From EUDET-type beam telescopes
Jump to: navigation, search

Telescope infrastructure

The sensor planes are mounted on a Hardware#Telescope frame. The whole frame is placed on a Hardware#Moveable (green) support. Therefore, the frame can be aligned, so that the direction of the test beam is perpendicular to all sensor planes. Using Hardware#PI-stages for moving DUTs, a DUT can be moved accurately and remotely.

Telescope frame

The telescope frame is made of a stable aluminium structure and holds the sensors. Sensors can be adjusted geometrically by changing the distances between two sensor planes or by moving the upstream and/or downstream towers. Rotating the frame by a \(\mu\)m-precise screw plus spring allows to align the telescope properly inside the particle beam. For transportation, the frame is fixed by brakes in order to disable the rotation.

Moveable (green) support

Usually as a user, you don't have to move the support. It is only required to align the telescope, see DESY testbeam web page.

PI-stages for moving DUTs

Between the upstream sensors (plane 0 to 2) and the downstream sensors (plane 3 to 5), a DUT can be placed. By using a stage system, the DUT can be moved.

Components and mounting

\(xy\phi\)-stage system

At the DATURA telescope, e.g., \(xy\phi\)-stages from "Physik Instrumente" (PI) are provided to move the DUT with a mum-precision remotely. The following stages are available:

axis PI name weight range accuracy push/pull force sketch link
\(x\) M-511.DD1 5.0 kg 102 mm 0.1 mum 80 N
\(y\) M-521.DD1 6.2 kg 204 mm 0.1 mum 80 N
\(\phi\) M-060.DG 0.94 kg \(2\pi\) rad 50 murad 6 Nm (torque)
PI angle M-053.10 as possible DUT mounting interface

The PI components are modular, thus, the x- and y-stage can be switched, if a bigger range is needed in x-direction. Both linear stages are connected by this angle: M-592.10.

Using the angle (M-053.10, 0.6 kg) at the vertical axis, users can mount their DUT (please find here a drawing as zipped stp-file, File:Angle bracket.zip, in order to adapt your setup to the angle). In this way the DUT can be moved relatively to the beam remotely.

In order to mount the rotation stage M-060.DG to the angle M-053.10, there is an adapter plate for in between: M-060.HP. (At DESY and in addition, we can provide these PI adapter plates: M-105.AP and M-403.AP2)

Concerning the mechanical mounting, the only constraints are:

  • the weight of the DUT: It is 8 kg in total, thus, minus 0.6 kg if using the angle and minus 0.94 kg if using the rotation stage.
  • the spacing between the upstream and downstream telescope plane is in maximum ~0.5 m

Controlling and operating

The stages are operated with PI components: Each stage is connected via serial SubD 15-pin (1:1) cables to a Controller. Two controller are in use:

Using the PIMikroMove software on a Windows PC (here the hardware#NI crate), you can operate all stages.

Sensors and periphery

Mimosa26 sensor

The sensors of EUDET-type telescopes are six Mimosa26 sensors each in a aluminium housing. A Mimosa26 (Monolithic Active Pixel Sensor (MAPS) with fast binary readout) consists of 576 by 1152 pixels with a pitch of 18.5 \(\mu\)m. This results in an active area of 10.6 X 21.2 mm. The size of the chip is 13,7 mm x 21,5 mm. The design process is from Austria Mikrosysteme AMS-C35B4/OPTO which uses 4 metal- and 2 poly-layers. The thickness of the epitaxial layer differs in the production history between 10, 20 and 15 \(\mu\)m. The sensors are thinned down to a nominal thickness of 50 \(\mu\)m in order to reduce the material budget and therefore a multiple scattering of beam particles. The pixel design is based on self biased diode radtol architecture and each pixel includes the amplification and correlated double sampling.

The data stream of the binary pixel information is zero suppressed on the sensor chip. A sensor is read-out in a continuous rolling shutter mode, columns in parallel and row by row. One readout cycle takes 115.2 \(\mu\)s. Two memories buffer are implemented to perform read and write operations simultaneously.


