RESULTS OF VIBRATION MEASUREMENTS

ON LA SILLA TELESCOPES

Philippe Bourlon 1, Thierry Ducros 2, Michel Faucherre 2

IOCA, Departement Fresnel, URA 1361 du CNRS, 06460 Saint Vallier, France

2 ESO, Garching

Abstract

With reference to VLT Interferometer requirements relating to mechanical vibrations, we describe the

instrumentation and procedures used for telescope testing at La Silla (1989). A picture of conventional (3.60 m) and less

conventional (NIT & CAT) telescope vibrational behaviour becomes clearer from a very large number of samples.

Guidelines are presented for telescope design and for the selection of instrumentation adapted to vibration measurements.

Introduction

This paper describes the results of a mission performed at ESO-La Silla in 1989 on large

telescopes. A number of issues are tackled here:

a) Will the VLT 8m telescopes be stable enough to be used in an interferometer? Up to now, all

interferometers (in operation or being implemented) make use of specially designed telescopes for

interferometry, except the VLTI : the VLTI structure is very large (see figure I) and can accommodate

a number of foci and instruments. Even though stiffness is a driving parameter for other reasons (first

eigenfrequency> 8Hz), it is not known how the structure will behave at submicrometer level.

b) Why do vibration measurements on large existing facilities? I) to identify main sources of

perturbation (drives, wind loading, effects due to nearby transformers, motors or fans...) 2) to obtain

an estimate of the power spectrum of optical path variations in a telescope after integration of mirror

distance variations along the whole path under different conditions.

c) Longitudinal OPD stability requirements for the VLTI 4: the budget allocated to vibrations

within one telescope in terms of Optical Path Difference (OPD) versus exposure time is:

OPD fluctuations = 14 nm rms for a . 10 ms exposure time

OPD fluctuations = 50 nm rms for a 48 ms exposure time

OPD fluctuations = 225 nm rms for a 290 ms exposure time

To be published in the Proc. of ESO Conf. on "High Resolution Imaging by Interferometry", Garching, Oct 14-18, 1991

GOALS

Investigate vibrational behaviour of telescopes

The PRIMARY goal of the testing mission was to obtain some numbers on the actual level of mechanical vibrations found on existing telescopes.

The SECONDARY goal was to identify some of the sources of vibrations.

An important TERTIARY goal i.e. "how to possibly improve the vibration status on these existing telescopes" is much more difficult and was beyond the scope of the work.

1) Instrumentation used. Noise

limitation and Calibration

a - Hardware and software

* 2 accelerometers Bruel & Kjrer 8306 ('Piezo), working in push-pull mode with

Fig. I : the VLT 8 m Telescope Structure booster amplifiers

* high-pass filter to eliminate the DC windcomponent

* second order filter for antialiasing (transforms accelerations into positions)

* acquisition board from "GW Diesel generator

Instruments". Sampling rates used: 260.1,360.9 and sometimes 5218 Hz. air condj­

tlonnlng:

* Data acquisition and processing with Macintosh computer

b . Noise limitations (mlcro)selsmlclty

a few potentlalWe identified the following sources sources of nols are shown here lllltlof noise in our push-pull set-up:

* sensor random noise

* electrical common mode (power supply drifts, 50 Hz...) are almost negligible

* electrostatic low frequency fluctuations, especially troublesome on electrically insulated

structures under wind load

The usual noise level is plotted on each graph. The measurements were very often "noise limited"

below 3 Hz, for quiet telescope conditions.

c - Calibration: we checked our system response against: 1) other accelerometers with their own data

acquisition system 2) a HP interferometer.

2) Vibrations from oil pumps

The oil pump frequencies (azimuth oil bearing of the NTT, hour-angle oil bearing of the 360)

appear clearly on almost all of the corresponding samples. The resulting amplitude remains small but is

certainly not negligible in terms of interferometric requirements.

