my work in jobs/atelier/guest/sylvia from my PC: U:\atelier (\\PCTS2X2660H01\Ucam\jobs)

Create New Project

  • Layer--> New Layer
  • To modify something the little square at the left of the layer name should me active (red)
  • Aperture-->Create and select the aperture you want (Use the ? taste to know which aperture it is)
  • Edit-->Insert-->Flash, Draw, Square etc...
  • View-->Select the window to select apertures etc.
  • Trasform (on the bar on the right) to move, copy etc.
  • To copy a Layer: select and activate it --> Copy --> Create --> Paste
  • To change the coordinate system: Activate all the layers where you want to chamge them, Transform Objects --> Rescale --> Set Origin and click on thge point you want the new coordinates
  • Cantour: filled objects. Tools-->Netlist--> Build Layer Netlist Selections-->Net Copy Paste on the Cantour Trasform--> Scale 0.5
  • To put something at an exact distance: Select, Tooling --> Distort
  • Create a rectangle: create an aperture and inside the menu of the aperture chose a rectangle and give him the dimensions
  • To know dimensions of a cantour: in the menu of measurements, instead of chose "points" chose "layer"
  • To change a cantour in a rectange: create a rectangel wit the same dimensions, then Tool--> Modify--> replace

Fill pattern with holes or strips:

  • Add the active area that you want to fill (with a rectangular aperture that then you transform in contour or import the dpf file from another project)
  • See its center with the measuring tool, clicking in layer
  • View-->Grid--> put as center the coordinates measured for the center of the active zone, as pitch (for exagonal pattern: x: pitch/2, y:pitch*0.866 (that is the sin 60)
  • Click in the column on the left: Fill Pattern, fill with pad, Odd, click or create an aperture we want to use, then do Apply

Select Window

  • Use also the +, - buttons to add or substract selected objects
  • Toggle: it does the complementar of what is selected

Final Detourage

  • Detourage interal: anticlockwise the versus of the machine, external: clockwise
  • Create an extra layer of kind route
  • Use the tools you can find in
     Tools-->Editing-->Root Manager  
  • To cut lines: Click on the image with crossed lines. Then click on the points where you want to cut your line. Then select the cut line and erase it
  • Click all the icons in Optimize path one after the other in the correct order
  • After the border is defined, check with the tools in "Optimize path" G?, G1G2G3... G? is the starting point of the detourage. It should be only one, otherwise means there are different overlapping lines and they have to be corrected. Check also there is only one versus of the total line
  • You can use the "simulate" tools to simulate the outcome of the final detourage
  • you have to use "Open path" to open the final path in some points in order to keep the board fixed in the machine, and then they will be cut in a second step. Double click (with no snap option) and then select the little line and delete it.
  • To connect open lines: create another aperture with one more decimal (in order to distinguish from the other) and close open paths. Pay attention to the versus of the lines (remember the versis clockwise or anticlockwise depending if external or internal detourage). Use the same versus as the others. Then simulate again the result. (In the final text file, put 111 instead of T7 (1 mm diameter) to identify the little final open lines or click on the aperture and change the name with 111, and put 111 in the Tool Editor window after Drill and Draw that one opens when generate the final file)
  • use the G1,G2,G3... in order to put in order every segment. you should click each segment and put the correct order (with "To line option")
  • After the detourage is defined click the "Save" icon in the Rout editor panel
  • If one wants 90 degrees corners, the machine will round them; in order to do corners, one can put holes, of the same aperture of the detourage, before starting it
  • In the same file there can be put also drill holes (but the aperture has to be put before the detourage). Three reference holes have to be done at the beginning
  • For frames for internal cutting the direction should be clockwise and in order to don't cross the starting end ending lined in the corner, on of the lines should be opened (with aperture 0) and then chained. Like that the machine will start from the center of the line and no problem of crossing
  • Outer frame countercloacksize. Again start from the center and put some attachment point to be cut at the end
  • For lamage (milling) wite in the sm3 file for katia that that aperture is the lamage, with z thickness XXX and put OFF in the routh and path the displacement of the drilling tool (we want to do the lamage exactly were it has ben designed, while all the other lines what we see is the final results, means that the tool has to be moved on the right (automatically done)).

