Service Task: Trasport boxes for Strips Modules

Member: Francesco Peri

Institute: HU Berlin

Qualification Project: Upgrade

Local Supervisor: Heiko Lacker

Technical Supervisor: Craig Sawyer

Project Description: Systematic characterization, evaluation and optimisation of transport boxes for modules and their packaging options (box-in-a-box, different dampening material, maybe Euro palette...). Quantification of limits concerning transport damage caused wrt accelerations, vibrations, temperature and humidity variations. Technical supervisor: Craig Sawyer (RAL) Local supervisor: Heiko Lacker (HU Berlin)

About this twiki:

The twiki is organized in the following way: introduction to the task, main part (Transport box design, air tightness tests, crash tests), summary of the current status, additonal information.

Introduction:

Starting from 2024 the LHC will provide pp collisions with a Luminosity much higher than the current level. This fact, together with the accumulated radiation damage, makes the replacement of the Inner Detector mandatory. New sensors will replace the old ones, guaranteering a much higher granularity and a stronger radiation hardness. The new Strip System consists of a Barrel and 2 Endcap parts, covering in total a pseudorapidity region up to +-2.5 eta. The Barrel consists of 5 cylinders (plus a "stub" barrel layer) made of 472 stave, each ones with 26 modules. The Endcaps have 5 disks on each side, populated with 32 petals, each one made of 9 modules (the Barrel Stave and the Endcap Petal are represented in Figure 1a,b). A module is defined as the basic electrical unit of the stave (or the petal). ABC130 chips are mounted on kapton circuits to make a hybrid. Up to 4 hybrids can be mounted on a silicon sensor to form the module. Each Module is connected by data input and output lines, low voltage, high voltage and slow control lines to an End of Stave (EoS) card.

Figure 1a (Barrel and Encap Strip Detector).

Figure 1b (Barrel Stave and Encap Petal).

Sensors:

Frames:

During the development phase it is fundamental to properly test all the modules. This is done using Module Test Frames which provide the connections between the hybrids, DAQ and Power supplies. Different frames must be used for different sized modules. In particular 2 frames are going to be used for testing and transport. They are rdescribed in the following links ( R0 V2.0:, R5 H0/H1 V2.0:). The first one can to transport only 1 module, while the second one has space for two of them. R0,R1 and R2 modules are going to be transported either with the first kind of frame or in pairs on the second frame. R3,R4 and R5 modules must be transported on the big frame.

Transportation

The final packaging needs to provide an air-tight environment to avoid moisture and ice formation on the silicon sensor. Moreover it must be strong enough to protect the modules and the frame during the traveling. Easy access to the frame would also be preferable to provide powering and HV.

At the moment the most plausible solution is the following. A rigid box containing the frame and the sensor will be the first barrier against shocks and damage. This box will be then placed in 1 or 2 cardboards filled with an appropriate packaging material.

Transport box design

This paragraph focuses on possible designs of the internal rigid box. This box must be strong enough to protect the frame, keeping the sensor in position.

A first solution follows the idea proposed by Craig Sawyer (see Fig. 3) for the transport box of the Barrel Modules. In this case the module is fixed on the frame using a 3D printed plastic support (in white) and 2 layers of plexiglass are then fixed under the frame and over the 3D printed support with 4 screws. The white FrameSpacer locks the modules along the z axis, while they are kept in position on the xy plane thanks to a small (ca 0.3mm) step created on the test frame itself.

Figure 3. Design of the transport box for Barrel Modules

Following this idea we have designed a similar transport box for the Endcap sensors. This also requires a redesign of the test frame, since the current Endcap test frame is completely plain. In the new test frame the area used to allocate the silicon sensor has a little step with respect to the rest of the frame. In this way the sensor is locked in the xy directions. The following picture is a 3D representation of the new desing (left), together with an aluminum prototype (center) and the orginal one (right). The blue part is the slightly hollowed part where the sensor goes.

Fig. 3b Modified Endcap test frame (left). The central area has a little step to fix the silicon sensor in the xy directions. Original test frame (right).

To fix the sensor in the z direction a frame spacer must be fixed on top of the frame. Four pins on the edges of the sensor will keep it in position (see Fig. 5).

Fig. 3c 3D model of the Endcap Framespacer (top). 3D printed version on top of the original test frame (bottom left), 3D printed version on top of the new test frame together with the module (bottom center), zoom on the fixing pin (bottom right).

On the top and the bottom of the structure, 2 plexiglass layers are placed. The following picture shows the final assembly.

Fig. 3d 3D assembly of the box (left) and picture of the first complete prototype (right).

