FRS Current Grids

Location at FRS
Construction of CG
Voltage settings
Pulser test
Gas supply

This page was last modified by Helmut Weick, 9th Nov 2022, contact h.weick(at),   Imprint (Impressum), Privacy Policy (Datenschutzerklärung)

Location at the FRS

The CGs are used in the FRS to measure x and often also y position in the focal planes. There are:

focal plane simple name nomenclature type
S0 CG01, CG02 TS1DG5, TS2DG2 small +/-38mm, X+Y
S1 CG11 TS3D2H large +/-92mm, X
S2 CG21 TS3DG7 small +/-38mm, X+Y
S3 CG31 TS4DG3H large +/-92mm, X
S5 CG51 TS5DG1 small +/-38mm, X+Y
S6 CG61, CG62 TE5DGCG, TE5DGDG small +/-38mm, X+Y
S7 CG71 TS7DG1 dismounted
S8 CG81,CG82 TH4DG4G, TH4DG5G small, other type

Picture of vacuum chamber at S5 with CG51 mounted.

Construction of CGs

The CGs collect secondary electrons produced in a window or foil just in front of the detector wires or in a gas volume surrounding the wires.
Some CGs have wires for X and Y direction others only for X. Wire spacing is 1 mm.

Photograph of CG61:

Left: sketch of type small size CG, right: type large size CG: (figures from [2])

Table with wire layer and wire thicknesses.

Voltage settings

As only low energy electrons (max. few eV) have to be collected a bias of about -10V on the cathode is enough to collect all.
To be safe also for for fast extraction with space charge typically -90V are applied. Set values from CAEN HV crate.

Some CGs were tested up to -500V. However, this brings no gain in signal height, but is dangerous (sparking).
As used in present the detectors are not foressen for amplification even with gas filling. The gas only increases the volume from where secondary electrons can be emitted. Otherwise only a very thin layer of a solid would contribute.
The original design was for real gas amplification and up to -400V on the cathode plane. But then also gas at much lower pressure must be used.


The readout is done via single amplifiers per wire. These amplifiers are then multiplexed an read out in sequence by the controller. The amplifiers are inside the tunnel (at S2) and the 2 controllers are located in the electronics room of FRS Messhütte.
Electronics scheme from [2]:

Left amplifiers (new type not used for FRS), right old type at S2:

Multiplexer at S2:

The wider (only x) detectors are read in 2 groups of 48 channels, the narrow (x+y) detectors use only 47 channels and are read on outside first then the inside channels (due to symmetry with one wire exactly in the center).

The LSA software to display the same result is called "Profile Grids".
The ON button means readout on, voltage is set on a separate HV crate. As the readout is limited to only one detector per cycle it can be helpful to switch off readout when displaying many detectors at once.
Detectors can also be moved from within the app even for step motor drives. But for step motors better check that the drive really ends up exactly at the wanted position in DeviceControl.
TEST starts the same test pulser pattern from remote. The same test pattern then looks like above.
You can chose different integration times (0.5ms or 5ms). The gain can be adjusted automatically or in steps by hand.
Watch that the time of "Last Measurement" updates and data are really new.
Example of peaks with beam on GTS1DG5 (=CG01) and TS2DGS (=CG02). The readout was stopped (yellow =off) to freeze the picture.

To save readout channels for |x|>10mm always 2 wires are combined to one channel, only for |x|<10mm the full resolution of dx = 1 mm is used. However, the old front-end computer already sorts them wrongly in the transition region and creates artefact peaks around x=10mm. to correct this bug there is a software option in the [Menu] button.
The menue also includes an option to export data to an ascii list and a screenshot, which are copied to clipboard (

Pulser Test

A full test without beam is not possible because detectors are inside the vacuum chamber. But the amplifiers and readout can be tested. For this a pulser is integrated into the two CG controllers located in the electronics room of FRS Messhütte.

Test procedure:
1.) Switch to manual
2a.) Set range (Bereich) to "nA/V" "x100", pulse length switch "=5ms" and "TEST", or
2b.) set range (Bereich) to "nA/V","x10", pulse length switch =0.5ms and "TEST".
3.) Push many times to select detector. A list is glued to the crate.

Connect oscilloscope to BNC connectors for X or Y direction. The sequential readout should produce a regular pattern in time. Missing peaks indicate a bad amplifier channel. Example: Picture on oscilloscope, 48 channels with different resistors for alternating height. Top x direction, below y direction, one bad channel (#16) in x direction.

Results of a pulser test for all CGs done in Dec 2019 (pdf-file).

Testing can also be done from within the ProfileGrid app, but this requires a running pattern.
Just select Test, the test pattern in "Profile Grids" then looks like this (the channels on the side are divided by two).

Gas supply

The detectors operate with 90% Ar + 10% CO2 gas at 1 bar. The gas flow can be controlled independently for the focal planes and is set to roughly 5 l/hour. Before the detectors are filled homogeneously gas flow has to be provided for some hours. The GSI detector laboratory (Dr. C.J. Schmidt, Mr. Michael Träger, Tel. 2639) takes care of the gas supply unit and has to be informed in advance. The gas flow is controlled, mixed and monitored from the cupboard in front of the cave entrance, see picture below.

However, as the detectors are usually not used in a gas amplification mode they can also run with no gas filling but just rough vacuum. For fast extracton this even is better as it avoids background and widened peaks. Instead of gas supply in this case a roughing pump can be connected via a hose.



[1]  R. Anne et al., "Development of Beam-Profile Monitors with Gas Amplification and Current Readout for the SIS Projectile-Fragment Separator ", GSI Scientific Report (1990) p.257.
[2]  M. Weber, E. Roeckl, K. Rykaszewski, I. Schall, "Beam-Profile Monitors with Gas Amplification and Current Readout for the Projectile-Fragment Separator", GSI internal report (1990).