Starting Dimensions

For the press there are standard dimensions. You can choose them from:
  • Panel Plus --> Results --> foil dimensions:
Name Dimensions
P1 223 * 177
P2 254 * 330.2
P3 279.4 * 406.4
P5 406.4 * 330.2
P6 508 * 457.2

Print Films with the photoplotter

  • Output-->SmartPlot
  • Choose layers you want to print --> Add Layers
  • Choose polarity, top-bot on the Mirror field (Top:X, bot: N), resolution (It depends from the withdt of tracks).
  • For copper lines the polarity is Negative (all film black a part from the lines where we want copper)
  • for Pillars: all film black a part pillars
  • For Screen printed resistive layer: Polarity POSITIVE, mirror N (as bottom)
  • merge Layers (Change also the left indicator that merge different layers on the same foil)
  • Check on the bottom of the page that the job has been processed and after a whyle, go on the dark room to take the film. Extract it from the photoplotter machine at the right and put on the development machine on the left (Push a bit when you insert the film on the machine because the first roll doesn't work well). The film will exit on the other room, ready to be used.
  • If some jobs fail may be it's because films inside the photoplotter are ended. Push two times the first botton on the photoplotter in order to open the cover. take 20 films and fill it. close the door. Go in the pc inside the dark room and follow the instructions on the shell. tell him that you have put 20 films.
  • if want to merge small fil with big ones, put priority 1 on the small and leave 5 on the big. If there is place it will assemple all, otherwise will leave out one of the big.
  • to merge high and low resolution need to put in high resolution everything

Output in the LDI computer

  • In the windows were there are all the layers, click layers-->SaveAs-->click on the layer you want to export to the LDI --> Computer Resources --> Select the disk called X: --> create a new folder and save the file there

Output Drill File

  • Set the origin on the bottom left part of the final dimensions of the foil
  • Activate layers where you want to do the output drill files
  • Output --> Drill/Rout --> sm3_drill (metallized holes), sm3_det (not metalic holes) (it depends which machine we want to use, for precision holes better the drill, if not metalized holes change the tool number loking at the table, for example: hole of 0.2: it put T14 because after metallization it will come 0.2. If we don't want metalize, change the number of the tool and put T13 that has as starting diameter 0.2)
  • always delete the name of the file that is in Tools before pusching the Edit button
  • Edit-->File-->Auto-->Flash&Draws and write the nale of the output (control the parameters before saving, width of holes and number of them)
  • Better to optimize the path with Tools--> Tooling-->Smart Drill Optimize so it optimize the path of the machine in order to reduce time
  • If in the tool for the drilling there is T13M31 is because there is a block somewhere. Better to expand the block (third icon in the left), generate the file and then do Undo if we want to keep the block

Output PDF and Gerbers

  • Create two new folters Jobs --> Parameters --> Add to the path name the name of the new folder
  • Output --> CAD --> Gerber274X or PDF

Other UCAM functions

  • Divide grouped Apertures Tools-->Editing-->Expand-->block
  • ready layouts for alignment, multilayers etch: U:/cnf/setup/panel plus. If one clicks on Panel Plus can chose the layout it prefers and dimensions for the press (better defining before which layer is inner, which outer so it can better design aligment, glue lines etch..) it put also date and name of the job *How to do a strange surface full: design the surface with lines all well connected and then change the aperture in CNT
  • How to modify a CNT: edit the CNT and say: circle. Modify it and then come back as CNT
  • In the left side: first is the contourize function
  • A block can be modified with "Edit" without breaking it, then in the main menu: Quit and save it

Insert cern logo

  • Go in the layer panel --> add DPF --> job/doc/logo

Open gerber file

  • save file in a folder
  • in UCAM: Job/SmartStart
  • click the file and do Analyze and check is a gerber file
  • Do AddtoJob and it create a layer

Débit matière

  • Ask the permission to enter in the room to Benilde or Julien
  • Compile the paper with written how much material has been taken


Name Thype Use Width
Polyimide (kapton)
Pyralux Dupont LF adesive Cu-Kapton-Cu press LF100 25um of glue, LF121 25u glue - 2*25u ka 25u glue
FR Pyralux Cu-Kapton-Cu flexible circuits  
Sheldahl Cu-Ka-Cu more precision, used for GEMS  
Pyralux LF100 25um of glue    
Pyralux LF700 13um of glue    
Pyralux LF121 25u glue - 2*25u ka 25u glue
Prepreg epoxy foils with glue   75um, 125um...
DUPONT PC 1025 (1020) (1010) coverlay photoimageable MM pillars 64 um (51um) (25 um)

  • Material: combination of acrylic, urethane, imide-based material
  • Dielectric strenght: 137 kV/mm

Glue krempel more strength that LF (epoxy acrylic) but LF tested at cryogenic temperatures whyle krempel not yet tested

  • Cu 18/5 means 18 um of protection and 5 um of copper

Pumice (Ponce)

  • neutre powder has sand smaller that acid one. Acid used to clean the copper if it's oxidated
  • to activate the pump:
  • put some more powder in the bath
  • opoen the pressure of the pomp @ 3 bar
  • put the yellow tube inside the bath and fix it to the box
  • mix the bath with the high pressure water that is exiting
  • use it
  • when finish clean the pump putting water in the other tank and put the input yello tube inside, leave running for some minutes


