Peter Buschkamp: Shooting lasers into space – for science!

preroll music

Herald: Our next speaker has studied in Bielefeld,

and he studied... laughterclapping

what he did is: He studied laser physics.

And now he is working at the Max Planck Institute
for extraterrestrial physics.

And today he will explain you

how it is possible to use laser light

to enhance distorted images

that were take from the earth

of stars and galaxies and nebulars.

So I want to hear a

really loud and warm applaus

for Peter Buschkamp with

"Shooting lasers into space -

For science"! applause

All right! Thank you for the nice introduction

Thank you, for coming here

this evening.

I'm very excited

to speak at the conference.

Finally I find a talk

where I can contribute

after all those years.

I'm not going to talk about Bielefeld.

You might want to hear something about that.

I'm not allowed to tell you... right?

Okay, so today I'm going to talk about

a bit what is in my field

of experties.

If there is one thing

I want to bring across to you

then it is

It's not about a single person

showing this to you this evening.

This is a team effort and a real team effort.

So most of the images are done by

a college of mine Julian Ziegeleder.

And the PI of the project,

so the leader of the project

Sebastian Rabien

has contributed some slides.

And I wouldn't be standing here today

and showing you these images

if it wasn't for a huge team

and many people.

I hope this is reasonably complete,

but I think there were even more.

Many people have tributed most and

long years of there career into such a project.

So this is never about something

which a single person does

and he or she finds something very cool

and then saves the world.

No, it's always a big, big team!

But before we actually see the lasers

then in working, we have of course to clarify

why we do this.

This is not just because we can.

We can! But there is a reason for that,

because if you want to get funding,

you have to write a reason and a reasonable
reason.

Not just because "We want to!"

So in the first part

I will introduce you

to the whole thing

and we talk about bit... about the problem

which we want to tackle

with this kind of technique.

I will mostly present only diagrams

not actual hardware blocks or relays.

So you get the basic concept.

So when we do astronomy

we do two types of things.

We either do imaging,

which is: We maybe produce a nice image

of a star - so that's the blop over there -

or we take this image,

maybe this little blop over there,

and make it into a spectrum,

so disperse the light,

and then we look at the differential intensity

between the diverse colors

or are there maybe

- for example you see black lines in there -

absorption bands and so on.

To do such a thing you need a spectrograph

and in a spectrograph

there is a thing called an entrance slit.

So this slit you have to

put over your objects,

so you don't get light from left or right next to the object

to what you want to observe or analyse

so that you only get light from where you
wanted.

The thing is now

this slit can not be made

arbitrarily wide or small,

because the width of the slit directly

determines what kind of resolution

you have in such a spectrometer.

as it's called. This is a quantity

Which needs to be above a certain value

when you want to do certain kinds of analyses.

So it has fixed width.

So now if we look at an image produced

of one of the most capable telescopes

on this planet

and we put a representation for this slit

over the star

- okay now its white, let's make this black -

then you see if you want to go

for that star over there,

you do have a problem already.

As said, you can't make this slit wider,

but the star is actually larger than the slit,

meaning that you lose light.

"Well you lose some light...." No!

If you want to quantitative measurements

you want to have all the lights

and all the pixels.

So you can't get rid of them

and just throwing something away.

So, but our image is looking like that.

It's maybe nice, so but can we do better?

Yes, we can!

And this is what we can achieve with

adaptive optics.

This is an image that has been produce

with adaptive optics with a

LASER AO assisted system.

And if I flip back and forth you see

there is a difference!

All right! So why is that?

Why don't we get this ideal images?

The reason is because there is the atmosphere.

The atmosphere is great for breathing.

It's not that great for astronomy.

So if you have a star up there somewhere

in outer space

- can be very far away - so the photon

have travelled for 11 Billion years

and now they finally hit the atmosphere

and then something happens

which you do not want.

Okay, first they travel freely.

There is a nice planar wavefront.

So it's not disturbed by anything,

maybe something but that's not the

scope of this evening. It's planar, it's nice!

And if you actually have a satellite,

it's very cool.

Because then you can directly record this

undisturbed light.

If you have something on the ground,

well, you do get a problem,

because the atmosphere introduces turbulence,

because, well, the air wobbles a bit.

There are stream coming from all directions.

There are temperature gradients in there.

And these all work together

and make from this nice planar wave front

a crumbled one.

If you have a perfect image

which you create

- This is called "diffraction limit".

This is just limited by the size

of your optics.

So the wider your optics is,

the nicer your resolution is of your image.

