Transcript Thermal compensation simulation with ANSYS

Thermal Compensation
System
Roma Tor Vergata
Outline
• Updates
– TCS positioning
– TC system
– Status of the quotations requests
– Installation planning
– Laser power stabilization
• Referee comments
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
After last Virgo Week, a meeting with all the people
(software, electronics, infrastructure and vacuum)
involved in TCS installation and the Detector Coordinator
on September 17th.
•UHV ZnSe window
•Vacuum cabling and feed-through
•Control software for stepping motors
•Electronic boards and cabling for stepping motors
•Laser chiller
•Acoustic enclosure
•Defined position of the optical benches
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Optical benches positioning
Oven
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
This allowed a final design of the Optical Imaging System
Definition of the optical components completed last Friday (October 12)
Defined list of lenses
E. Genin
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
UHV ZnSe viewport
DN100CF
Clear aperture 90mm
Size of the first in-vac
steering mirror increased from
140 to 190mm
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Dimensions of the optical tables defined
(E. Genin, A. Pasqualetti)
750mm
1800mm
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
All quotes in millimeters
NOT IN SCALE
492
Beam Dump
L7
L8
AOM
500
CO2 Laser
l/2
waveplate Polarizer
Power
Meter
264
198
AXICON
50
Components in the blue box are used for laser
power stabilization, will be installed by mid 2008
30
537.5
L1
L2M
237
810
(762)
Beam
Dump
L4
LX
HeNe Aiming Laser
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
To in-vacumm
steering mirror
Two more CO2 optics companies contacted:
ULO Optics
Laser Research Optics
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Provisional installation planning
Starting date depends on arrival time of components
Duration of the single tasks agreed with all the people involved
Duration of the whole TCS installation approx 3 weeks
Manpower:
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
CO2 laser power stabilization
(TCS phase 2 – Mid 2008)
25W CO2 Laser
PD
AOM
Critical components
PD
RIN 
Area
Pinc  D*
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Referees’ comments (R. Flaminio)
Referee: “The second important element of the discussion is the system foreseen to provide the required thermal
compensation. Originally two systems had been considered: CO2 laser and heating ring. The main reason for choosing
the laser system were two:
1) it is more flexible i.e. it is possible to heat the mirror not only at the border but also in the centre.
2) it is 'easier' to implement as all the components stay in air.
In practice both these reasons do not apply completely. First according to the simulation I had the impression that we
need to heat the border of the mirror not the centre.
Second, as it is described in the document, the implementation of the laser system is rather complex; two remotely
controlled relatively large mirrors need to be installed inside the vacuum system as no window are available to look at
the mirrors. Two optical benches (2mx0.75m) need to be installed outside of the ovens. The beam has to be sent from
these benches into the ovens and then into vacuum chamber. As one wants to have power fluctuations at a relatively
small level both the bench and the optical path need to be protected with an appropriate acoustic isolation.
Defending the idea that this is easier than installing a ring heater inside the vacuum chamber is a challenge.”
We agree with the referee in considering the installation of the CO2 laser TCS
more complicated than it has been on LIGO. We considered the possibility to use
a ring heater, as it seems a much easier task. However, there are a couple of
considerations that make this solution impractical on Virgo/Virgo+
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Ring heater
Markers
Shields
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
To solve the previous problem,
the heating ring could be embedded
inside the reference mass, but...
…this would require a deep investigation:
geometry of the Reference Mass itself; the
path of the thick cables carrying the ring
heater high power supply; the behaviour
of the Tekapeek CF30 under such high
thermal load (plastic deformation, melting,
creep, emission of unwanted pollutants…)
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Ring heater
Referee: “It is said that only 100 mW are required in LIGO because the mirror
are precurved. what are the numbers ? What is the LIGO nominal curvature?
What was the pre-curvature? How much power was expected to be absorbed in
the LIGO input mirrors? How much was actually found to be absorbed?
I'm asking these questions because some time ago some persons of the LIGO lab told me
that the precurvature was wrong as the absorption was considerably different than those
initially expected. Can the authors clarify these points?”