  • To be used before laminate or before the press, to clean the circuit. (Only rigid circuits)
  • It's possible to change the pression of the brushes.
  • If the circuit is little (10 cm) put on diagonal otherwise it will be blocked in the middle of the machine


Riston Liquid

used for circuits with metallic holes, in order to protect them during the etching phase. Put on the oven at 80C for 5 min to dry, then laminate.


resistive type: fx930
  • 17 um Copper --> 1 Layer
  • 35 um --> 2 layers
  • when you switch on, wait until the temperature is >110C. Close the two rolls, start the "Convoieur".
  • Clean the circuit before inserting it
  • If it's not a flexible circuit, insert it and then put pression
  • Make sure at the exit the photoresist is very well stretched.

For small size lines use thinner photoresist (15 or 20 um) with lower exposition time (<50) and higher development speed (~80)

UV exposure (Insolation)

LDI ( Direct Laser Imaging)

  • pw: dp100
  • upload the program from the pc, look for dimensions etc. (it takes ~ 2 min to trasfer informations to the machine). Name of the job, dimensions, resist type (fx930), Panel type (IL: inner layer if multilayers, OL: outer, if only one layer or if the outer layers), width of the board (if OL the total width, if IL it depends). Polarity is the contrary of the film. If lines are diseigned as red in UCAM, leave the polarity positive, it will design directly your lines in the photoresist.
  • delete processed layer files --> select top, bot --> APPLY (If you click ok then the window will disappear and for the next file you have to reopen it from the desktop-->IGU)
  • If there are not holes in your circuit you can disable the record function both from the starting pc or directly from the LDI
  • From the screen on the machine click on the botton with two arrows on the side in order to wake up the machine. It takes some minutes saying "waking up" Then wen you finish, click again in order to go in standby. Sometime it takes so long that we have to restart the machine. Shut down... or if it doesn/t work, with the switch in the front panel of the machine..
  • From the screen of the machine: take layers top, bot and move to the grey windows on the top. Press Start
  • If two layers of photoresist I should change the exposure time from 60 to 90 or 120, 200.. More is the exposition time, more resistent is the photoresist during the etching process and it doesn't risk to broke.
  • Positioning circuits: TOP: three holes on the top right, BOTTOM: top left (!!! pay attention the board doesn't take UV light after the application of photosensitive material)
  • Upload top or bottom files
  • Press the vacuum and clean
  • press both buttons on the side to start
  • then do the other layer
  • le mire sono a 2.54 mm dai bordi, se non son standard, posiziona il pezzo in modo che sia a quella distanza, se no non le trova

Manual UV machine

The lamp is stronger than LDI. For high irradiation is better. Used also for "portefeuille" techinque for alignment by eye.
  • portefeuille: cut the films bigger thank the circuits. The bottom one bigger than the top. Make many holes on two sides only on the top one and put tape. Align by eye.
  • Clean with alcool at the first use in the morning, the other times clean well both the plane and the film and the circuit.
  • if you want to change the exposition time wrt to the one already set, do: imput --> code --> number you want to put (the code is written on the machine)

Developpouse (carbonate de sodium, 35 ^)

Take out the photoresist where it has not been exposed to UV light.
  • Take out the trasparent film of the photoresist before inserting the circuit in it
  • SPEEDS: 1 layer of photoresist: ~ 58, two layers: v=70*2times or v=35 (speed used for flexible of Pb_MM, one layer: 35)

Etching (Gravure)

Perclorure de fer, it etched the copper that is not protected by the polimerized photoresist
  • Automatic machine for low res circuits. Results are visible only at the end. Manual machine for high precision (GEMS). Make some seconds, wash, look with microscope, and again untill it's perfectly etched.
  • Before etching make sure (with the microscope) all the lined are covered with photoresist. If a line is interroupted, correct with the black pen, If two lines are touching, cut the connection with the cutter. If there are metallic holes, make sure the photoresist is not destroied.
  • If you have to put more times in the machine, change the side and the orientation
  • Control with the microscope the lines are well marked and no copper is visible any more on the side
  • If there are some pieces of copper that don't go away is because the development has not been well done. Strip, reput the photoresist, UV exposure, development and etching again.
  • If the liquid has been recently changed, check the temperature of the machine. It should be ~35-40C. If it's too cold, wait if you have high precision circuits. If it's too high increase the speed