Each sensor has its individual configuration file (JTAG file). The relevant threshold DAC values are determined by using a characterisation bench developed by IPHC.

Housing and cooling

The sensors are cooled to keep them at a stable temperature, and thus, a stable noise level. This technical note shows the heat up of a Mimosa sensor (File:160812 temperature of Mimosa26.pdf).

Sensor aluminium jigs

For a proper lightshielding, Kapton CB (0,025 mm thick) is used.


The tubing system incorporates 12.5 m long tubes with a 6 mm inner diameter.


A minichiller (by Huber) is used which needs approximately 4 litres of cooling liquid (50 : 50 = tap water : Ethylen-Glycol). The normal operating temperature is 15 $^{\circ}{\rm C}$. Please, do not reduce the temperature to avoid condensation water and corrosion.

Minichiller for cooling sensors.

Condensation can especially be a problem during the summer at CERN. Thus, you might have to set the temperature up to 20 $^{\circ}{\rm C}$ in order to avoid condensation.

At room temperature it takes about 20 minutes until all six sensors are cooled after the cooling system is switched on.

Power supply

The power supply for the Mimosa sensors and all supporting boards is usually an Agilent E3644A. The power supply can be connected to the NI crate via a GBIP cable, so that you can control and monitor the current.

Aux board

The Aux board is attached on top of the sensor housing and handles the communication of signals to and data from the sensor. On top, the three RJ45 cables are attached:

  • JTAG signals (yellow)
  • Clock signals (green)
  • Data stream out (red)

plus one DC power cable.

JTAG and Control boards

JTAG card

JTAG distribution board

Control/clock distribution board

Data reduction board

The data reduction board is an aggregator of the sensors data stream and an adaptor for the NI FPGA.

The round connector on the left of the front panel is an optional grounding connector. It is not used in the standard configuration!


  • a broken AUX board or wrong JTAG files (too low threshold) can cause too high data flow ("Event size meter" of MimosaDAQ > 1000)
  • a wrong (red) data cable cause no header and thus no proper data flow


Each telescope is equipped with four PMTs plus scintillators having a 1 mm x 2 mm area which matches the sensor area as trigger devices. Two crossed trigger devices are placed directly before the first plane 0, and two crossed trigger devices irectly behind the last plane 5.

PMTs are from Hamamatsu, model H5783 or its successor H11901-110.


A scintillator incorporates:

  • 2-3 layers of 3M Vinyl tape (single thickness 0.177mm)
  • 1 layer of 0.033mm thick PET radiant light film
  • 3.0mm thick BC408 (scintillator)
  • 1 layer of 0.033mm thick PET radiant light film
  • 2-3 layers of 3M Vinyl tape (single thickness 0.177mm)

(Some scintillators are wrapped differently, using black shrinkable tubing e.g.).

Trigger Logic Unit


Tlu front.png

The hardware and the functionality of the EUDET TLU is described here: File:EUDET-MEMO-2009-04.pdf. Double triggers are a known problem, during using the autotrigger option. For beam operation, a patch for the discriminator board is availabe, which is integrated in most of the TLUs: File:TLU discboard 180913.pdf.

The operation of the TLU, you find in the EUDAQ user manual.


The File:Aida tlu note.pdf ("miniTLU" and "fullTLU") will be the successor of the current TLU. Basically, it is a hardware upgrade having more options and providing higher trigger rates; find here the online documentation pages for the AIDA mini-TLU.


Telescope control pcs scheme.png

Terminal PC

The PC is located in the hut ("control room") to access to the Hardware#RunControl PC and Hardware#NI crate to operate the telescope hardware and the data taking with EUDAQ. At DATURA and DURANTA, it is connected to the local telescope network and the DESY network for external access.