The Inventor files are available in the attachments: endcap_test_frame, endcap_framespacer, frame_top, frame_bottom.

A possible alternative to this, is to use 3D printed pins fixed on the frames using small magnets. This pins can be properly shaped to fit the corners of the module, fixing it in 3 directions without the need of the 0.3mm step on the frame. The following pictures show the design applied on a R0 module.

Fig. 4a Magnetic pins design for the transport box applied on a R0 module.

The advantage of this design is that we don't need to have different frames depending on the module, just the position of the magnets need to change. The following pictures show the R1 and R2 modules fixed with magnetic pins.

Fig. 4b Magnetic pins design for the transport box applied on R1 and R2 modules.

The inventor files are available at: endcap_mag_test_frame( v1, v2), magnetic_pins, magnetic_framespacer (also in this case a framespacer is necessary to prevent the upper surface of the module to be touched), frame_top, frame_bottom.

Air tightness tests

This paragraph focuses on testing possible solutions to prevent the formation of moisture on the silicon sensors.

The easiest solution to avoid moisture is to use moisture barrier bags. Unfortunately this means that the transport box must be sealed inside the barrier bags but, as previously said, an easy access to the frame would be preferable to provide powering and HV. For this reason we decided to test a couple of alternative solutions. The idea was to use commercial boxes together with a certain amount of silica gel and investigate if we could mitigate the moisture problem without having to completely seal the frame. In a perfect air tight box, after reaching thermal equilibrium, it is expected to measure a constant humidity percentage, i.e. a constant dew point. Therefore, to test the airtightness of the boxes, temperature and humidity sensors have been placed inside the boxes and then they have been placed inside a high humidity environment created with a climate chamber.

Instrumentation

Humidity and Temperature sensors:

• low storage: testo-174H-Mini.
• high storage: testo-175-H1.

• ISI_HPL_834
• max-grip-003

Testing

Different tests have been performed. In the following lines we summarize the most relevant results.

First of all we wanted to get a feeling on how rapid was the raising in terms of moisture when the boxes were simply placed in a high humidity environment without silica gel. The two boxes ( ISI_HPL_834, max-grip-003) were placed in the climate chamber at a constant temperature of 30°C and 90% humidity for ca. 3 hours. For both of them it's possible to see a progressive increase in terms of humidity and dew point inside the box, even after the temperature reached the equilibrium. The measured data are available in the dew_point_excel (sheets 1,2 for the ISI_HPL_834 box, 1b,2b for the max-grip-003). In both cases we can see the dew point increasing of ca. 2°C per hour.

A second test has been done using a Static Shielding Ziplock bag ( ANT013SSB) and repeating the measure. The advantage of this kind of bags is that they are not completely sealed, but a zip lock allows to easily open them. Unfortunately also in this case the dew point grew constantly for the entire duration of the measurement (see dew_point_excel sheets 3)

To mitigate the rising in terms of dew point a third test has been performed inserting the shielding bag inside the box and adding silica gel both inside the box and inside the bag (20 silica bags in total). The climate chamber has been programmed to keep 30°C and 90% humidity for 23 hours and then to cool down to -10°C and 0% humidity in 1 hour. The cycle has been repeated 3 times. The results are summarized in dew_point_excel sheets 4,5. Sheet 4 corresponds to the inside of the box (10 silica bags inside) while sheet 5 corresponds to the inside of the bag (other 10 silica bags inside). It is possible to notice how in this case the dew point stayed constant when the temperature was constant.

The 4th test performed was similar to the previous one: 48 hours at 30°C and 90% humidity. Results in Sheets 6,7. Again the dried silica gel seems quite powerful in mitigating the humidity problem. Longer tests are required to understand which is the proper amount of silica gel to be used (1 week?) and to verify that after cooling down there is no ice inside box/bag.

The 3 results obtained with the max-grip-003 box are reported also in the following picture. The temperature and dew point are reported in the 3 cases: box (left), box with silica gel inside (center), box with the shielding bag and silica inside (right).

Fig. 5 Dew point and temperature variation when the max-grip-003 box was placed inside the climate chamber (left - 1 minute sampling), when silica gel was added inside the box (center - 5 minutes sampling), when the ziplock bag was added inside the box together with silica gel (right - 5 minutes sampling).

In the Dew_Point2_Excel it is possibile to find more results about this tests. We repeated the test putting the box and the ziplock bag in the climate chamber for 6 hours without silica gel, for 6 hours with the silica gel and for 2 days with the silica gel. Here we report just the last couple of plots correspoding to the 2 days measurement (left - bag, right - box). Again the silica gel performs quite well keeping the humidity level and the dew Point constant and low. The 2 days test has been repeted twice. In the Excel file you can find also the Information about the second measurement. The box seems to perform always better with respect to the ziplock bag.