For delicate circuits alcool baths, otherwise the machine
  • take out the pen with alcool before putting in the machine
  • speed of machine: 1 layer photoresist--> ~ (Temperature - 10), 2 layers: v=T/2
  • After stripping if the circuit is finished, put 1h on the oven to make evaporate vapours. No oven for multilayer circuits.
  • Alcool bath in the gravure room: open the three rubinetts and swith on the red switch to start pumps


If putting many layers of prepreg, put all of them on the same way (the curvature from the roll on the same directions, otherwise there will be distortions. Solder the prepreg on the 4 corners before putting on the press.
  • Switch on the press pulling on the red button in front and turning the red switch on the side (take out also the cable tie on the manometer, used to work with open doors)
  • Put boards inside in the middle with tapis blanche on the top (!!! remember to put spacers on the planes you are not using with your circuits otherwise it will break!)
  • to go to the primary interface push many times F8
  • F5 --> chose the Pxx format of your board --> enter (It sets already the pressure for that kind of dimensions)
  • ( for prepreg 100 micron and copper on the top, SSAL50: sans halogen, 50 minutes of flat top)
  • F3: load files --> yes
  • F8 --> F4 to check the temperature and pressure graphic. Chose Temp&pressure --> Enter
  • F8
  • close the doors, turn the vacuum turning the key with the symbol with a big arrow going on the left
  • Turn the authomatic vacuum: key with symbol of 4 arrows arranged in a square
  • F6 --> reset alarms
  • F2 --> reset warnings. Password: cern --> Enter
  • Start press: push the green button with 4 arrows arranged in a square and a vertical line in the center
  • F8 --> F4 --> chose the graphic "Theoric temperature & pression all superposed" and it starts to display the theorethical curve and measured points, one per minute
  • When it's finished there will be an alarm
  • Turn back the two keys for the vacuum
  • Open
  • Take gloves and remove your circuits


  • switch on the laminator (no heating)
  • take isotac from the fridge
  • cut it (stiky part on the top)
  • wash and ponce the surface to laminate
  • clean well with water *deposito on top the isotac and with the antistatic roll take out every bubble with water running below
  • go in the laminator, excess of water should exit



Optical Test


Electric Test (UCAM)

  • Create a Test folder
  • copy inside all the project and erase everithing is not going to be tested (for each layer leave holes). (Use the Toggle botton if you have to erase a lot of objects)
  • Create an EXTRA layer, called OUTLINE of the type Outline
  • create a circular aperture of 0.1 and do a rectangular box around the interesting zone for the electric test
  • Duplicate this window on every circuits in the project. Use the button with the 4 windows --> Edit-->Select Chain
  • Tool --> Electric Test --> Loc8 (name of the electric test machine)
  • Net List --> build
  • Test points (use mask)
  • outut MNF2 (all) It will create test points
  • Check if test points it has created coincide with mechanical holes it has to check

Drill Machine (perceuse)

IT use also a foils of cartonpaper and one of aluminum to drill

Manual Drill machine

  • Precision in the alignment with the camera and the viewfinder: +- 50 um
  • Open the machine with the black button at the bottom right
  • Control the dimension of the tool (or change it in case)
  • Close the machine
  • Put a piece of white cardboard plus the trasparent plastic and control they can fit (otherwise unscrew and change dimensions)
  • Try to drill (you can arrange the two speed of rotation and traslation). (Kapton foils have to be drilled with high speed)
  • Start drilling with your feet
  • Switch on the lamp of the camera
  • look in the scrren and align your board
  • Put the plastic on the top and fix it
  • Drill
  • For the next hole move the white paper and the plastic where there are not other holes
  • Check with the microscope the firt hole you did (alignemnt ans boards)

Manual drill machine to drill metal

  • adjust the speed in order to be very low (for metal). One has to go in the machine located chez Marina, but with our tools. Use the liquid (take from Xavier) in order to take out heat while drilling
  • Do the hole and then change the tool, put the conic one in order to round the edges

Flexible Circuits

Difference on production steps with respect to the rigid ones:
  • débit matière: cut by hand, don't use the machine
  • No bresseuse, use the manual powder
  • Laminatours without pression
  • Developpouse, Etching as the others but putting the rigid part attached with the tape
  • Stripping with alcool


  • Clean only the interested part with acid powder and brush until the copper is pink (or neutre powder if it's not copper)
  • 1 min in cloridric acid + 1 min in clorure d'ammonium (catalizator) (without rinsing between the two)
  • rinse well with water
  • 15 min in Ni @ 90 degrees (after 2-3 min look at the color, it should be uniform grey otherwise restart cleaning with the brush) Ni deposit ~ 5 um/15 min
  • Rinse very well (Ni with chanure is a poison)
  • 10 min Au @ 90 + cianure
  • Rince
  • look at the microscope
  • For deposition of Nickel on the Silver: look with the scaplel untill it's not possible to scratch. Nickel is enforcing the Silver. Before the Nichel dry. To start put catalizator many times avery 2-3 minutes to be sure it's uniform, but see by eye at the beginning (CHECK EVRY SOME MINUTE!!!)
  • After 5 min it should be uniform on the surface in order to check