On the other hand there is a possible filtering

effect by the oil support on some of the

mechanical resonance lines. (The following

numbers are orders of magnitude, given in nm

rms over a 0.2 Hz band)

NTT @ 85.9 Hz & 86 Hz (the 2nd harmonic

components are visible but very small) :

M2 (parallel to the optical axis) approx.' 3 nm rms

Nasmyth. vertical approx. 2 nm rms

MI (parallel to the optical axis) approx. 0.1 nm rms

Fork base, horizontal approx. 0.3 nm rms

3.60 m telescope @ 173 Hz (the 2nd & 3rd

harmonic components are visible but very small) :

M2 (parallel to the optical axis) approx. 0.2 nm rms

MI (parallel to the optical axis) approx. 0.02 nm rms

Horse-shoe. vertical: approx. 0.6 nm rms

3) Vibrations from the M2 spider

support on the NTT

Surprisingly, most of the measurements, at

various places over the NTT, displayed a

similar spectral signature in the 43 to 50 Hz

range. After some investigation the source was

identified to be the transverse vibration of the

secondary mirror support spider, transmitted

through the telescope structure.

Fig'."2 : The N;iT structure just before it was disassem­

bled and shipped to Chile; last lime it could be secn entirely.

NIT : MII/optical axis Sample # 115 IOlokodl9 21 UL'l

Air condllionning; OFF ~l suppon : ON

AlJmul: OFF AI",udc: III "/, Sampling .: bO I Hz

Buildmg OPEN Wind : N~W 5 m/" AIOIudr IFNml

Jnmmu/.Jlil.

1.' j'-~-~ '0

i-- Spo:lnJm -;;-fI mccharucaJ Ouauanoru '6'.1

" :: J

45".1.­

'-IF·7 "4

,".. Fig. 3 : Vibrations coming from the Mirror 2

support spiders. seen on Mirror I.

Figure 3 : the accelerometers were installed on the side of the primary mirror of the NIT, which produced this typical PSD under quasi normal observing conditions.

rms fuctuations versus exposure time 1:

From measured Power Spectrum Densities (PSD), the rms position errors a during the exposure lime 1: were computed using the following3 : ~ "F 4 10 . The nrc erence on Ig. to

NyqUlst . . . 2 corresponds to the requirement 02(1:) = PSD(Slgnal + NOIse) [1 - SlOe (7t'tf)]df given in Int.(c) with a=aopd/2f1 ffz .

Ideally the integration should be from 0 to infinity and applied to the signal above the measurement system noise. Since our PZT accelerometers are less sensitive at low frequency (cf. §7), measurements were detector noise limited below ... 5 ffz when the environment was quiet (especially on Fig. 9). In this paper the PSD corresponds to (Signal + Noise) and isinlegrated from 1 Hz (measurements valid above == 1 Hz) to the Nyquist frequency.

4) Contribution of the wind to mechanical vibrations

(rmsflucluations given below are computed/or a 15 msec exposure lime)

a - On the NTT; measured on the side ofM I, along a direction parallel to the optical axis

see fig. 4 : sample # 111 (building closed) and sample # 115 (building open).

Average over 5 samples building closed: 14 nm rms for as mls open air North wind

Average over 3 samples building open : 42 nm rms for as mls open air West-North-West wind

Spectrum of mechanical fluctuations

1000� :: ; j 1�

~ .111�iii iii I� Ir~o\ut~o~ : 27 mHz

100� Sample # III�

10 Mardl 89 21:33lrr GlI8 SampIinc: 260.1 Hz-~ 10

Airconditionning: OFF Oil support ONE c

Azimut:OFF Altitude: 13.1 "Is....ooo~---~--j . Building: CLOSED Wind: N6m/s

tOO ···················t:::....-t.. _.

- .� Attitude: ZENITH 0.1 10� . ..•..• :�

• -"",_( ..1141)�

·_···· ..f.,.~

:. ...---- Wind Test on the NIT,.0.01� .000_

1Hz 10Hz Ml /I to the optical axis� 1000 :: ; i : i j j j�

1 l j i" 15� Figure 4 III 1 resolution 127 mHz .__-+-_+--+--+! i i i~;-+i ~ '------------__--J

N:x:� Same, excepted that the building is OPEN...,..-

100

·---+-""~r-_+_+_fr~f~\-------ic----;--H \llJlL__-r­~ 10 Sample # 115

E� 10 March 89 21:SS lrr GlI8 Samp\ina: 260.1 Hz c� 111111 .._~---~--

':L/j//�0.1

." ---t··.. ...,_ 1 i ...