Originally the LIGO 1 Interferometer was designed to be a point design with
respect to input power. The radius of curvature of the Recycling Mirror was ground
to be to concave to match the effective curvature of the Input Test Masses when
the instrument was hot. It was anticipated that power absorbed in the ITMs would
create a thermal lens that would alter the effective radius of curvature of the ITM
as seen from the power recycling cavity side such that it matched the recycling
mirror. This relied on accurate knowledge of the absorption coefficients of the ITM
substrates and coatings. These are poorly controlled parameters. The Hanford 4K
interferometer has been shown to absorb excess power in the substrates and
hence achieves an optimum power-recycling cavity at only 2.5 Watts of input
power. The Hanford 2K and Livingston Interferometers do not reach the optimum
operating point with the designed six watts of input power.
For this reason, LIGO people faced two different problems: on some of the optics
they needed central heating; other ITMs showed an excess of absorption (about a
factor of five for the worst case) and needed annular heating.
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Referee: “It is said that the losses (better to say absorption) "resulted to
be about a factor 6 and a factor 2 higher than the nominal value ..".
Would it be possible to have the actual numbers?”
According to our simulations, the absorptions of the Virgo ITMs are 7.7ppm and
2.3ppm for the West and North input mirrors respectively. We must not forget that
all the simulations based on the frequency shift method done so far rely on the
knowledge of the elastic parameter dependency on temperature.
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Referee: “It is written that the power needed to compensate for Virgo is 2W. the
comparison with the 100 mW needed for LIGO is impressive. Especially if one
considers that the finesse of the LIGO cavities is 4.5 times higher than Virgo (should
make the thermal problem in LIGO worst). Maybe the answer to question 1) and 2)
will clarify also this point.”
Besides the above considerations, there is a mere geometrical factor to take
into account. There is an empirical formula to determine roughly the power
needed to compensate, given a certain amount of absorbed power:
2
PTCS
R 
 Pabs   ITM   1.4
 s 
Where Pabs is the power absorbed by the optics, RITM is the radius of the mirror
and s is the beam spot size on the ITM and 1.4 is a “refinement” factor found
with the simulations. In the case of Virgo the numbers are: Pabs=36mW,
RITM=0.175m and s=0.03m, thus PTCS~2W. While in LIGO, RITM=0.125m and
s=0.04m, so for the same amount of absorbed power, PTCS is ~0.6W, a factor
of 3.5 lower than the Virgo case.
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Referee: “It is said that the power stabilization needed for Virgo+ is 3·10-7 at 30
Hz. It might be interesting to give the specification as a function of the frequency
(or a reference to a document containing this graph). what is the security factor
taken in giving this specification?”
 21
2
1 10
10 2 2
ztot (  )
z1 (  )
z (  )
1 10
23
zrad(  )
Sh( fv)
 24
5 10
1 10
24
10
10
100
        fv
Black: Virgo+ sensitivity
(with monolithic suspensions)
Red: total TCS noise
Blue: TCS flexural noise
Green: Thermoelastic+
Thermorefractive
3
Pink:1 TCS
rad.pressure
10
3
110
When performing noise calculations, the RIN has been considered to be flat over
the whole frequency range of interest (10-1000 Hz). In this case, the security
factor is about 2 @ 30Hz and about 3 @ 100Hz. This is why we need a power
stabilization level better than 3·10-7/√Hz @ 30Hz.
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Referee: “About the TCS servo it is said that two bull eye detectors have been
implemented in LIGO. Even if their experience is that one is sufficient, this may
depend on the differences between the losses in the two input mirrors. I would
recommend implementing two readout systems as they did.”
We agree with the referee. Besides the bull’s eye detectors, LIGO uses the so
called “pick-off port”, placed inside the recycling cavity, as the second error signal.
In Virgo there is no such port. We are investigating the possibility to use the signal
from B2.
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
Referee: “The description of the commissioning activity looks like 'once the TCS
is installed, the interferometer is locked and the CO2 is switched on ... then we
will try". I agree that things will certainly goes like that.
On the other hand one can try to foresee some commissioning time required
to characterize the TCS (e.g. measure the time response of the system, ..).”
Together with E. Calloni (responsible for the TCS Servo Task), we agreed in the
following commissioning provisional planning: around two weeks will be necessary
to characterize the actuator, i.e. to measure the open loop transfer function (apply
some TCS power and measure the resulting change in ITM ROC and time
constant). One week is needed to set up the error signal. Finally two more weeks
are necessary to develop and implement the control loop.
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review
A. Rocchi - 16.10.2007 - 2nd Virgo+ Review