20 s in bagno pilitore di rame, pulire con H2O, 30 s nell'acido XXX (vasca in entrata) pulire bene dall'alto se no cola e fa un segno sul circuito


Standard (Julien) MM design in UCAM:

Resistive strips

Used paste (carbon mixture): 100 kOhm/sq for 1D readout, 1MOhm/sq 2D readout. Values written on the ID number of the paste (116 --> 1*10^6 Ohm, 12116 used paste for 1Mohm)

Standard way (photolitography)

Do a circuit on top of resistive lines with the complementar of resistive strips, done with copper, then fill with resistive paste the holes and etch the copper. Advantage: more precise than screen printing. edges very well defined by copper lines.
  • For standard 10*10 cm^2 micromegas, now the readout PCB with already the complementary copper lines it's produced by ELTOS. 4 layers: bottom: all copper a part from holes for connectors, XY readout and top layer with copper were not resistive strips
  • wash the boards in metalization room (first bath + rinse in the second one and with high pressure water). Dry with compressed air and in the clean john's oven. During the washing and drying processing, leave the board with the interconnection between strips on the bottom so dust can be better washed away and it's not so critical bacause potential is lower on that side.
  • Leave degasee the board at 120 degrees for at least 1h
  • Enter in the clean room to put resistive paste:
    • fix on the table max 4 chambers, one behind the other with paper tape (again the interconnection on the bottom side: were you are)
    • kapton tape, width 40 um on the border, parallel to the strips (pay attention to put well the tape, without bubbles or dust because it keeps the width of the paste)
    • put kapton tape all around just to avoid the paste spreading on the board
    • clean well the spatula and mix well the resistive paste
    • lay the paste with the spatula in a way to not scratch the board and not put too much (the tape on the border will help)
    • clean the boarders from the resistive paste
  • Leave the paste to dry on air for 1 day
  • look that there are not holes or bubbles on the paste in the active zone. You should see an homogeneus layer and not see the individual lines (they will be seen after the cooking)
  • Cook 1h at 120C (after slow T increasing). The Copper on the side should be opaque while the R paste should shine. Now you should see the pattern of strips
  • Use the poncouse machine to take away most of the R paste (before it will become too strong, after the final cooking) over the copper, up to almost reaching the copper strips. (The copper strips now should shine, while the R paste will be opaque). Now the active zone should have an almost homogeneous layer, without seeing the pistes.
  • Final cook at 180C
  • final poncouse to take away the last thin layer of R over the copper strips
  • laminate, put the mask with next pattern, and UV lights
  • Etch the copper

  • take out the resistive paste from the fridge the night before
  • At the end clean with acetone and chiffon
  • put at 120 deg in the oven of the clean room
  • then cook in th little oven near the ponceuse

Screen printing (Serigraphie)

Print directly on the insulator layer on top of readout strips the pattern with resistive strips. Advantage: much faster than standard technique, just one step; good for big productions and industrialization.
  • Print the film with positive polarity and no mirror (if you want 0.3 mm lines, print the film with 0.2 width because they will spread a bit with the printing)
  • Photoresist on the mesh put by hand on the clean room with a photoresistive glue
  • Dry on the oven (not too high T, max 50C because the mesh is stainless steel and the frame Al, they have different heat expansion coefficients)
  • Exposure with the UV machine in the clean room with the film on the top
  • development with just water
  • Dry again on the oven at 50C
  • Antonio does the printing
  • After checking with the microscope there are no cut lines and the alignement is ok, dry on the ovens of the clean room at 100C for 30min, then put on other ovens at 180C for ~2 h (increasing the temperature 20 C by 20 C every 10 min)