If you then build a large facility with

maybe two 8 meter mirrors on the ground,

well, you only get your seeing limited image.

Seeing limited. The Seeing is this wobbling

of the atmosphere as it's called.

And that's about it.

You can make it arbitrarily large.

You won't get a better resolution

then a backyard telescope

of having 20cm in diameter.

So yeah...

What to do?

There have been people, of course,

thinking about this problem longer.

And the first idea came up in 1953.

And some guy Palomar Observatory

in California said: "Well, if we have

the means of continuously measuring

the deviation of rays from all parts

of the mirror and amplifying and feedback

this information so as to correct locally

the figure of the mirror

in response to schlieren pattern,

we could expect to compensate both

for the seeing and for the inherent imperfections

in the optical figure."

Ehhh... what?

So if we could somehow get rid of this wobbling

or conteract that,

then we could get this perfect

diffraction limited imaging we get in space

also on the ground.

In the 1970s the US military started

to experiment on that.

Well, I guess the Russians too,

but it's not... it's known that the US started

at Starfire Optical Range.

In 1982 they build the first AO system,

adaptive optics system.

The "Compensated Imaging System" on Hawaii.

And in the late 80s the first astronomical
use,

adaptive optics system "COME-ON"

as it was called was installed at the

Observatoire Haute-Provence

and at ESO at La Silla.

That's the European Space Observatory.

All right so that was:

Yeah, we get for we found that this

fussy blob is actually not a fussy blob,

but two fussy blobs.

laughter

Well it's a binary system as I would say

if this was at an astronomical conference.

But yeah, you disentangle things

you could not see before.

Okay! How does this AO system look like in
principle?

So again we have this star somewhere,

we've learned already that

we do have... - actually you see this slight

schlieren pattern in the air

for the warm and the exhaust from the...

Yes, there is a bit flimmering in the background.

That's seeing. Okay?

So the image is not as sharp here as

it comes from the projector.

Okay, that comes from somewhere

and then we need a system

which has three components.

One is a deformable mirror,

the other is a wave front sensor

and the third one is a real time computer.

We need something to actually measure

what is going on.

Then we need to take this measurement

and extract some information from

this measurement

and then we need something

which can correct this wave front,

straighten it out so to speak,

'cause we want to have it straight again.

So the wave front sensor sends some information

to the real time computer.

This some information namely is:

What is the curvature?

How does this wiggled thingy look like?

- The wavefront -

And that real time computer computes

then information that goes

to the deformable mirror

and that in real time shaped

in an arbitrary shape

conteracting that incoming wave front

and then straightening it out.

So we do have a light path like this.

First it goes on the deformable mirror,

goes on something else,

which I will come to in a minute,

and then this wave front sensor.

And of course this means if you run it

you do have a control loop,

meaning measure something here,

the wavefront,

you put the information into there feeding

that into the deformable mirror,

that deforms somehow,

modifies this wave front that comes

from above and then of course

you want to have a feedback loop:

Is that what I did enough?

Do I have to do more?

And also: Of course in the next second

or split second this pattern

will have changed,

because the atmosphere is dynamic.

If it wasn't dynamic we don't need

to do this in real time,

but we have to do it in real time.

Real time meaning we have to do this correction

and calculation and sensing at a rate of

about 1 kHz, so a 1000 times a second.

Then we have a scientific instrument

because actually we do want to see

what is in there.

And so this thing in the middle

is a beam splitter.

It takes some of the light,

puts it to the wave front sensor

not all, because most of it should go into

the scientific instrument

and there, as you see here,

then the wave front is straightened out

again and then I can focus it

into my instrument.

To do actually that

I have to do...

- This is the one slide in this talk

with a Greek symbol -

You have to this incoming wave front

which is shown in orange

and then you do a piecewise linear fit

which is an approximation

of the slope.

Of it actually how it looks like.

It's put into linear pieces.

And the size of what is normally

can be taken als a linear fit

Piece is roughly 10 - 15 cm

for good observation sites

while this thingy here

so this is the primary mirror of the telescope

which collects all the light

that comes from outer space

is usually for the big telescopes

at this point 8 to 10 meters

Okay, but how do we get this slope?

Now we know that we can approximate it

in pieces, but how do we get

the slope?

Because we need theses slopes of course

fed into this deformable mirror

to maybe okay:

If it comes like this, I go like this

and it comes in nicely

or comes out nicely.

So is where the sensor comes in.

There are different types of these sensors,

but the one we are using

is a so called Shack-Hartmann-Sensor.

And it looks like this.

We have... this is the ideal case of course.

So we have an incoming planar wave front

- straight on.