Fig. 6 Dew point and temperature variation when the ziplock bag(left) and max-grip-003 box (right) were placed inside the climate chamber for 2 days with silica gel inside.

Still the most powerful and safe alternative is to use a Moisture Barrier Bag (http://www.farnell.com/datasheets/1485137.pdf, http://www.alibaba.com/product-detail/Zip-lock-moisture-barrier-bag-from_60376819637.html?spm=a2700.7724857.29.3.NB8lwq&s=p) where the frame and sensors would be allocated before packaging. This would obviously require every laboratory to buy a proper instrument to seal the barrier bags but the final result would be much better.

Crash Tests

This paragraph focuses on how to mitigate shock damage to the silicon sensor and frame. The idea is to test different packaging methods to find out which is the most appropriate one to prevent the sensor to break or to slip out from the proper position.

Since at the moment the silicon sensors are still in production phase, we decided to start characterizing the typical acceleration which can be suffered while moving and transporting objects with a similar size and weight of our transport box.

Instrumentation

Packaging material:

Bubble wrap:

Air cushions:

Studded foam:

Pesky polystyrene peanuts:

Instapak:

Accelerometer: pce-vd3.

Testing

Different tests have been performed. In the following lines we summarize the most relevant results.

The easiest way to test a packaging is to drop it from different heights, measuring the different accelerations involved. To prevent the original frame from being damaged, we decided to use an alluminum plate of similar shape and weight, with the accelerometer fixed on top of it:

Barrel box (down/top plate+frame+spacer) weight: ca. 430g

Aluminum test frame: ca. 560g

Fig.7 Alluminum plate used in place of the original test frame

First of all we decided to perform a couple of quick tests to get a feeling on how the different materials were performing.

For the first test the aluminium plate, together with the accelerometer has been dropped from different heights, without any protection and then with a simple packaging (pesky polystyrene peanuts+bubble wrap+cardboard).

Fig.8 First test performed: pesky polystyrene peanuts+bubble wrap+cardboard

The results are reported in sheets 1 and 2 of the crash_excel. It is possible to see how the added protection does not help too much in increasing the height before reaching the 16g (maximum acceleration per axis readable with the current accelerometer).

Another test has been performed placing the accelerometer and the plate inside the max-grip-003 box, together with some foam. The box has then been placed in a cardboard, bubble wrapped. Further protection has been added using polystyrene peanuts or additional foam (2 distinct tests).

Fig.9 Second test performed: max-grip-003 box (with foam) in a cardboard with additional peanuts or foam.

Results are summarized in excel crash_excel sheets 3,4. No rilevant improvement. It is possible that the box is too light and therefore the deceleration too rapid. Increasing the weight could help.

Starting with the idea that the lightness of the box could be the problem, different tests have been performed, comparing the acceleration felt by the instrument when attached to the alluminium plate (ca. 500g) and when attached to a steel plate (ca. 5 kg). Also different amounts of packaging material have been used.

First of all we decided to test different amounts of the same material with the light plate. We used the following 3 packaging solutions: Foam, polystyrene peanuts and air cushions. 5 cm and 15 centimeter of material have been used to absorbe the impact.

The following pictures compare the results (left=5 cm of packaging material, right=15 cm of material). There are 5 buches of oscillations, corresponding to 10,20,30,40,50 cm. For each height an average of 5 drops has been performed.

Foam:

Polystyrene peanuts:

Air cushions:

No relevant improvement can be seen.

Then the same test has been performed with the 5kg steel plate. Again the following figures summarize the results (left=15 cm of pakaging material, alluminium plate;right=15 cm of pakaging material, steel plate). In this case it's not been possible to test the air cushions since they would explode due to the additional weight.

Foam:

Polysterene peanuts:

In both cases it is possible to see an immediate improvement. In particular it is also possible to notice how the foam performs much better than the peanuts and keep the acceleration under the maximum of 16g in all cases. Even at 50 cm the acceleration stays lower than 13g. Since the result obtained with the foam was particularly good, we decided to repeat the test as a double verification. In the following Figure we report the results we got with the 2 tests performed with the 5kg steel plate and ca. 15 cm of foam.

It is possible to see how the results show the same trend. Again the maximum acceleration stays below 13g.