Making Pillars

  • carefull wash the substrate with the neutre powder
  • Laminate the coverlay only on one side (for the ATLAS big chamber they laminated both sides to avoid distorsions). Pay attention to take away the transparent plastic films while laminating
  • Between laminating different layers of coverlay take out the plastic film (it's easy if you put on the fridge for some minutes).
  • When doing many layers of coverlay, to avoid waves: after taking out from the fridge, take out the plastic film and live in open air to warm a bit (pay attention to dust impinging on coverlay!). Cut some parts of coverlay on the borders were the trapped air can exit the next step.
  • Laminate the mesh (after cleaning it) with the coverlay for bulk micromegas
    • To clean the mesh: cleaning in a bath in the metalization room, then wash it with compressed water. Dry on jonatan's clean oven. Do not dry the mesh with compressed air because it's not clean enough
  • Then irradiate with UV and the film mask (UV @ 4000)
  • develop until the coverlay is gone on the active area. Don't develop too much otherwise pillars will go away, and not too less otherwise the coverlay do as insulator in the active area
    • First step of development with the coverlay UP otherwise it will stick on the bottom rolls of the machine. Wash it whyle inserting in the machine
    • other steps DOWN (the liquid is better because it doesn't create the "swimming pool effect" on the active area. It just hoes directly and then it drops by himself
    • Rotate the board every step on the developing machine
    • Stop when the green layer is gone (check with the microscope)
    • MM done with Julien, steps on the deveppouse: spped of the machine: 20UP, 20Dw, 20DW, 20DW, 30Dw, 40DW ..
  • Control with the microscope
  • Cook gradually increasing the temperature otherwise will be bubbles on the coverlay

Image in the coverlay

  • 1 layer of coverlay, if not fine pistes, use 360 of irradiation, while 120 for piste fines
  • Acetone to take out coverlay
  • development starting at 25
  • If little pistes to do in the coverlay, UV, take out protection, oven 10 min 2 120 deg to polimerize also the bottom of the coverlay who saw few ligh, development as usual
  • Holes of 100 um in the coverlay used for spiral detector: 50 um coverlay, irradiation 120 (before leav the vacuum for 10 min otherwise light pass under the film and conic holes), take out plastique, oven 10-15 min @ 80-90 degrees, development @ 40 and maximun leaving the surface on the bottom of the machine (rinse with water before) and before pass a green board in order to be sure there will not be pieces of coverlay on the rolls of the machine that will attach to the circuit
  • Cooking 90 deg 30 min, 100 30min, 120 30 min, 150 1h, 170 1h
  • Holes of 100 um in the coverlay used for spiral detector: 25 um coverlay: tried insolations of 60, 80, 120 --> 120 the best but still not ok

Electrical connections

Silver paste (look in the fridge if there is already something done). Put with the "scalpel" and then dry it in the oven 30 minutes @ 130C . Check at the end if it is dried enough.

Soldering connectors

Clean the surface (usally one solders on the Ni/Au, not on Cu because it oxides) put a bit of tin (etain in fracais) on the two extremities of the connector (PCB side) in order to attach the connector (cut away the little tips of the connector if there are not holes in the circuit). clean with the pumice stone, put the flux mixed with alcool with the paint brush (it reduces the surface tension of Cu). Solder one by one. Se viene troppo spesso e connetto due piste, per levarlo usa calza rame con ferro, assorbe lo stagno. At the end clean with alcool and painting brush

Etch only a piece of circuit and protect the other Cu parts

put a piece of film were you want to protect, place the kapton tape all around and put the green paste XXX etc after etching use the solvent XXX (use the mask to breath), then acetone and alcool

Test and clean a micromegas

  • tester @ 500 V between mesh ad pillars, should have R>10G and I<10 nA, if not one shuld look at the mocroscope if there is something below the mesh and clean with high pressure water (once we saw something and Rui did a little cut in the corner of the mesh in order to make the dust exit, then with high pressure water very close to the mesh (10 cm), and we dropped it also in lessive and Saude caustique (basic) (some seconds) + the acide at the end of permanganate for some 10s, then high pressure water again
  • invert polarity with tester, should behave the same way the velocity of going down with current, if not it's dust who make behaviour asymmetric
  • 700 V direclty, the if ok, 800 and 850V (900) the current strongly depends from the room humidity
  • If problem that in the middle no source and in the borders yes: may be the mesh is not well attached to the pillars, sligltly bend the micromegas and see if all the pillars follow the mesh *

Test and clean a gem

  • short --> permonganate, finisheur, eau demineralize
  • sparks random, leakage current --> lessive, demineralized H20 flat, oven 80
  • big triple GEM stack: H20 demineralized everywhere, clean ultrasonic bath with demineralized water. 20 min, move everytime, don't leave in the same position otherwise demlamination of Cu, after 10 min clean again water with new one on top. After 20 min oven 80 one hour. Clean also the other components. When putting in the ulstrasonic bath go very slowly inside, pay attention air bubble don't stick in the holes

Etching polyammide (kapton) for GEMs

ethylene diamine. Wear the integral mask with ir taken from the back so te filters will not see directly vapors

Acide Sulfurique (H2SO4)

  • Used to clean from the krempel glue for example.
  • !!!! Remember to wear big gloves, mask for drops and protection for vests
  • In the "image trasfer room" take the bottle with writtem "new" on it. Put in a little vessel. Put water in another vessel. Take two new paint brushes, one used for each vessel.
  • Impinge the circuit in the acid and work with the paint brush some seconds. Then impinge immediatly in the water vessel, also cleaning with the painting brush. After the use rinse the painting brushes and trow away