And we do have an array of lenses,

so it's just 1.. 2.. 3.. 4.. lenses

and then in an array like 4 by 4.

And they all focus what is coming in

into onto a detector and this wave front

that is coming in is planar

like this on the left.

Then you do get a regular spaced grid

of focus points, in this case 4 times 4

so 16.

If now this incoming wave front

is no planar it looks like this.

So the focus points do move a bit,

because, well, it came in like this,

so the focus is offset.

I will flip it back and forth again.

So it's looking like this and you see

of course you do know what is perfect

meaning they are
at their designated grid points.

If its imperfect, well, then just measure

the deviation from their zero position

so to speak

and then you do have a proxy for the slope.

Of course it's a bit more complicated
than that.

There are matrices involved which are not

necessarily in a square form

and you have to invert them

and if you don't... yeah... ...

There are pretty clever people

and programmers working on this type of

problems.

And this is actual current research.

This is far from done, this field.

Okay, so suppose we do have the slopes.

Then we take a deformable mirror

and this is the zeros order approximation

of a deformable mirror.

Let's say the wave front looks like that,

well, then take just a mirror which is

maybe reset a bit in the middle

the other tipped forward.

It bounces on this mirror

and because there is something sticking out there

and in there

well if this approaches there goes back

and in the end the whole thing

when it has been reflected is planar again.

Okay, that as said,

that is the easiest order approximation

for that. It's a bit more complicated.

Your incoming wave front doesn't look like that

It's normally a bit more complex.

And that means you do have to have

more wobbling in your deformable mirror.

You could do this.

That's in the upper diagram.

You could do this with a membran

which is continues

or maybe it's also in pieces

and this segments are driven up and down

or maybe tilted by piezo stages

that are put underneath.

Remember they have to do like

a thousand times a second

or you could do something like

you take a two piezo electric wafers

they have opposite polarizations

put electrodes inbetween

and then when you apply a voltage to this blue

electrodes then you have local bending.

So the one thing will bend up,

the other ones will bend in the opposite direction.

And then you do have changing curvature

on this whole thing.

It's not that easy of course in reality,

because they are not completely independent

one cell will influence the other

and yes...

But this is the basic principle.

Okay, now you have seen

there was this beam splitter.

So most of the thing goes into the

science instrument

and some goes to our wave front sensor

of the light.

If the object we want to record like

a galaxy that is 11 Billion lightyears away

then this galaxy is to faint.

We can't analyse it's light.

So what do we do?

We need maybe a star that is nearby.

So our galaxy, which we actually do want

to observe, is the red thingy

the bright star is the yellow one

and if there are reasonably close together

- reasonably close meaning

about 10-20 arcseconds.

If you stretch your arm and look at

your little finger at the finger nail,

this is about 30 arcminutes.

1 arcminute has 60 arcseconds so it's

very close!

It's not like the galaxy is there

and the star is there. No!

It's there!

Because if you have a large separation

then they do sense different turbulence.

Simple as that.

Now the thing is

that less than 10% of the objects

you have on sky

which you are normally interested

do have a sufficiently close and bright star

nearby.

So what to do?

And now we come to the lasers.
laughter

Because if don't have your....

If the don't wanna play nicely

build your own themepark with yes ... you know.

So make your own star!

This is what we do.

Because if the star is not nearby,

a sufficiently bright one,

well, why has it to be sufficiently bright?

Because if you want to do this computation

a thousand times a second, well,

then the time for your CCD
when you record this image

for your wavefront is a thousands of a second.

And if you don't have enough photons

in a thousands of a second, well,

then there is no computation of this offset

of this little green dots on that grid.

So you need a lot of photons.

So let's get enough photons!

And there are actually two things

what you can do.

There is a conveniently placed sodium layer

in the upper atmosphere.

laughing

It's 19 km above ground

and there is a sodium layer.

And what you actually can do is

you can take a laser on ground here,

and then shot laser which corresponds

to the energy transition of this sodium atoms

which is 589.2 nm. It's orange.

And excited those atoms up there

in the atmosphere and they will

start to glow.

And if you have a focus,

if you focus it in there,

and than you have a blob of sodium atoms

lighting up in the upper atmosphere,

maybe... what ever some hundred meters long

and some meters wide
as big as your focus is there.

This can be done with a continuous laser.

This has been done in the past.

Yes, of course.

And actually the first instruments

were build like that.

The thing is

in those days they were very, very expensive.

There is no sodium laser.

There are only Di LASERs and they are messy

and expensive.

Nowadays we can build this as fibre laser

but not ten 10 years ago or 15 years ago.