RunControl PC

If installed, PC for EUDAQ components euRun, DataCollector, LogCollector and OnlineMonitor. At DATURA and DURANTA, it is an 6-core PC with Ubuntu including a 2Tb RAID to store EUDAQ data located in the PC rack in the beam area.

NI crate

NI Max indicates the properly installed NI hardware.

Operating and DAQ PC for Mimosa26 sensor (EUDAQ NIproducer) and PC for TLU PC (EUDAQ TLUproducer). It is located in the PC rack in the beam area.


  • NI PXIe-1082: 8-Slot PXI Express Chassis (780321-01)
  • NI PXIe-8135: embedded controller (former 8133) (782340-04) + 250Gb SSD
  • NI PXIe-7962R NI FlexRIO FPGA Module (Virtex-5 SX50T, 512MB RAM) (781206-01)
  • NI 6585 Digital I/O (200MHz, 32 Channels, LVDS) FlexRIO Adapter Module (781071-01)

OS, drivers, software

  • DESY Windows 7 (32bit)
  • NI drivers
  • Software
    • LV Runtime Environment
    • Visual Studio
    • EUDAQ


Telescope add-ons are hardware components which were tested for a while or are integrated in some of the telescopes. These add-ons are not part of the standard configuration.

FEI4 reference plane

HV (upper PS) and LV (lower PS) power supplies
FEI4 mounted between 4th and 5th plane of the telescope
FEI4 mounted between 4th and 5th plane of the telescope

The FEI4 plane is an additional pixel sensor, which can be used as a trigger, a ROI (Region of Interest) or a time stamping device. The latter will be described in the following, which is available for users coming to DESY and which can be integrated in DATURA or DURANTA.


  • (singel chip) FEI4 to a 3D sensor incl. power cable and CAT-RJ45 to UsbBix, LEMO incl. adaptor for high voltage, see below. Mount the PCB to a Mimosa jig or an extra empty jig. Place it between 4th and 5th plane or behind the last (5th) plane (depending on scattering considerations).
  • USBpix board incl. USB cable, CAT-RJ45 cable to TLU, connections
  • 2 Power supplies
    • Low Voltage: Agilent E3644A (or similar) for analog and digital voltage supply for the chip. Connected to the USBPix. Set to 1.7V (or 2.0V). Pulling 450-600mA in operation.
    • High voltage: Keithley 2410 1100V SourceMeter which provides the muA-resolution. Set up to NEGATIVE 10.0V in steps for scanning (IV curve) or operation.


On the Win 7 Laptop "atlaslap08":

  • c://usbpix_svb_5.3/stcontrol_eudaq.exe (working with EUDAQ version 1.6 and lower)
  • Config file is in c://usbpix_svb_5.3//bin/workingtdac.cfg.root
  • Set local IP adress of EUDAQ RunControl (in the lower right)

Add to EUDAQ config file, if we have these connections:

  • TLU CH0: Telescope
  • TLU CH1: USBPix
DutMask = 3
DUTInput1 = RJ45
TrigRollover = 0 
ReadoutDelay = 10

SkipConfiguration = no
UseSingleBoardConfigs = no
boards = 121
modules[121] = 1
config_file = C:\usbpix_svn_5.3\bin\workingtdac.cfg.root
fpga_firmware = C:\usbpix_svn_5.3\config\usbpixi4.bit
lvl1_delay = 26
tlu_trigger_data_delay = 10

Functionality test

Check IV curve:

  • by increasing high voltage from 0 to -10 V in steps

Iv dark.png

Pixel scan, if needed or as a experience user (according to Andy Blue):

  • Open stcontrol.exe
  • File->open cfg file.
  • Choose /config.3200e10TOT
  • Choose ‘Initialize in dcs objects’….
  • Ignore the warning,. Related to not having Keithley attached by GPIB for HV control
  • Switch on voltages on gui
  • Sanity check: go to pixscan (right panel)
  • Choose digital scan and make a root file to save histograms to
  • Should return a map of 200 hits

Further references