Similar tests have been performed also using the InstaPak packaging material. This is an automatic growing foam which takes the shape of the object you want to protect ( https://www.youtube.com/watch?v=8-tJRSnGN40). The following Images Show the results obtained with the alluminum and the steel plates using about 15 cm of protection in the vertical direction. No real improvement. The performances in terms of deceleration of the plate seems quite bad, in particular if compared to the foam.

All the results, together with the original files and the corresponding excel version can be found in Al_Steel_Comparison.zip.

Since the good results obtained with the studded foam, we decided to proceed ordering more of it. It was not possible to buy the exact same type that we used for the previous tests. The one we bought is the http://www.mercateo.com/p/103-611183/Noppenschaumplatte_Staerke_40_mm_LxB_600x400_mm_VE_26_Stk.html .

Before proceeding with further tests we decided to verify that the different material could provide the same results in Terms of deceleration. The following Images present the results we got with dropping tests similar to the ones above (ca. 15 cm of foam, steel plate, drop from 10 to 50 cm).

Similar results are observed. Again no acceleration above 13 g.

box in a box tests

To allow a further layer of protection we decided to tests box in a box solutions. The idea was to have a very small and compact box in the inner part, to guarantee immediate protection to the sensor, plus a big box filled with foam as the outside protection.

The first attempt was done using a cardboard in a cardboard solution, both of them filled with studded foam. The small inner cardboard with dimensions of 24x25x13 cm, the big external one with dimensions of 60x40x35 cm. The following 2 Pictures compare the results obtained when we used the box in a box solution (right) and when the plate was simply placed in the big box filled with foam (left).

As you can see it seems that a direct Impact with a bigger amount of foam guarantee a better protection than a box in a box solution when there is only a little amount of foam as the first layer of protection.

Seen These results we decided to give a second chance to the instapak material. In this case the small box was filled with instapak, while the outside one was again filled with foam. The 2 pictures below Show a comparison between the results obtained when the 0.5kg plate (aluminum) was place in the inner box, with respect to the case when the 5kg steel plate was used.

Again the right case (steel) is clearly better than the left one, but still the results are worse than when we simply package the plate inside the big box completely filled with foam.

So it seems that a direct impact on a big amount of foam gives much better protection than any other case. Following this idea, we decided to put plate and accelerometer inside the max-grip-003 box, filled with a tiny layer of foam, just to guarantee that the plate was well fixed and that the Impact was mostly transmitted to the outside foam the big cardboard was filled with. Following picture shows the result.

There is an evident improvement with respect to any other box in a box solution. The 16g limit is never reached. All box in a box test data are included in boxinbox.zip .

So the way to go seems to be make the box heavier and trasmit the impact to the biggest amount of foam possible. Of course placing the naked Frame inside a cardboard is not a good idea, having an inner protection against Shocks is anyway preferable. Since the maxgrip box showed as quite resistent and good in the moisture resistence tests, we decided to use it as the first hard protection where we would place the Frame (or the accelerometer in our case). To make it heavier we attached the steel plate on the bottom of the box itself (see Picture below) and then we fixed the accelerometer in the box with screws.

The following tests show the results obtained with different configuration of foam packaging and weights using the 60x40x35 cm cardboard. First of all we compared 2 different weights (3.5 (left) and 6 kg (right)) placing the maxgrip box on top of the usual 15 cm of foam and performing vertical droppings. Again it seems that the heavier the better but the results are worse then the ones obtained with just the naxed plate on the top of the foam.

This effect could be due to the fact that the box is larger than the single steel plate and therefore the pressure is distributed on a larger surface. In this way the foam can apply a larger resistence to the box decelerating it faster. With this idea we decided to make the foam softer. This can be done creating a sort of web structure instead of using solid layers of foam (see Picture below).

The following Picture Show the result obtained with this Kind of web foam structure. There is a clear improvement with respect to the previous case. We are back to the naked plate Performances.

We also tried to understand if Fixing the accelerometer (or the frame) is better than letting it free but protected with foam or bubblw wrap within the maxgrip box. The following Pictures Show the results obtained in the 2 cases. No real different with respect to Fixing it to the box. The foam plot shows only the last 3 heights (30,40,50 cm).

Dampers

An additional layer of protection could be added with the usage of dampers. There are many kind of dampers available (see: http://www.mcmaster.com/#helical-vibration-isolators/=12nfiha or http://de.aliexpress.com/item/AC1420-1-Pneumatic-hydraulic-shock-absorber-damper-damper-AC1420-Specifications-M14-1-5-High-speed-and/32626787658.html ). We decided to produce one of the first kind. The basic idea is to fix the box on top of a springy structure which can absorb the shocks and vibrations. We have two designs ready at the moment (see pictures below):

The rigid box can be placed on top of 1 of the left kind or 2 of the right kind. Inventor files available: damper_large, damper_long.