An other solution is to actually

use Rayleigh scattering in the atmosphere.

You use a Nd-YAG LASER

which is 532nm. It's green.

It's easily available, it's cheap

compared to the other one.

And then you focus it in the atmosphere.

The only thing is:

You will do have backscatter of photons

all along the way.

So you have to think about

how can I only record light from

a certain height above ground?

Because otherwise I don't have a spot,

I have a ...ehhh... a laser beam column

somewhere there.

Okay!

How do this things look like?

Can we dim these lights actually a bit?

Or is it only an off switch?

Can you check on this? Let's check on there...

Just push the button... come on...

No? No. No!

laughing

Nooo!

It's still on here...

gasp

All right, it's looking like this.

Who has been at the camp?

There was an astronomy talk at the camp

from Liz.

Actually if this talk had been tomorrow

we would had have a live conference

to that side because Liz is right now here

and she send me that picture

just some hours ago.

That is how the just do things on

Paranal in Chile.

The thing I will talk about

is the green one to the right.

That's the thing I have been involved with.

Yea, let's look into that.

So if you shoot the laser into the atmosphere

of course you do have problem.

The star is very far away,

it's infinitely far away.

And the light that comes down

is in a cylinder.

And if you shoot a laser up, it's a cone.

So you only probe the green region.

The unsampled volume of turbulence
is to the side.

That is a problem with our laser AO.

An other problem we face is this one.

When we take a star to measure the wave front

then it passes only once through the atmosphere.

The laser beam goes up and down.

And so there is a component

called tip tilt component

which is actually just the thing moving around

It's not just the phase

that gets disturbance introduced

in the wave front but this moving around.

So not the bright and more
or less bright twinkling

little star thingy,

but the moving around.

And that can not be sensed
with a laser guild star.

So when ever we do laser AO

We do need an other star

to get this component.

But this star can be a bit further away,

like an arcminute or 2 arcminutes or so.

So it's that... is wide. There are enough.

And then we should think about

actually what we have to correct and so

we should make a profile of the turbulence

above ground.

And this is how it looks like.

And for example for the side

where we are there in Arizona

we see that most of the turbulence

is actually just above the ground.

So we maybe should care mostly

about the ground layer.

It's not so much about the high altitude things.

So and then what we do is:

Well we want to sample

the ground stuff nicely

so we don't take one but 3 lasers.

So to fill this area nicely.

And yes, of course, we can also combine this

and this looks like that.

This combination we will not talk about today.

We will only talk about that.

This is how it looks like.

So this is our telescope, the primary mirror

which receives the light from outer space

it then deflects on the secondary, tertiary

and than somewhere here.

But first we need to have to shoot the laser up.

And it's launched from a laser box

onto a mirror behind that secondary mirror

over there into the atmosphere

and after 40 microseconds it reaches

an altitude of 12 km.

And then of course it comes back.

After 80 microseconds it's here

in our detector again.

So the star then lights up,

has this cone, get's focused there, focus,

reflected to here

and we do have our signal
in our detector after 80 ms

and as said, because of course

the laser has scattering all along its path,

you want to gate this information to 12 km

and well then you just -just- look at

when your laser pulse started

wait. wait. wait. wait. wait.

open the shutter for the detector

for short time after 80ms,

close it again and then analyse

and read out what you just did.

Easy, huh?

So we are done.

Thank you for coming to my talk

and now go out and build your own lasers

with... to...

laughing

Now we are going to look at this thing

which is actually build and which works.

So this is called ARGOS.

It's a ground layer AO system.

That's what we want to build.

It has wide field corrections.

That means you can not correct

just a tiny patch on sky but for for astronomical use

a huge area, meaning it's not just

a circle of 10 arcseconds but

this thing can correct 4 by 4 arcminutes

which is huge,

so all the objects that are in there.

We have a multi-laser constellation.

We have seen that why we need this,

because we want to fill

the complete ground layer.

So we have 3 laser guild stars per eye.

Why per eye?

This will be clear in minute.

And we use high power pulse green lasers.

And this deformable mirror is actually

build in the telescope system already.

The secondary mirror is the deformable mirror

which is very convenient,

because then all the instruments,

that sit on the telescope can benefit from

this system.

It's installed at this telescope.

Look's pretty odd. Yes, I admit that.

That's the Large Binocular Telescope.

It's two telescopes on one mount.

One primary, two primaries.

It's roughly 23 by 25 by 12 meters.

It sits on Mont Graham in Arizona.

And it has an adaptive secondary mirror

which is this violette coloured thingy

up there in the middle on top.