0.01 10.. 100.. 1000_

1Hz� 10Hz

Conclusions from this test: the vibrations on the side ofMl increase by a factor of 3 at 15 ms

exposure time, when the NIT building is opened, for a 5 mls open air wind. This increase, when the

building is opened, occurs mainly at frequencies below 10 Hz. The 14 nm rms value (building closed)

would certainly be lower under a no-wind (and closed building) condition because of suspected wind

influences on the telescope through the foundations even with a closed building.

b- On the CAT (3.60 m telescope, Coude Auxiliary Telescope)

Two samples are available to estimate the wind effects on the secondary mirror (vibrations are

measured along a direction parallel to the optical axis) : these samples do not show any significant

variations between an OPEN dome and a CLOSED dome, for a 7 m/s open air North wind. The

measurements done on the CAT do not distinguish wind effects on the tall tower or through the

opened dome (cf. above, § 4 a). ­

5) Contribution of drives and bearings to the mechanical vibrations

(The rms fluctuation given below is computed for a 15 ms exposure time.)

a - NTT : effects measured on the side ofM 1, along a direction parallel to the optical axis (see fig. 5)

average over 3 samples (drives OFF) : 14 nm rms average over 2 samples (drives ON) : 16 om rms

Specnum of mechanical fluctuations

1000 ++----+--+---+---+--t-+--+-++------+---+----+--+-~-,~,,f-------i:­

~ 1106 Sample # 106! resolution 127 mHz t • i ~ ~ i ~ !

100 10 March 89 21:l3lTT G118 Sampq: 260,1 Hz 1- 1- I ·j-Il-n·--------·

- : 1 i ; ; ~ l

~ 10 Air conditionning: OFF Oil support: OFF

E Azirnut : OFF Altitude: OFFc

Building: CLOSED Wind: N 5 mls

Attitude: ZENITH

0.1

Drive Test on the NIT, 0.01

MIll to the optical axis 1Hz 10Hz

1 000 -rt----+----+---+---+--t-t-+,+.+-,---+---+----+-~,'-+,-!,---+-H-----b­! ! ! . III II~ 1~ • • ,

! r&solution 127 mHz

Figure 5

100 : i j l ! ! ! !

-~ 10

E c

N :I: "?"

i Jill I ! ; ; !

! I I! I

11111~---··· Altitude: 13.1 "'s

Sample # 113

Same, excepted that: Azimut: 1350 "'s

0.1

I I i i

--+---+-iIIII;: i . 10 Much 89 21:55 lIT G118 Sampq: 260.1 Hz

1Hz 10Hz

- The difference on the nns value for short exposure times is negligible.

- There is a clear difference in spectral signature, when the drives are applied: the mechanical noise

increases markedly in the 1 to 10Hz band and a second line appears at 8 Hz.

0.01

0.01

b - NTT : effects measured on the side ofM2, along a direction parallel to the optical axis

The difference between: DRIVES ON (with the tracking rate set for a small zenithal distance, ie Az

: 1350 "/s ; EI : 13.1 "/s) and DRIVES OFF, is hardly significant in terms of rms fluctuation for short

exposure times. These effects were certainly shadowed by the fairly strong wind (7 m/s) with the

building open at the time of the measurement.

c- 3,60 m telescope: effects of the drive measured on the side ofM 1, along a direction parallel to

the optical axis (fig. 6)

Specaum of mechanical fluctuations

1000 Sample # 177 : .177 i resolution 176 mHz

.... 14 March 89 14:07 UT G4t9n Samphn,: 360.9Hz100 ····1··-·1····1···1··1-·_·-···__·····

:J:-,.- Air condition'g: OFF Oil support: ON§ 10

Air:ON ~

Hour-angle: 0 ../s Declination: 0 "/s···4····..--·.··.-······----·-···-··

...ooo~---~-- Dome: OPEN 45 0 Wind: NNW 5m1s ~ ........�

'00 ..·--·····-·1:::~.-'~······· Attitude: ZENITH 0.1

, 0 .....:_._:~ ~ . -""' ..... I~'.. )

._.•.. ,.I.rene.

, [-­0.01 Drive Test on the 3.60 m Tel.,

1Hz 10Hz M II/to the optical axis

1 000 -:H--~--I------4--+-+-+-~++------+--~---I----+---+--'~+-+----+-

#180 resolution 176 mHz Figure 6

100

N I-,. -~ 10 Sample # 180 E c .OOO~---[--

j .. 100 ••.... _-_._-.•._.... t : ~._ .. __ .

Same, excepted that:

0.1 '0 /.:::::.•....:J . . == :::::J:1Hl) ~+ + ~ - ~, _..;.. _............•..........;... J.IL'PHlIt

Hour-angle rate is: 150 "/s

14 March 89 14:07 UT G4t9n SamplinJ: 360.9Hz

1 i-­

1Hz 10Hz

Tracking in hour-angle at a rate up to 150 "/s does not show any significant changes with respect

to the OFF status in terms of rms fluctuation. Nevertheless, with increased tracking rate, new lines are

visible on the Power Spectrum.

d - CAT: effects measured on the side ofM I. along a direction parallel to the optical axis (see fig. 7)

Average over 3 samples (drives OFF) : 12 om rms Average over 4 samples (drives ON) : 36 om rms

There is a clear increase of the level of mechanical vibrations, mainly for frequencies below 10Hz,

when the hour-angle tracking SPeed is increased.

Spectrum of mechanical fluctuations

1000 i : :

'199 resolution 176 mHz

100

-~ 10

E c

.~··-·-+····1·····!-········"···········-·· .

.. ':~-~---;-~.:' ..•.....--:.... . .0.1

'0 ..-.• -"_C.'Hz) ·_...•• NftCe

1 : t+-.;--+--+.-+....,..----+---r----1~:-+-_+_ii_ri---__+_0.01 '000...

1Hz 10Hz

1000 Tt------f~--+-___t---t--+---t--l!---!!-+;------t---+-----t----tl-.+~O-t~-!,H:H,-----+-

100 N :I: "?>

§ 10

E c

0.1

0.01

111 !r~~ut~~ 176 mHz

.OOO~--~--

·:.:::T·/~ . --""I,,'"'J._.... ,.••,.ft(.

1· l....-- ... ,Oto.

1Hz 10Hz

Sample # 199

15 March 89 15:30 UT Gtt912 Sampling: 160.9Hz

Air conditioning : ON

RolI:ON,O"/s Pitch: ON, 0 "Is

HA: 0"/5

Dome: CLOSED Wind: N 2 mls

Attitude: ZENITH

Drive Test on the CAT,�

M 1 /I to the optical axis�

Figure 7�

Sample # 209

Same, excepted that:

Hour-angle rate is : 50 "Is

15 March 89 14:07 UT 0-912 Sampling 360.9Hz

6) Estimated cumulated mechanical fluctuation along the NTT optical path:

The intention is to derive, from point accelerometric

measurements, an estimate of the contribution of the NIT to optical

path fluctuations as ifit were used for interferometry. (fig. 8)

We used measurements on the primary mirror (Ml), on the

secondary (M2) and at the Nasmyth focus. These measurements

were performed under normal observing conditions; the

fluctuations were assumed to be uncorrelated, which is most

probably true except for the lower frequencies, and therefore were

added quadratically. The result does not incorporate air index

contribution and refers to mechanical fluctuations (not optical path

differences).

NTT

estimated cumulated mechanical

fluctuations�

(above I Hz)�

(under typical oper.ling conditions)�

om nns fluctuations vs expo time 1 OOO+-----'-----'--''-'-'--~_L..--L............................~

100

1 0 ---I· --.. td"emx:e 1---­

Fig. 8 I exposure time

1 +---.----,.-r-r-rTTTT-...---...-r-,...,.-r,m­

lams lOOms 1DOOms

7) Validity of the measurement technique and possible improvements

- the hardware and the software were extensively tested and calibrated ; the overall response

(fluctuation values in om) is estimated to be correct within ± 20 % between 3 and 110 (or 140 Hz) ;

the lower boundary comes from possible slight changes in the high-pass filter; the higher boundaries

take into account the rather low sampling frequency.

- in quiet conditions, measurements were in most cases limited by the system noise up to a few

Hertz (typically 5 Hz).

- single point measurements make it impossible to distinguish between translation and rotation of

the mirror under test; we assumed in this paper that the motion was mostly translation but this needs

further investigation, for example with more accelerometers at proper locations near the mirror under

test.

- the method used neglects the effect of air index fluctuations in the telescope. The lack of phase

information does not allow to add linearly the contributions ofeach mirror. Therefore one would need

laser measurements provided a very stable reference is available. We strongly recommend that this be

considered for further investigations (the drawback is that the set-up for laser measurements takes a lot

more time than point accelerometric measurements).

- when the environment is noisy, as it was often the case at the time of our measurements, an

evaluation of the contribution ofeach source ofvibration is difficult.

CONCLUSIONS

1- Conclusions on wind

We noticed a small increase in ems fluctuations -by a factor of 1 to 3 (computed for a 15 ms

exposure time)- for wind speeds of 5 mls. Wind influence through the outside building and the

telescope foundations may be of importance.

2- Conclusions on drives - we found:

- a small increase in nns fluctuations, by a factor of 1 to 3 (computed for a 15 ms exposure time).

The increase is spread over the 1 to 180 Hz frequency range, especially below 10Hz.

- a clear but not striking change in spectral signature

- a difficulty in distinguishing between drive errors in acceleration (spectral lines at lower

frequencies?) and bearing / gear noises (wide band noise at higher frequencies ?).

3- Conclusions on overall behaviour

See fig. 9 (quiet conditions) & fig. 10 (typical observing conditions) for measurements on M1

(leftcoJumn) & onM2 {right column), fortheNTI (upperrow), for the 3.60m{middJerow)and for

the CAT (lower row).

- no striking differences between telescopes; the 3.60m telescope may be a little quieter than the

NIT ; the impression is that vibmtions are less on secondary mirrors than on primaries.

- the NIT, considered as one component in an interferometer, would probably be above the

provisional specs by a factor of 3 for "normal observing conditions", but the secondary mirror by

--

QUIET CONDITIONS (Fig. 9)

NT[ : MIll optical axis Sample # 106 NTf: M2 /I optical axis Sample #I 23 IO_19·11:l3lTT Gil '_191I,'llTTAir conditionning : OFF Oil suppon : OFF 0<7

Air oonditionning : ON Oil suppon : ON Arimul : OFF Altitude: OFF Sampling: 260.1 Hz

Arimut: OFF Altitude: OFF Sampling: 260.1 HzBuilding: nOSED Wind : N_5 mls Anilude : ZENI1l1

Building: a.oSED Wind : ESE 2 mls Altitude: z "" 45°

SpeclJUm of m«hani~ fluctuations SpeclJUm of mechanical fluctuations

-r+----'----'--..-...~~--!-t_t_+-~-_t_- 1000.........-!__t_~_+_H_---\,_I1000 1106 .023 r.oIution 127 mHz resolution 127 mHz

100 100

~ 10 ~ 10

._~--:-­'10 ._..- .._ ..J-......~~:...

..J.... 0.1 0.1 .. :,..,".~ ..~~..... ,

i-­0.01 0.01 1

, .. t.... ,....­

1Hz 10Hz 100Hz 1Hz 10Hz 100Hz

------------------------------------------~---------------------------------------------~---

360 : M 1 1/ optical axis Sample # 175 360 : M2 /I optical axis Sample # 186 oil boost pump OFF .'_I9·I];56lfT oil boost pump OFF 14~I9-P 09UT

09n G9I1

Air condo : OFF Oil pads : OFF Air:OFF Air condo : OFF Oil pads : OFF Air: OFF

HA:LOCK Dec: LOCK Sampling: 360.9 Hz _ HA:LOCK Dec: LOCK Sampling: 360.9 Hz

Dome: OPEN 45° Wind NNW 3 mls Altitude: z.E.NmI Dome : OPEN 45° Wind N9m1s Anitude: ZE.NITH

SpeclJ'Wll of mechanical fluctuations Sp:cuum of mechanical fluctuations

1000 1000

100 100

~ 10

.•. , .. ----_._.-.._~.;.:.~:.::~.'Ei--:-"­...•~.---

. ' <~~~ ~__C_I�

0.1 0.1 , i _

'... ,00- , .....

0.01 0.01

1Hz 10Hz

CAT: Ml II optical axis Sample # 199 CAT: M2/1 optical axis Sample 11191 t4_19·1O:31lTTtS_I9·IS:JOUT 09...209...2 ....... :]60.9Ha�

Aircond.:ON ......... : -, Ha Air c:ond. : ON�

Roll: ON. 0 "/5 PilCh: ON, 0 "/5 (HA :O"/s) Roll:GUIDE PilCh : GUIDE (HA :50·/5)

Dome : Q.OSED Wind : N 2 ml5 Altitude: ZENmf Dome: OPEN Wtnd : N9m1s Anilude : 2ENfIH

Sp:cuum of mechanical flue:tuationsSpeclJ'Wll of mechanical fluctuations

1000 i i ; iii1000 . i : : 1! ! !i.191

~utio~ 176 mHz

100 i. .! .! ! !.~_! q~J,j~-l-L~J~~_~ -, -+---+--~I i i i-ii ---­100

I ! i II iii I ; I' j!!I I . \ ! I!.! II~~~---

. i ! Iff J..-~ 11/]--- --.. j I '!!:--- t---H-i4··.._~-- . I I . ! i ji; ..... I... -_._----;.-......~.- ':W--/J~'III 0.10.1 .........~. ,. ....,.:~..­

. - __ c·.... ),. _·=·~.... I

, 1-- 0.01 .... ,-- '......... '..... '..... 1Hz 10Hz

1Hz 10Hz

0.01

-----------------------------------------------

TYPICAL OBSERVING CONDITIONS (Fig. 10)

NIT: Mill optical axis

Air condilionning : OFF Oil suppon : ON

Azinwt: OFF Altitude: 13.1 "'5 Building: OPEN Wind : NNW 5 m/s

1000

100

0.1

0.01

1Hz

360: MI /I oplical axis

Air condo : OFF

HA: 150 ·/s

Dome : OPEN 45°

1000

100

SpeclJUm of mechanical fluctuations

Oil pads: ON Air: ON

Dec: O"'s Sampling: 360.9 Hz

Wind N5m1s Altitude: ZENlTI-I

SpeclJUm of mechanical fluctuations

Sample # lIS 10 MMdlI9 ·11;" \TT COl

Sanlpling : Z60.1 Hz

Altitude: ZENInl

: :!!

"15 'esolulion 127 mHz

Sample # 180 .4 MatQ 19 - .4 :29 lIT 0912

~ 10

0.1

0.01

CAT: Ml II optical axis Sample # 209 U_19·1Z:J1lTT 09_2

Air cond. : ON SoIofIioII' MO., 1\&

Roll: ON Pitch: ON HA.a .. SOOts

Dome : a.osED Wind :SW3m1s Altitude: z '"' 20"

SpcclJUm of mechanical fluetualions

I , ! ,;o~ ; !I' .�

i , ! ! NSOluIion 176 mHz�

~

100

tHlt~i~ 10

1'1~ d.

100 .. ~ _~ 10

~

NIT : M2 II optax Sample # 101 IO_I9·~,"1TT

C,#eAir condilionning : ON Oil suppon : ON

Azimut: 1350 "/5 Altitude: 13 "'5 Sampling: 260.1 Hz

Building: Open Wind tHI VI 7 m/s

SpeclIUm of mechanical fllICluations

1000

: : 'i! II ,"0', ,esolution '27 mHz

.. _ ..~. __.)._..~. __ .~ ..l .._~ ...i100 -+~ H1ij-­ i; :

-';--'--+-'+-~-+H \

.._~--~--........ _--_..,-:,..:::::~

i"· 0.1

":f""'.:.::.:.~i__ c.,....

0.01 . .... ,--;--......

1Hz 10Hz 100Hz

............

360 : M211 optical axis

Air condo : ON

HA: IS"'s ->S

Dome : OPEN 4Sdeg

Sample' 185 14_19·16,4I1TT+ oil boost pump 09/2

Oil pads : ON Air:ON

Dec: IS "Is ·>W Sampling: 360.9 Hz

Wind N ~ .../6 Attitude: ZENITI-I

Spccuum of mechanical fluctuations

1000

: '185 resolution 176 mHz

100

0.1

0.01

1Hz 10Hz

CAT: M2/1 optical axis Sample MI91 IC_19·1O,SlITf 09_2 SoIofIioII' MO., 1\&Air ClON1. : ON�

RoU:GUIDE Pildl : GUIDE (HA : SO "Is)�

Dome:OPEN Wind :N9m1s Altitude: ZENI1li�

1000

Spccuum of mec:hIIIical fluetualions

1 iii i i

i'UlI r.olution 176 mHz

+'i..---+--+---r-+-r++++-----i----+-t-+-:-4-+~-H---I III! I \ \ ill \

-ti---+"~rl--+--t-+++-t+----+--+_r_t-+.!~il' . I

I! !' I ! !

! III \ ! iii II

0.1 0.1

0.01 0.01

1Hz 10Hz 1Hz 10Hz

1000

itself would meet our requirements for exposure times> 100 ms, i.e. for wavelength approximately

larger than 3.5 J,lm.

- The vibration level of the tested telescopes -which were not designed for interferometry- is

above, but still close to, the tentative specifications and we can be reasonably optimistic for the future

of the VLTI in that respect.

4- Conclusions on vibration testing methods

Accelerometers are a good choice for an extensive survey of telescope vibrations, but their low

frequency response should be improved. For deeper and more accurate measurements on actual

Optical Path Distances, accelerometers and laser metrology should be combined.

5- Recommendations for engineering

- investigate the respective influence of errors in tracking speed smoothness and of the noise

generated by solid bearings and drive actuators.

- introduce damping in the structures ofbuildings and telescopes

- evaluate the ground transmission ofvibrations induced by the wind on buildings around or in the

vicinity ofa telescope.

- avoid pulsating oil pumps

- compressors, air conditioning fans, etc... should be "far" away

- select carefully the location ofpower supplies for instrumentation and telescope control

Acknowlegements

We thank the European Southern Observatory, 1'Institut National des Sciences de rUnivers

(CNRS) & rObservatoire de la Cote d'Azur for their contributions to the program. The assistance of

the ESO staff on La Silla was extremely valuable for the success of the mission.

We appreciate especially the support of Pierre Lena and Jacques Beckers and the extensive and

fruitful discussions with Bertrand Koehler.

References

1) P. Bourlon, P. Lena. Vibration testing of telescopes and interferometers, NOAO-ESO Conference, March 88,

F. Merkle, Ed.

2) P. Una. Requirements for stability. Note for the VLT~ working group, 23 April 89.

3) B. Koehler, P. Bourlon, M. Dugue. Active vibration filtering: application to an optical delay line for a stellar

interferometer. Proceedings of SPIE Symposium on OFfAerospace Sensing, Orlando, 1-5 April 1991

4) H. Jorck, R. Maurer, J. Kiise, M. Faucherre, J.M. Beckers, G. KUhn, H. Hupe ; The design of the delay lines for the

VLT interferometer. In these proceedings.