VISAR & Vibrometer results

Goran Skoro (University of Sheffield)

UK Neutrino Factory Meeting

Lancaster, April 2009

Part I: VISAR

02

VISAR tests03

VISAR tests have been performed with 0.3 mm diameter tungsten

wire

Idea was to measure the VISAR signal and to extract the longitudinal oscillations of the pulsed wire

Two characteristic results (shots 3 and 5) shown on the left*

Yellow – Current pulse

Green, Purple – VISAR signal (2 channels)

* Note the different time scale

3

Current pulse

VISAR signal

VISAR signal

5

Current pulse

VISAR signal

VISAR signal

Wire Laser beam

VISAR signal obtained for the very first time - nice agreement with simulations

results -

But, noise is an issue here!!!

Analysis shown on the following slides will try to address this problem

04

3

Current pulse

VISAR signal

VISAR signal

Frequency analysis of the VISAR signal

FFT method (MATLAB) has been used

– from time to frequency domain -

~ 16s

~ 60 kHz

Dominant signal frequency clearly seen in frequency spectrum (noise -

negligible)

We need this region too in order to describe the wire motion properly (noise

~ signal)

5

Current pulse

VISAR signal

VISAR signal

05Frequency analysis of the VISAR signal

~ 60 kHz

Dominant signal frequency (not so clearly) seen in frequency spectrum

~ 16s

~ 8s

~ 125 kHz

Real effect (friction of wire’s end) or noise?

Compare with previous plot – looks like a noise

(here: not enough data points for noise)Really powerful method; we need more data from VISAR

FFT method (MATLAB) has been used

– from time to frequency domain -

06Frequency analysis of the VISAR signal - Filtering

A few examples how we can filter the data

Filter set 2

2

Filter set 1

1

07

Frequency analysis of the VISAR signal and LS-DYNA results

~ 60 kHz

Dominant frequency that corresponds to longitudinal motion of the wire is clearly present in both frequency spectra

~ 60 kHzshot 3

VISAR tests with a shorter wire08

New tests have been performed with a shorter wire

Current pulse

VISAR signal

VISAR signal

Wire Laser beamDifference between green and

purple

Current pulse

VISAR signal

VISAR signal

Difference between green and purple

Current pulse

VISAR signal

VISAR signal

Difference between green and purple

A few characteristic results

First conclusion: no signal here!

But, interesting ‘coincidence’ in frequency spectrum…

09

FFT analysis of each shot – no signal seen (expected, if we look at

previous slide)

Small statistics – so we can say it’s a coincidence (but will be interesting to collect more data)

Frequency analysis of the VISAR signal

Tests with a shorter wire

But, by averaging the frequency spectra, the ‘structure’ starting to appear exactly at the right position

We could expect to see this dominant frequency (~ 80 kHz)

Experiment

averaged

VISAR tests – data collection10

VISAR tests have been performed with 0.5 mm diameter tungsten

wire

More than 130 shots have been taken between 17 and 19 March

The measurements have been performed at different temperatures as

a function of applied current.

Wire Laser beam

First wire has bent during the test at 1750 C. Only single shot at this

temperature has been recorded (shown on the left)

Current pulse

VISAR signal

VISAR signal

Difference between green and purple

1750 C

Wire has been replaced and measurements continued – excellent

consistency between two sets of data!

Each temperature (and current value) – between 5 and 20 shots (FFT analysis of

each shot then averaging)

Scripts* have been written for a quick analysis of such a large ammount of

data

*Particular thanks to Jelena Ilic for her help.

VISAR tests – Results (Part I)11

Frequencies seen in spectra:~ 20 kHz;~ 40 kHz;~ 60 kHz;~ 80 kHz;~ 100 kHz;

~ 140 kHz (weak).

Peak current = 7.5 kA except for T = 850 C where it was 8.1

kA

- as a function of temperature -

We expect* to see 80 kHz as a result of wire thermal expansion

What about the others?

Answer (part I):20, 60, 100, 140 kHz are the

fundamental frequencies of the wire vibrations

*See: FFT_Visar_ver2.ppt (Slide 9)

l

cnfn 4

)12(

C (for tungsten) ~ 4 km/s

L (wire full length) ~ 5 cm

f0 = 20 kHz, f1 = 60 kHz, f2 = 100 kHz,f3 = 140 kHz…

12VISAR tests – Results (Part III)

- as a function of current (for T = 1100 C) -

Only the magnitude of the 80 kHz spectral structure increases with

increasing current!!!

Legend:

1 = ff (1st harmonic)2 = ff (2nd harmonic)3 = ff (3rd harmonic)4 = ff (4th harmonic)

b = bending frequency

Arrow = 80 kHz (thermal expansion)

13VISAR tests – Results (Part IV)

- as a function of current (for T = 1300 C) -

Again, higher current –> higher magnitude of the 80 kHz spectral

structure!!!

Legend:

1 = ff (1st harmonic)2 = ff (2nd harmonic)3 = ff (3rd harmonic)4 = ff (4th harmonic)

b = bending frequency

Arrow = 80 kHz (thermal expansion)

80 kHz thermal expansion

14

P R E

L I

M I

N A

R Y

- Young’s modulus of tungsten wire -

)1/()21()1()2( 2 lfE- density; – Poisson’s ratio; l – length of wire (heated part)f – measured frequency

Errors estimated from the widths of the characteristic spectral ‘line’.

VISAR tests

Part II: Vibrometer

15

Vibrometer tests16

Laser Doppler Vibrometer tests have been performed with 0.5 mm diameter tungsten

wire

Only a few shots taken - shown on the left and below*

Yellow – Current pulse

Green, Purple – “Displacement”, Velocity

* Note the different time scale

6

Current pulse

“Displacement”

Velocity

Wire

Laser beam

But, noise is an issue here!!!

5Current pulse

“Displacement”

Velocity

4

Current pulse

“Displacement”

Velocity

17Frequency analysis of the Vibrometer signal

FFT method (MATLAB) has been used to analyse velocity signal

– here: short time scale -

~ 1s

~ 1 MHz

Current pulse (and reflections) Radial oscillations of the wire

~ 160 ns

~ 6 MHz

We expect to see radial oscillations

6

Current pulse

“Displacement”

Velocity

18Frequency analysis of the Vibrometer signal

FFT method (MATLAB) has been used to analyse velocity signal

– here: medium time scale -

~ 150 -300 kHz

Expected frequency of longitudinal oscillations should be in this region

(even lower)

We ‘see’ radial oscillations of the wire

~ 6 MHz

We expect to see ‘longitudinal oscillations’

4

Current pulse

“Displacement”

Velocity

Very noisy here – is the previous results (see slide 3) in this frequency range

realistic?

19Frequency analysis of the Vibrometer signal

FFT method (MATLAB) has been used to analyse velocity signal

– here: long time scale -

We expect to see violin modes of wire oscillations They are here

5

Current pulse

“Displacement”

Velocity

20Frequency analysis of the Vibrometer signal - Filtering

What is ‘wrong’ with these pictures?

~ 6 MHz~ 6 MHz

6 MHz signal seen in both frequency spectra

Velocity decoder VD02 has been used to extract the surface velocity of the wire- but it’s upper frequency limit is 1.5 MHz -

To answer this question we need a decoder which works in higher frequency regime (DD300)

21Frequency analysis of the Vibrometer signal and LS-DYNA

results

~ 6 MHz

~ 6 MHz

Radial oscillations of the wire!!!

VISAR – noisy but a decent level of information can be extracted (FFT, filtering, …)

Vibrometer has been purchased (just arrived to RAL)

More results soon…