This is how it looks like.

This is the control room

where you sit.

This stays fixed.

All this shiny part rotates.

That's the actual telescope,

the red thing that moves up and down.

So the whole building rotates and it moves

up and down.

It's from ceiling... the ceiling is at level
11.

So when you actually sit there,

you can watch around a bit

... this is outside... it's winter... yuh!...
let's see...

There is a ladder...

Yes, this thing is huge...eh.. nice.. cool

Okay, that's what it's looks like

when you are actually there.

Okay, our system layout is like this.

We have this adaptive secondary mirror

which is the deformable mirror.

We have the primary, tertiary.

That is clear already.

So we have a laser box.

The green things is the lasers themselfs.

So that's how it looks like.

We produce some laser beams.

We have steering mirrors in there

to get them into the right pattern on sky

of course.

We do have control cameras,

if : Is the focus right?

Is the position right?

This is one control loop

another control loop, another control loop

an other control loop.

The black thing is the shutter.

Because we have to close this whole thing,

when aircrafts are overhead,

when satellites are overhead.

So if you want to use this system,

you have to, 6 weeks in advance, you have to

put out your list of observable targets

to some military agency.

And they will tell you: Okay! Not Okay!

Okay! Not Okay! Not Okay! Not Okay! Okay!

Not Okay, meaning something is passing overhead.

Hmm... what could this be?

laughing

Of course, at some point the lasers

come down again in this cone shape.

They will reach the primary mirror

and ultimately it will end up

in the wave front sensor

which is much more complex than just this box.

I showed you before.

So there are aquisition cameras

which detect are we at the right spot.

Do the spots get onto the detector

in a nice fashion.

We do have to do this gating, remember?

We have to open this shutter

for the CCD when we want to record the light.

This tiny fraction after 80ms.

After the laser pulse has been launched.

It's done in here.

These are Pockel Cells.

So its an electro optical effect.

And then there is also something

in addition because I said

we can't do without the tip tilt

and there is another unit in here

that sits right in front of the science instrument

that detects this tip tilt star,

this additional star.

So you have the laser wave front light,

the green one, you do have this tip tilt light,

the blue one,

and you do have the actual science light

from the object you want to observe on sky.

That goes directly into this scientific instrument

in the end.

And then you have a lot of control things.

Of course, you do need a common clock

for this synchronization of all this pulses

and the gating and what not.

And of course you need the information

for the tip tilt component and for the wave
front

into this computer

which sends then all the slops

- you remember we have to do this

linear approximation pieces wise, yes -

into the secondary mirror

which than deforms in real time.

And does this a thousand times a second.

This is how it looks like.

So when I am there I am roughly that tall.

The two black tubes right in the middle,

those are the two tubes which go up.

Looks like this.

So, this is how the components are distributed

over the telescope... once back.. okay

primary mirror, primary mirror,

some instruments in the middle,

some tertiary mirror,

the secondaries, the adaptive ones up there.

Yes, I hate to use this laser pointers.

laughing

Because I am always going like this... eee

(green laser pointer on the slides)

laughing

That's my man! laughing

So okay!

So we do have the adaptive secondary

up there and then it goes back on the

tertiary down there and then it goes over

into the science instrument,

all the wave front sensors and what not.

Again, we do have a laser system.

We have to place somewhere a launch system

for the laser, a dichroic to separate

between the laser light, the tip tilt light
and the science light.

We do have to have a wave front sensor

to check how the wave front looks like.

We do have to have this tip tilt control.

We have calibration source.

A calibration source would be nice

to calibrate the system during daytime,

aircraft detection, yes, satellite avoidance,

-also an issue here- and a control software.

There are many people just writing...

...just haha... writing software for this.

And this is really hard.

Some are also on the conference.

They don't want to be pointed out

as I learned, but you will find them

at the conference, if you look at the right places.

That's where the laser box is located.

Just next to it is the electronics rack.

How does this thing look like?

So that is one of our lasers.

It's about 20 W. Don't get your finger in there.

laughing

It really hurts.

(Did you try?) No!

There is a mandatory annual laser training of course.

Yes, if you want to have something
like this at home,

you do need a huge refrigerator next to it

just for the cooling of that thing.

This is nothing you want to have at home.

Just because it's... that bulky... no..it's
not..

but actually when you do

this green laser pointer thingy

then there is always this always this:

"Don't use this for more than 10 seconds."

Because why? Because the crystal inside

heats up.

And if you can't dissipate that heat

the crystal at some point breaks

and then your laser pointer is broken.

This thing gets continuously cooled.

So, therefore it's a bit more expensive.

laughing

If you than put it up,

so this is still on the lab table

when it was integrated and tested

and than at some point it gets put all

in a box with all this control mirrors

and cameras and what not.

But finally you see in the middle

on this picture there is

a focusing lens and then you see

these 3 tiny little beam coming out of there

which than expand on sky in size

of course when they are in 12 km height

but that's how they come out of it.

And if you install this in the telescope,

you actually have to tilt the telescope,

because otherwise you can't reach it.

And then you need your climbing gear.

So once you have produced the lasers,

you need to propagate them to a through

a dust tube onto a launch mirror,

a folding mirror and from there to

a launch mirror.

Yes and then it looks like this!

Okay, so the lasers come from here into that

and then over to the other side

over the secondary mirror and then

being shot right up into space

like this.

Okay, so if you want to have that at home,

.... eh... but I can tell you the whole facility

does cost less than one fully equipped Eurofighter

laughing
applause

Thank you for taking the hint.

Yeah, that's how it looks like.

It's.... yes it's... laughing ... yeah...

laughingapplause Okay?

okay... I have to admit this are a bit longer exposers.

It's not that bright and green

when you are actually at the telescope up
there.

But if you have been in the dark long enough

around ten minutes, then I really becomes bright.

There is a little telescope that observes,

where actually the spots are on sky.

And if we have clear sky,

then we have this constellation on the right.

So that is how the lasers come up.

As I said you do see them all the way up,

but we are interested in the little dots

at the end.

You can barely see them.

If there are high clouds,

well than we produce something like this.

We have the dichroic when the light comes
back down

as said.

Which separates the science light in red

and the laser light in green.

This is how it looks like.

Actually the dichroic is right in front of
Sebatian there

and from there it gets then reflected

on a reflector and then up into the

wave front sensing unit.

So there is the dichroic, there is the reflector,

and it goes over in this unit

which is the wave front sensing unit

which sits there, at the side.

That's how it looks, when it gets installed.

And that is how it looks inside.

So you have the 3 laser beams coming

from the side, from the sky, of course.

You have patrol cameras

which monitor where are these?

Are they at the right spot?

Do we have to steer the lasers a bit?

Than we have some control for the position

of the laser spots and the field.

The Pockel cells are the ones

that do this opening and closing in front

of the shutter.

You can't use a mechanic shutter in front

of the CCD.

We have to do this electro optically

So you have a polarization of the laserbeams.

And you have a polarizer... a cross polarizer

and then you just turn the polarisation

of the crystals.

It's an electro optical effect

and then it gets passed through

or it gets blocked.

Then you also of course have this lens slit arrays

in there and then the CCD

which actually records this dot pattern.

You remember, this 4 by 4...

well it's not 4 by 4 in our case we do

have a bit more resolution.

The sensory looks like this.

This is actually a custom build CCD.

Very special.

The imaging area is in the middle

and when you read out the thing,

you split the image in half,

you transfer it to the sides

to the frame store area and than read it out.

'Cause read out is slow, transfer is fast.

And you have to do this a thousand times

a second at very low read out noise,

which is only 4 electron read out noise.

For the experts here in the audience,

this is very good.

It's not many pixels but it's more than enough for us.

So how does this look like?

It looks like that!

So there you have your pattern again,

regularly spaces pattern of course

from 3 laser guild stars you get 3 patterns

and then you analyse, well, the position,

the relative position, the absolute position

of those stars on their grid,

and somehow compute this slopes

from there feed them back, compute then

actually electrical information from them

which you can than feed into your

deformable mirror again

which sits on top of the telescope

and then hopefully everything works.

This you can digest at home.
laughing

It's in the stream now so it will be

saved for all eternity

and all the aliens

which record all the electromagnetic field

from Bielefeld... (mumbling)

laughing

Anyway, so, just in short.

There is down in green there is this thing

that goes up from the lasers through

some steering mirrors.

We have diagnostics, then we got to focus

check launch mirror one and launch mirror two

onto sky and then we go back

up there N1 is the primary mirror.

And then we go through this whole chain

and there are various control loops

sitting in there.

And all this things have to talk together

on very high rates.

Sometimes you see 1 kHz other things are a bit slower.

This all needs highly sophisticated control software.

And the programmers can be real proud

of what they did in the past

with all this control loops.

The tip tilt is very... much much much easier,

because all the...

you remember this tip tilt

so this all is moving around.

So you have 4 quadrants at a little cell

and it moves to somewhere up, down,

left, right.

You can easily detect that.

That is feed into an array

of 4 Avalanche Photon Diodes

to actually record this and for that

we don't need many photons.

So this tip tilt star can comparably...

be comparably dim.

The calibration unit for the daytime calibration

can be put into the beam,

so this arms can swing over,

over the primary mirror and then we can

inject artificial stars via a hologram

into the whole unit during daytime

and calibrate this whole thing.

And than yes, we are back here.

This is how we look like.

Maybe concentrate on this two areas first.

I will flip back an forth many times.

But, yeah, what is this?

Are this two stars which are just fuzzy

and dim?

Or is this an extended object?

The upper one may be a galaxy because it's

elongated.

Okay, concentrate on that.

Well, it actually just a bunch of stars.

And this is over a huge field.

So the correction is not just in the middle

but you can see also at the very edges

of this image, we do see this improvement

in image quality.

Of course you can have the diagram, if you want.

So the blue line is without the thing beam activated,

open loop,

and if we close the control loop, to do

this measurement and correction in real time

we do squeeze all the energy into a few pixels

which of course also means

our signal to noise level in a single pixel

goes up tremendously.

Meaning you can decrease

your exposer time.

Which is important if you want to observe
galaxies

at this telescopes

it's 200 Dollars a minute.

laughing

It's not cheap.

Okay, good so... the thing...

just last week there was
another commissioning run

testing commissioning run for this system.

And my colleges José Borelli and Lorenzo Busoni

have done a nice video.

The music btw. "hallo gamer"

it's royalty for ears...

If it was now darker therefore I asked,

this would come up nicer,

but let's see!

There is sound hopefully,

so the sound guys, let's see!

applause

Of course this a longer exposure.

It's not that starwars like

I would have loved to use some starwars

tones along those. But you know, all those rights

and... what not... yes... anyway!

That's how it looks like.

So you have 3 laser beams per eye.

Remember, we have 2 telescopes on one mount.

They look roughly in the same direction

but still...

So if you observe two telescopes

at the same time it's only 100 dollars a minute.

Yea, This is not so much the shiny part

on the dome itself, but if you actually

do stand on the mountain during night

and are a bit dark adapted,

you see the laser beams like that.

And don't be fooled!

If you are at the valley,

or very far away you hardly see them.

You don't see them at all.

You see them there.

If you are two kilometers off side already,

it's merely a dim greenish something.

If you are down in the valley 10 km off,

you don't see them any more.

If you take a camera, 5 minutes exposer, yes!

But otherwise, No!

There is no such thing as

"The people in the valley down can see like

these lasers pew pew every night.".. and no.

Ok, which gets me to the last part.

How, do you become

and how do you work as a laser rocket scientist?

Yes, I put this in the talk directly,

because I do get this question in the Q&A, normally,

when I talk about these things,

and it's always like:

"What do I need to do if I want to do this?"

Maybe you have already an idea about this

because you have seen
how complex this thing is.

And, there are so many things to do in these
kind of projects

and on various levels, also in the administration,

also for senior people, new people, maybe
master thesis works on that

or bachelor, or PHD or then as a post-doc.

It's very complex.

Yes, and it's not only about just shooting
lasers in the end.

Sometimes it's just about checking the cables

It needs to be done.

There is a tremendous amount of electronics
and electrics involved.

There are all the mechanical components in
such a system are custom built.

Either the institutes built it themselves

or they give it out of house.

There are these real time computers, for example.

this is by the way our real time computer

from micrograde, if you want to look that up.

it's company. It builds these things.

They need to be programmed.

Oh, if actually somebody is here in the audience

with real hard core experience on

real time computing, coding and such things,

do talk to me!

laughing

Yeah, this is how our software system looks like.

A very small part of the GUIs. It's a lot of code

and a lot of work and a lot of sleepless nights

in front of these computers
and just testing it and testing it

and then testing some more,
and testing even more.

And, to be involved in these kind of projects,

you don't need to be a laser physicist,

because there is no one thing.

If you want to take 3 messages
out of this, it's:

it's a team effort, there are many tasks,

and there are many jobs,
and you have to pick one.

Because in this one job you do in these projects

you have to be very, very, very good.

Because there are other people that are very,
very, very good.

If you work in these kind of projects, if
you meet a new person for the first time

just assume that he or she knows
everything about this

and you know nothing.

You will quickly realize if that is true.

But otherwise, if you assume it
the other way round,

you just make a fool of yourself, okay?

Don't do that.

People in science, second most important thing
if you really want go into this,

people in science are just like
people outside science

meaning you will meet nice people
and you will meet.....

laughing

just like in life.

It's not that these things are spheres
where people are, you know

floating above the lab surface and nice coloured.

No, it's hard work.

And if you actually go into this
like study physics

or maybe if you want to construct this,

of course all the drawings are done by

people how have learned this in there studies,

so "Maschinenbau" what ever...

Go for that one.

Building optics needs optics experience.

If you want to actually build stuff,

well, there are many people in this institutes

or universities who work

in the mechanical fabrication departments

or electronics departments.

They just do PCB layouting all the time.

But this things do need sophisticated electronics

and this all custom built.

This is nothing you can buy of the shelf.

Nothing of it! Almost nothing.

And this means you might end up

with something equally cool.

It's not that you can have this one thing

and then BAM ten years later you will be

the laser-rocket scientist. You won't!

You might become one

and then even after 10 years,

you might realize this is not the thing

you want to do forever.

So I have to correct

the introduction in one point:

I'm no longer working there.

I recently left.

I'm now have my own company.

I'm still involved in these things.

I do calculations for this kinds of things,

but I'm not at an institute any more,

because I decided for example for me

that the contract conditions in this type

of scientific work are not of the type,

which I want to live with any more.

Like one year contracts.

applause

And so there are many ways
of being involved in this

and don't just... don't just
focus on the this!

Focus on what you really want to do and

you might end up in this

and if you don't,

well you do something equally cool.

All right! Questions?

applause

Herald: Okay, first of all

thank you for our daily dosis of lasers.

I have said... Ich hab keine Zeit...

cause we have really not much time left for Q&A,

so I'm first asking the signal angel,

if there are any questions from the internet,

because... was that a 2? 2! ok.

because this people can't ask questions afterwards,
soo...

Peter: I'll be all congress and
if you want to reach me

directly 7319 is this telephone.

Herald: Ok, the signal angel questions.

Signal A.: Yeah, the first question from the
internet was:

How strong the laser actually is

or if it could be any danger for something

in the vicinity?

Peter: Actually, no!

So we shoot up around 15 to 20 W

per laser beam.

If there was actually a plane flying through

our laser beam,

then nothing happens to the pilots.

They don't get blinded or what not,

because it's di... the beamsize at that altitude

is so big already.. they will of course look like:

"Errr what is this?"

And that's what we do not want,

because then they might push some other buttons

which they are not suppose to push.

laughing

If you of course work directly at the system,

you have to maintain it,

you open it, you have to align the lasers

and what not beyond there self aligning capabilities,

you do have to wear
all this protective laser goggles

and what not, because if you do...

if you don't you do have instant eye damage.

It is not... no its instant.

You might not see it instantly.

But the instant... it's there instantly, period.

So really, folks, don't experiment on this

laser stuff at home,

if you are not following basic
laser safety rules.

Not prying this things from the DVD burners

or no blue ray thingys "uuh does it really work?"

Just, just don't!

Your eyesight is not worth it. period.

It's not!

Herald: Please remember to cover
your still working eye!

Peter: Yeah... only look into the laser
beam

with your remaining eye.

Herald: The other question?

Signal A. :And the second question from the internet

was... It's actually commenting that,

this was a very cool concept already been used

and where do you see this going

in the next 10 years, so what's the outlook

for observation from the Earth's surface

in the next 10 years?

Peter: Oh, of course

the telescopes will get bigger and bigger.

The next generation of the telescope is coming up

in the 2020s.

The European Extremely Large Telescope

will be about roughly around
40 meters in diameter.

These are so huge they can't work in

seeing limited operation any more.

They do have to have laser AO all the time.

It will look similar to this.

So this is in that sense also

a technology demonstrator.

There will be a combined thing.

You may remember this diagram

with the one sodium laser in the middle

and the others outside.

So these combined things.

And then you can also imagine something,

that you probe different heights

in the atmosphere,

because you do have different turbulence layers

and all of these then have their own

deformable mirror.

So it's a very comp... gets a very complex
set,

a multi conjugate AO as it's called.

And then there are of course

new... there is research being done on

how to detect this wave front

most efficently.

And there is a so called thing called

the pyramid sensor.

You can look for that, also

we do have one in our system.

And this is very efficient.

So it takes much less photons

to get to the same signal to noise level.

This is active research and... well...

Every major telescope of course now has this.

And every big telescopes in the future

will have this all over the place.

Herald: Okay, we're completely out of time.
Again.

Again, so thank you very much.

Peter: Thank you!

applause

postroll music

subtitles created by c3subtitles.de
in the year 2016. Join, and help us!