Potassium permanganate (desmearing)

to clean detectors with big holes. lessive, red solution (base), green acid

Silver electrode deposition

  • 1 component silver paste: 1906
  • To fill a surface with little holes: take the green couciu` "racle" (before protect the borders of you surface with kapton tape)
  • For big and regular surfaces we used the screen printing because was difficult to lay the paste very uniformly by hand, then when the surface is flat (after cooking and scratching with snd paper) I filled the little holes with the coucuiu and re cooked again (don't clean the paste on the coverlay image with acetone, even if already cooked the coverlay have been displaced!)

Big Circuits Machines

  • Exposure: 1.4 m * 2.2. m
  • Laminator: 1.2 m width
  • Oven: 2.4 m * 1.4 m
  • Etching: 0.6m width
  • electrochemical etching (GEM) 2 m * 0.6 m

Tutto quello che avreste voltuto sapere e non avreste mai osato chiedere (bè.. io l'ho fatto) sui MPGD

  • Perchè mettere poliuretano in ThickGEM? è messo con uno spray, penetra nei fori fatti con la drill machine in cui ci sono ancora residui di fibra di vetro e fa la superficie piu' smooth. IN piu' è idrofobico quindi rende il materiale pu' isolante (nell'epoxy con piccole percentuali di H2O diventa conduttore e gli ioni si spostano, diminuendo il gap tra gli elettrodi--> scarica)
  • Com'è fatta la mesh di stainless steel delle micromegas? come i tessuti, coi fili intrecciati. Usata nelle industrie come filtro grossolano (resistente e si pulisce bene)
  • Perchè tutti usano GEM e solo ATLAS MM? ora che c'è la protezione resistiva, MM sono comparabili. Prima eran meglio GEM e ora la gente ha ancora l'idea precedente. Joerg e Rui han studiat per un paio d'anni la protezione da mettere per le spark. Le camere Rxxx solo dall'ottava in poi si son visti i diversi risultati. Prove senza simulazioni. Provavano a fare spark e vedere per quanto tempo detector reggeva. Problema calore dissipato., Con strato resisitivo si dissipa meglio. Se si lascia piu' tempo tra una spark e l'altra il materiale ha piu' tempo per recovering ma tempo morto per detector.
  • MSGC studiate per 8 anni per CMS e poi non han funzionato per via delle spark che rovinavano il circuito. problema che strip troppo piccole, effetto Joule fondeva l'anodo. Non abbastanza materiale per dissipare. GEM inventate proprio come preamplificatore per MSGC. Ormai troppo tardi per CMS.
  • In compass GEM funzionano ancora, le MM già cambiate tre volte. non resistive e si rovinano con sparks.
  • a saclay le prima MM le facevano senza pillar e mettendo fili da pesca come distanziatori. Al cern han iniziato a mettere i pillar, prima sulla mesh, poi meglio su pcb e fare bulk
  • MM di T2K fatte al cern, bulk non resistive. Ok perchè bassissimi rate. prime MM fatte abbastanza grandi
  • MM è il primo detector a gas che è totalmente industrializzabile e facile da fare una volta che l'inidustria ha capito come. Si fa come coi silici per l'industrializzazione
  • Perche' RETGEM (resistive thick GEM) non son mai usate? perche` non funzionan bene, resistenza non abbastanza alta da stoppare spark e materiale in mezzo (epoxy) si danneggia comunque
  • Com'e` fatto kapton resistivo? e` gia` comprato cosi` con del deposito di carbonioi in mezzo
  • Sealed detectors: per usare le standard gem bisogna prima fare degasare il kapton mettendo ad alta temperatura in una struttura totalmente metallica e senza colla (se no degasa lei), se si riesce a fare questo poi si riempie del gas che serve e si sigilla, dovrebbe funzionare. Nessuno usa perche` vantaggi gas detector e` possibilita` di fare grandi detector, per piccoli e` ancora piu` conveniente il silicio. Glass gem non degasano ma costan di piu` del silicio. Buchi nel glass lo fanno i giapponesi con glass photoimageable, lo ha solo una compagnia in giappone e non puo` venderlo in europa e quindi non lo possiam provare

Good to Know

  • 1mils = 1/000 inch = 25.4 um



Cleaning Micromegas with leakage current

  • test if there is leakage current @ 500V with tester, then if it's ok with power supply @ 950V should have <10 na
  • If current, open from the bottom of the screwed frame, take out the drift, look at the microscope if the mesh is ok or if in the corners there is some bad stuff
  • Clean with high pressure water and dry, retest
  • When some stuff is under the mesh, Rui cuts a little hole in the corner and then cleans again a lot in order to make it exit
  • Whyle testing, invert the potential, if the behaviour is not the same means there is some dust that makes things asymmetric