The following picture shows the damper with the rigid box on the top.

Of course the dampers must be fixed on a rigid structure. The following is a first possible design of a rectangular frame to be inserted in a cardboard.

A first prototype for a Frame has been created (see following Picture). A weight of approximately 500g has been placed in between 2 dampers. The cable used is simply a 3 conductors cable which showd interesting springy properties in previous tests.

Unfortunately the shock tests do not seem very promising. The following plot shows the acceleration measured when dropping the package from the usual heights 10-20-30-40-50 cm. As you can easily notice the 16g limit of the accelerometer is reached already at 10cm. The foam still seems the best option.

Up to this points results didn't seem too promising. Most probably due to the cable. For this reason we decided to use an elastic cable (bungee shock cable) and repeat the tests. Below you can see the kind of cable we tested ()3mm Diameter) and how it was fixed to the frame and the box (500 grams aluminum box).

The results of the drop tests from the usual 5 heights, (10 to 50 cms) are reported in the following plot.

The plot Shows promising results. The acceleration stays always low, similarly to the best foam case. Of course there are some disadvantages in this case. First there is quite a lot of movement. The box is almost free to move in the lateral direction, and it could possibly hit the internal surface of the cardboard. Second there could be Problems due to the resonance frequency of the bonding, if this is Close to the oscillatory frequency of the box during Transport. As already said the cable tested had a Diameter of 3 mm. Below you can see 3 different tests performed with 3,4 and 5 mm sizes, dropping the box from 40cm.

If 3 and 4mm perform almost the same (the 4mm seems slighlty worse), the 5 mm cable is consistently worse. This is due to the extreme elastic force combined to the fact that the box itself is quite light, i.e. it's immediately stopped by the cable if this is too strong. To avoid having to increase the weight of the box too much, the smaller cable seems the proper solution.

Now let's come to the problems. The following plot shows the acceleration felt by the box when dropping it from 40cm, in the lateral direction, where there is much more space for movement. As you can see the box oscillate for a much longer time, even if the acceleration itself stays decently low.

A first possibility to solve this problem could be to add an additional force in the lateral direction. We achieved this connecting the box to the edges of the frame with some additional bungee cable. See next picture.

Using this configuration we performed the usual drop tests (10 to 50 cms) in the lateral direction. Results showed in the following plot show accelerations similar to the previous ones, slighlty lower oscillations, but we consistently reduce the risk of hard inpacts on the external container.

Also a smaller frame has been built, to allow some space for additional foam between frame and cardboard. We tested this frame using both vertical and lateral support and making the box slighlty heavier (ca. 700 grams), with the 3mm cable. Results reported in the next plot show vertical and lateral drop tests (first group and second group of 15 drops each from 10 to 50cms). We get the best results ever achieved in vertical drops.

Following suggestions in the group, we decided to perform a simulation, using Al wire bonds, to estimate their resonance frequency, so to be sure that the oscillation of the box couldn't cause problems to the frame and sensors to be shipped. The following pictures show the results obtained for the first 6 eingenfrequency. The values are 4982,13236,13892,16426,27080,39954 Hz. They are all very far away from the order of magnitude (10Hz) of the oscillations produced by the box.

The next step was to solve the problem on the lateral side. Again following suggestions we started to think that a cubic structure could have been more symmetric and could have possibly solved the problem. The following picture is a rendering of the first prototype designed with Autodesk Inventor.

The box can be completely closed, frame and sensor can be attached to the internal planes, which can be slided in and out. On the external part there are some "pins" (4 per face) to fix the elastic cable. They can be eventually removed if the final decision would be to use the box as a protection but to ship it with foam.

Following pictures show the real box, already attached to the frame.

It is easy to see how there is a lot of space between the frame and the box in 1 direction (ca. 15cm), but much less in the other 2 (ca 5cm). This is of course a problem caused by the frame shape which wasn't designed for this structure. For this reason we tested it dropping in the direction of maximum space. Results are reported below (3 drops per height 10,20,30,40,50 cm.).

The good results seem confirmed with low accelerations even at 50 cm. A new frame with larger space in all directions is under consideration.

To-be-done

Shipping tests

References:

ATLAS Letter Of Intent Phase-II Upgrade: https://cds.cern.ch/record/1502664/files/LHCC-I-023.pdf

Frames and Tools: https://twiki.cern.ch/twiki/bin/view/Atlas/Petal130nmHybrids

FrancescoPeri - 2017-02-01