1. First results with new stave support prototypes
26/06/2014
François-Xavier Nuiry
Fernando Duarte Ramos
Wolfgang Klempt
Kyle Spindelman
François Boyer

2. Prototyping short history
Mid March we got simulations results and X0 estimates for new full sandwich staves made of very thin prepregs.
Simulations were showing at least same stiffness as current prototypes
X0 estimates revealed that 0.05%X0 objective was reachable with this new material
8 staves have been ordered the 24/03/2014 to North Thin Ply Technology:
4 staves with nomex honeycomb (C ) core
4 staves with PMI (polymethacrylimide) Rohacell (IG-F51) core
Glue film
As thin as possible max 30gsm
Skins:
High modulus fibers
M55JB 30gsm – 37 micron thick
Core: Nida Nomex or Rohacell
Glue layer present on ~200% of the stave surface

3. Prototyping short history
The 30th of May, we received:
6 staves with nomex honeycomb core (including 2 lower quality)
6 staves with Rohacell core (including 2 lower quality)
Several samples of cured M55JB skins

4. Reception tests
We started reception tests and mechanical characterisation in June:
Stave dimensions measurements
Stave mass measurements (even after drying)
Stave X0 estimates based on manufacturer data
Stave X0 measurements
Flatness measurements
Stave flexural stiffness measurements

5. Features Stave label Proto 1 Nomex Proto 2 Proto 3 Proto 4 Proto 5
Rohacell
Proto 6
Proto 7
Proto 8
Mass (g)
1.53g
1.58g
1.78g
1.65g
2.07g
1.97g
2.05g
2.11g
Mass after drying 96h at 3%RH
1.52g
1.57g
1.75g
1.64g -
Default observed
Local skin buckling on both skins.
Stave not very flat.
Extremities a bit crushed are slightly compressed.
Small holes on skins.
Very minor defaults:
Very small delimitation at extremities
Small shearing on ends

6. X0 estimates
Radiation length evaluation (based on measures and manufacturer’s data)
PROTO 1
Nomex
PROTO 2
PROTO 3
PROTO 4
PROTO 5
Rohacell
PROTO 6
PROTO 7
PROTO 8
Both Skins
0.0271
0.0279
0.0309
0.0313
0.0252
0.0221
0.0267
0.0233
Core
0.0145
0.0146
0.0143
0.0268
0.0260
Both Glue layers
0.0155
0.0126
0.0166
0.0120
0.0109
0.0069
0.0092
Total %
0.057
0.0568
0.0581
0.0622
0.0640
0.0597
0.0603
0.0584
After some discussions with the supplier it appears that there way to measure material thicknesses includes some inaccuracies.
Besides, Rohacell was considered as PMMA whereas it is in reality PMI (polymethacrylimide).
Still waiting for the exact formula of the glue film and prepreg matrix.
(C5O2H8)n
(C8H10O2N)x
X0=40.8g/cm2
X0=42.12g/cm2
X0=7846mm
X0=8100mm

7. X0 measurements X ray counter Beam Beam Source
Thanks to Hideyuki Oide CERN and Mauricio Sciveres (LBL)
Concept: measuring X-ray absorption by staves.
We employed a fully-depleted Si-PIN X-ray counter X-123 with a 500 μm-thick, 6 mm2 sensor produced by Amptek[4]
We use a Cd109 (4354RP) source with an activity of 3.1MBq (19/06/2014)
Tests currently performed in SR1 (Bd ATLAS).
X ray counter
Beam
Beam
Source

8. sample object with X-123 detector
X0 measurements
Process:
Measure of the amount of detected photons over a certain period. No Stave on the beam trajectory.  N0
Measure of the amount of detected photons over a similar period with the stave located on the beam trajectory. N
We get the X0 of the full stave sandwich thanks to the following law:
N/N0= e-X0
22keV X rays
25keV X rays
A narrow beam of monoenergetic photons with an incident intensity Io, penetrating a layer of material with mass thickness x and density ρ, emerges with intensity I given by the exponential attenuation law :
#entries/ch/sec
x = material mass thickness
ρ = material density
µ/ρ = attenuation coefficient
I0 = photon incident intensity
I = photon intensity after the material sample
MCA Channels
Example of the measurement of absorption of 22 keV and 25 keV X-rays from 109Cd by a
sample object with X-123 detector

9. X0 measurements - uncertainty -
Expectation:
Stave is thin and made of very light materials. 0.05%X0 expected.
The photon absorption will be small  N/N0 close to 1.
Measurements could be long…
𝑁 𝑁 0 = 𝑒 − 𝑋 0
𝑋 0 =− ln 𝑁 𝑁 0 =−ln(𝑟) 𝑟= 𝑁 𝑁 0 ~1
𝑑 𝑋 0 = 𝑑 𝑋 0 𝑑𝑟 .𝑑𝑟= 1 𝑟 .𝑑𝑟
𝑑𝑟=𝑟. 𝑑 𝑋 0 =𝑟. 𝑋 0 . 𝑑 𝑋 0 𝑋 0 ~ 𝑋 0 . 𝑑 𝑋 0 𝑋 0
𝑋 0 =0.0005
𝑑 𝑋 0 𝑋 0 = =0.2
𝑑𝑟~1. 10 −4
𝑁 𝑁 ~1. 10 −4  𝑁~
Attenuation law 
Attenuation law 
given
Targeted X0
Proposed uncertainty on the X020%
Amount of photons to be counted:
Photon beam counting statistical error 
Photon rate ~ 4.72/s  245 days of measurements…

10. Current measurements and getting faster results
X0 measurements
Current measurements and getting faster results
 About 144 hours of measurement already taken.
~ photons detected.
Photon absorption in one rohacell stave close to ~
To be continued…
Ongoing studies:
Trying to measure 4 staves in one go.
Hypothesis: staves are very similar. Problem: X0 measurement for 1 stave not got.
- Trying the tests with higher activity sources.