--++ Cleaning GEM

  • if there is a short: lessive + permonganate + rinsing with demineralized water (horizontal flow, in galvano)
  • no short, no permonganate
  • if there is frame no karshare


in the cern store: two kinds of standard GEM: 10*10 and 5*5. the F at the end of the code means "frame" Standard GEM UCAM file location:
  • 5*5 cm active area: rui/g50%2A50 whyle the frame is didier/art/cadre74_54.sm3
  • film: for double mask print two films in two separated foils otherwise not enough space for the portafoille. Resolution of film: not 5080 as usual but 10160. Films should be printed in the same time otherwise there a re fluctuations during printing
  • Diameter holes in the film of 55 micron in order to have 70 after the etching
  • cut top film at 2.5 mm from the foil and the bottom at 3 cm, every 3 cm in the top do holes and put pieces of tape (well cut)
  • debit matiere: Sheldal 5-50-5 in clena room (use the tapis blanche below and gloves to cut it)
  • photoresist: fx920 (20 um) in clean room
  • alignement: fix the bottom film at the UV machine, align and then fix the top with blue tape. Verify the alignement after the vacuum
  • Isolation @ 35 , vacuum 5 mn before
  • development @ 90 (put in a bar before entering in the machine)
  • rinse with water and the demineralised water
  • leave in air untill it's dry
  • check at the microscope the diameter of holes (need to be calibrated with the pitch of 140 the measurement with the aquisition system)
  • put kapton tape all around the foil over an epoxy board
  • etching: rinse very well before, then 10s one side and 10s the other side. Rinse well
  • strip in alchool and then rinse. Leave dry in air
  • Check the diameter of hole, repeat for the other side of the board
  • Permanganate to take out the crominium: put in a rack, then 30 s lessive, 30s permanaganate 30 s gree bath
  • Jose checks the four corners of the GEM after every etching step, and measure with the microscope. They are ok if the difference in diameter is maximum 5 micron, usually is 1-2 micron
--+ Epoxy gas box
  • after it has been drilled one has to manually drill the two gas connector holes with a 2.7 tool (but longer than the ones used by katia)

Etch copper around the GEM (Little GEM)

  • Put the GEm in the rack and put in the lessive (30s) + H2O + alchool
  • Insolation (35), development max speed, etching ~ 10s
  • After stripping the kapton has a silver color. It's the crome, one has to get rid of it with some seconds in the permanganate + rinse
  • Put GEM foil in a "cadre" with tape, lessive 30s, high pressure water, permanganate + green bain on its right, water, acid cromique 30s, water, high pressure water, oven 1h @ 80 degrees or 30 min @ 120

Thick GEM

  • Trieste method: 2 hours of hand cleaning with pumice it clean only the copper not epoxy but works, (more than 1 day in the oven at 170 to dry the epoxy)

Sand Blast machine (sableuse)

  • switch on the big vacuum cleaner, open the pressure, chech the blue tank is full of sand

Soldering connectors

  • For difficult ones better to ask to Nicole Wauquier or to EXCOFFIER, Stephane (1-015). Protect with kapton tape the lines around the soldered part. Put the no flow, solder, then clean from it otherwise it oxides

Cooking temperatures:

  • THGEM: there is epoxy that degases. It melts at T>200, normaly cooked at 130 deg
  • GEM: kapton can go more than 300 deg. Put in the LHC beampipe after a baking at 300deg for degasing. cook at 300deg for sealed detectors but without oxigen otherwise copper will oxidize.
  • honeycomb only at 80 deg because has been cold glued
  • resistive paste ~ 1h at 170 slow increase
  • silver glue 120


  • Can be removed with acetone before cooking
  • Put after having tested the THGEm that doesn't spark <1500V
  • Put one layer with the spray, check by eye don't close the holes, left open air a while, check again not closed holes (while drying it becomes thinner) and the oven to dry
  • Remove from the copper with sandpaper 1000


  • Macchina per fare buchi (manuale dall'ato) se i buchi son grandi sul Cu vengono sbavature. per evitarle prendere uno straccetto (2 cm2) e bagnarlo, metterlo sopra al posto da bucare. entra nel buco e non sbava
  • sadare connettori HV: mettere stagno sia sui capi filo che HV, scaldare HV e appoggiare filo vicino al buco --> lo ciuccia dentro appena e` caldo
Edit | Attach | Watch | Print version | History: r72 < r71 < r70 < r69 < r68 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r72 - 2015-06-09 - SilviaFranchino
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright &© 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback