 days. This is under the assumption that the tracking station in Green Bank is then available. It is hard to say how many observations really will be performed. I expect at least ten imaging experiments, maybe more. The KSP observation program is not yet fixed. I think there will be no imaging experiments until september because of severe visibility restrictions in the summer. So for the first two months of the KSP there will only be simple fringe detections for AGN, pulsars and masers, like the most time during the ESP.

RN: When will the tracking stations in Green Bank and South Africa start their work?

KS: The tracking station in Green Bank will start soon. Under consideration of possible delays our current best estimate is that it will be available in september. All papers between USA and Russia are signed, the electronic equipment is currently tested in Pushchino. A potential delay risk is the russian toll, but the work is under way. The control system of the 43m-telescope in Green Bank was tested, so I think it is ready. After the data recording equipment is delivered there will be of course some time needed to install and test it. So the most optimistic guess is july, the more realistic is september.

The south african station seems to be still on the level of official talks and agreements. There is currently no technical work done for it. So I think the south african station needs at least a year, maybe longer, until it works.

Kirill Sokolovsky in the control room of the Effelsberg radio telescope

[Salo]

http://forum.nasaspaceflight.com/index.php?topic=27335.msg1069205#msg1069205

websquid  пишет:

 I hope that this mission will get some broader coverage when results are released. A paper published in Nature or Science would for sure give a little boost. At the moment there are some papers under preparation about the results of the ESP, so maybe then the Reuters article (or whoever similar) may come...

Talking about the Green Bank Tracking Station, here some news:
The equipment was shipped some days ago to Green Bank. If all goes well it can be installed until july 14. Late july and august are planned for tests, the station is expected to be fully operational in september. Then the  observation time for RadioAstron will roughly double.

The 43m-telescope in Green Bank - the new tracking station

[raputor]

Иванов Петр пишет:
Тогда еще один вопрос. Зачем такой неинтересный диапазон электромагнитного излучения выбрали, как радиоволны для космического интерферометра?

Как раз-таки, выбраны самые интересные и оптимальные диапазоны для наблюдения того, что наблюдается.

Иванов Петр пишет:
Почему не взяли оптический диапазон, чтобы к примеру сфотографировать у близких звезд внесолнечные планеты?

Радиоприёмников оптического диапазона (в классическом понимании Радио) не существует в природе.
Для создания "фоток". Нужно делать оптический телескоп и заставить его работать совместно с наземным оптическим, в режиме интерферометра чтобы увеличить разрешение. Задачи по поиску экзопланет совершенно другие. Другой телескоп, другая аппаратура. У Радиоастрона же задачи лежат совсем в другой плоскости.
Здесь ведь нет понятия "крутизны" инструмента или "интересности" диапазона. Каждый инструмент заточен под решение своей задачи.
Например, Вы не сможете "увидеть" в оптическом диапазоне линии излучения водорода. Точно так же, воспользовавшись оптикой, нельзя получить необходимую информацию об изучаемых Радиоастроном объектах. Поэтому для каждого класса задач существуют свои инструменты. Для отрезания проволоки - кусачки, а для забивания гвоздей - молоток.
В данном случае, ответы на фундаментальные вопросы физики наблюдаемых Радиоастроном объектов не менее (а возможно и более) важны, чем вопросы составления списков экзопланет и их фотографирования.

[dmdimon]

Иванов Петр пишет:
Почему не взяли оптический диапазон?

потому что нужно обеспечивать когерентность измерений. На наземном VLTI в оптическом дивпазоне с помощью высокоточных линий задержки удается обеспечивать когерентность на время единиц минут. База там около 200 метров, все неподвижно (во время измерений) и промерено, ЛЗ подогнаны и отстраиваются. Радиоастрон летает на расстоянии сотен тысяч километров на скорости в километры в секунду. Для обеспечения когерентности нужно контролировать фазу с точностью заведомо лучше четверти длины волны, т.е. иметь синхронизацию меток времени лучше четверти периода колебаний волны плюс знать дистанцию опять-же с точностью выше четверти длины волны сигнала. Это конечно упрощенно. вот и прикиньте - реально ли это для оптического диапазона.

[G.K.]

raputor пишет:
Например, Вы не сможете "увидеть" в оптическом диапазоне линии излучения водорода

Аш-альфа. Даже глазом.

[Stalky]

dmdimon пишет:

Иванов Петр пишет:
Почему не взяли оптический диапазон?

потому что нужно обеспечивать когерентность измерений. На наземном VLTI в оптическом дивпазоне с помощью высокоточных линий задержки удается обеспечивать когерентность на время единиц минут. База там около 200 метров, все неподвижно (во время измерений) и промерено, ЛЗ подогнаны и отстраиваются. Радиоастрон летает на расстоянии сотен тысяч километров на скорости в километры в секунду. Для обеспечения когерентности нужно контролировать фазу с точностью заведомо лучше четверти длины волны, т.е. иметь синхронизацию меток времени лучше четверти периода колебаний волны плюс знать дистанцию опять-же с точностью выше четверти длины волны сигнала. Это конечно упрощенно . вот и прикиньте - реально ли это для оптического диапазона.

Всё ешё хуже. Для того, чтобы выделить и сохранить фазовую информацию (без которой невозможно синтезировать картинку) существует только один способ - принимаемое световое поле от обоих объективов интерферометра необходимо непосредственно подать на общий приёмник...НЕПОСРЕДСТВЕННО...чем в Южной обсерватории и занимаются со всем сопутствующим геморроем...

PS Интерферометр интенсивностей обсуждать отказываюсь...  :evil:

PSII Но для тех  правильных мальчиков и девочек, кого орешек знаний не страшит и масштабы не пугают,  добро пожаловать на страничку Антуана Лабейри на сайте Laboratory of Stellar and Exoplanetary Interferometry:
http://www.oamp.fr/lise/Antoinesspecialpage.html

[Stalky]

Коллега Salo, любезно продолжает нас приучать к английскому языку. 

Для ленивых - оборудование для организации ППИ в Грин-Бенк поставлено несколько дней назад и может быть смонтировано к 14 июля, ожидаемый срок готовности ППИ к работе - начало сентября, переговоры о ППИ в Южной Африке не вышли из стадии трёпа малоответственных лиц...

В течении первого года основной научной программы ожидается примерно до десятка сеансов работы наземно-космического РСДБ по результатам которых возможно получение столь ожидаемых заинтересованной общественностью  детальных изображений не пойми чего, основной объём работ не связан с получением изображений,  происходит в режиме получения интерферограмм в узком диапазоне базовых линий и использовании результатов для проверки различных моделей компактных источников радиоизлучений...настроение у всех бодрое, так как удалось преодолеть проблемы с ПО корреляторов и фишка попёрла...

[cross-track]

dmdimon пишет:
Для обеспечения когерентности нужно контролировать фазу с точностью заведомо лучше четверти длины волны, т.е. иметь синхронизацию меток времени лучше четверти периода колебаний волны плюс знать дистанцию опять-же с точностью выше четверти длины волны сигнала.

Получается, что дистанцию до Радиоастрона меряют с точностью не хуже 1.5 см? Охренеть!..

[ZOOR]

Меряют ЕМНИП десятки см (где-то в теме пару раз было).
А потом сдвигают и ищут корреляцию.
Чем лучше намеришь и напрогнозируешь - тем легче найти корелляцию и достовернее результаты.
Посему и за лазерную локацию бьются

[cross-track]

ZOOR пишет:
Меряют ЕМНИП десятки см (где-то в теме пару раз было).
А потом сдвигают и ищут корреляцию.
Чем лучше намеришь и напрогнозируешь - тем легче найти корелляцию и достовернее результаты.
Посему и за лазерную локацию бьются

В смысле прогнозируют орбиту и находят наилучшее приближение? Кальманом фильтруют?

[ZOOR]

Лучше kirxу на АФ такие вопросы задавать. Вон Сергей какое бодрое интервью с ним запостил

[G.K.]

Stalky пишет:
PS Интерферометр интенсивностей обсуждать отказываюсь... :evil:  

     Что-то мне этот топик начинает напоминать

[G.K.]

Stalky пишет:
чем в Южной обсерватории и занимаются со всем сопутствующим геморроем...

И как у них успехи?

[raputor]

G.K. пишет:
Аш-альфа. Даже глазом.

Раз так, предлагаю запустить Вас в Космос и ввести в режим интерферометра
Будете изучать одну из спектральных линий.

[Stalky]

cross-track пишет:

ZOOR пишет:
Меряют ЕМНИП десятки см (где-то в теме пару раз было).
А потом сдвигают и ищут корреляцию.
Чем лучше намеришь и напрогнозируешь - тем легче найти корелляцию и достовернее результаты.
Посему и за лазерную локацию бьются

В смысле прогнозируют орбиту и находят наилучшее приближение? Кальманом фильтруют?

Спросите у phobos24 - это его законная кафедра. http://www.kiam1.rssi.ru/~den/index.html

[Stalky]

G.K. пишет:

Stalky пишет:
чем в Южной обсерватории и занимаются со всем сопутствующим геморроем...

И как у них успехи?

Успешные, зайдите на их сайт. Есть его версия (неполная) на русском языке http://www.eso.org/public/russia/news/eso1227/
 но в конечном итоге всё равно  следует перейти на англоязычную страницу  http://www.eso.org/public/

и там розыскать страничку The Very Large Telescope Interferometer (VLTI) http://www.eso.org/sci/facilities/paranal/telescopes/vlti/index.html

и там найти всякое разное описание, например, The VLTI conceptual design http://www.eso.org/sci/facilities/paranal/telescopes/vlti/tuto/tutorial_bko_VLTI_Concept_and_technical_aspects.pdf
в котором очень наглядно представлено, каков этот монстр...