11. Stave flatness measurements
A side
B side
The presence of different materials types in the assembly sandwich (glass and alu) may explain the flatness values measured. An alternative process is under study, also at CERN.

12. Stave flexural stiffness measurements
Objectives:
Comparing stave stiffness versus design, geometrical imperfections, etc…
Comparing results with FE simulations.
Getting an idea of the stave natural frequency.
Standard used: ASTM D790-02
Configuration:
Loading nose and supports radius: 5mm
Support span : 260mm
Loading nose speed: 57.4mm/min
Test stopped when 0.3N are reached for honeycomb staves and 0.8N for rohacell staves
Polymer lab facility
Load cell:
Capacity: 100N Accuracy of 1% of applied load (~+/-0.025N)
Position measurement accuracy:
Not well known (0.001mm ! Announced by the manufacturer…)

13. Stave flexural stiffness measurements
Results
Honeycomb staves more than 2 times softer than rohacell ones. (Not expected).
Possible reasons:
Honeycomb staves are about 5% thinner than rohacell ones.
Honeycomb stave skins are already locally buckling (60 microns waves, skin not flat).
Staves are not very flat.
 Rohacell staves flexural stiffness values are the one expected since the beginning. But it does not fully correspond the FEA results if stave exact dimensions are updated.

14. Stave flexural stiffness calculations
Shell element for skins.
Solid elements for the core.
Glue layer not modelled.

15. Stave flexural stiffness calculations
Real geometry and materials - STAVE 5
Stave core Thickness impact (0.1mm)
Stave skin Thickness impact (0.03mm-0.037mm) (same total thickness)
Stave rohacell modulus impact (30 Mpa - 70 MPa)
Geometry
Full sandwich
Skin Mat
M55J (TPT) 2*1 layers 0°
Core Mat
Rohacell
BC UX=0 on one support
yes
Lesize
0.6
0.600
Stave width (mm) 26
25.937
Skin thickness (mm)
0.03
0.037
Core thickness (mm) 2
2.196
2.1
Span (mm)
260
Displacement (mm)
Core E modulus Mpa 30
70.000 70
Core G modulus Mpa
11.54
26.92
Reaction force in the center (N)
1.1741
0.8859
Flexural stiffness (N/mm)
1.24
1.96
1.36
1.48
1.34
Bending stiffness (EI) N.mm^2
5.24E+05
7.80E+05
5.77E+05
6.41E+05
Theoretical Bending stiffness (EI) N.mm^2
5.22E+05
7.78E+05
5.75E+05
6.49E+05
approx Bending stiffness (EI) N.mm^3
FULL Bending stiffness (EI) N.mm^4
X0 estimate % (Ansys)
0.0488
0.0568
0.0497
0.054
0.0484
X0 estimate % (Ansys + glue(2*20microns))
0.0585
Difference
57.70%
9.37%
18.99%
8.23%
Values we expected
Values we should measure (if the model is correct…)

16. Conclusion
Ultralight CF stave have been prototyped in collaboration with an external company.
It seems to be possible to build full sandwich staves (very simple design) with a radiation length very close to 0.05%.
The production quality has to be improved.
Rohacell staves have an apparent quality and robustness acceptable. The flatness needs to be improved.
Honeycomb staves are demonstrating that it is possible to glue 37µm skins on nomex.
But the quality has to be improved:
- The stiffness is under the expectations
- The flatness is not acceptable
- Stave extremities can be better.
It is proposed to re-launch a serie of 8 staves with a new assembly process CHF.
The objective is twice:
-Getting flatter staves
-Improving the reproducibility
It is also proposed to make such stave at CERN. Normally we know all the process. Prepregs and glue film have to be ordered.

17. Staves prototyping stage 1 2mm sandwich staves
SANDWICH SHAPE (2mm core thickness)
SANDWICH SHAPE (1.8mm max thickness)
Composite design
Design
Skin
1 lay. M55J per skin (0°)
3lay. T800 of 30 µm per skin
Core
Nida nomex
2mm thick
Rohacell (E~30MPa)
1.74mm thick
Rohacell
Mass with glue g
1.5g
1.44g
1.32g
3.17g
Radiation length
0.022%
0.064%
0.014%
0.025%
0.012%
0.020%
Glue
(40 µm* 200%)
0.019%
(20 µm* 200%)
0.010%
0.042%
Total
0.055%
0.057%
0.053%
0.054%
0.126%
Flexural stiffness
N/mm
3.45N/mm
3.26N/mm
2.69N/mm
2.56N/mm
2.12N/mm
Bending stiffness
N.mm2
5.21*105 N.mm2
3.97*105 N.mm2
3.5*105 N.mm2
Approx. natural frequency
Clamped Hz
~203Hz
With modules
~178Hz
~180Hz
~150Hz