SlsDetector camera



The PSI/SLS Detector Group has developed a family of X-ray detectors: Mythen, Pilatus, Gotthard, Eiger, Moench, Jungrau, among others. Most of them are controlled through Ethernet interfaces, with optional dedicated data link(s). A common protocol has been developed to control these detectors, based on the slsDetector class. A separate software entity receives and dispatch the data: slsReceiver. The SlsDetector LIMA plugin instantiates the necessary software objects to perform data aquisitions with the detectors supported by the slsDetectorsPackage.

The current implementation only works with the PSI/Eiger detectors.


The slsDetectorPackage-v2.3.x is needed by the SlsDetector LIMA plugin. As explained in Installation of Eiger computer at ESRF, the slsDetectorPackage is included as a submodule in the SlsDetector camera plugin. It will be automatically compiled and installed during the LIMA build procedure.

In addition to that, a configuration file, containing the commands necessary to initialise both the slsDetector” and *slsReceiver instances, is required.

The library protocol uses Unix System-V IPC shared memory blocks to exchange information between processes. The segments, referred to by keys matching hex 000016xx, must be owned by the user running the plugin, if it is not root. The following command, which removes the existing segments, must be run by the segments’ owner (or root) so they can be deleted/created by another user:

ipcs -m | \
  grep -E '^0x000016[0-9a-z]{2}' | \
  awk '{print $2}' | while read m; do \
    ipcrm -m $m; \

High-performance Acquisitions

High-performance acquisitions require a specific backend computer setup. Please refer to the Installation of Eiger computer at ESRF.

Installation & Module configuration

As a reference, see:

Initialisation and Capabilities

In order to help people to understand how the camera plugin has been implemented in LImA this section provides some important information about the developer’s choices.

Camera initialisation

The SlsDetector plugin exports two kind classes: one generic SlsDetector::Camera class, with the common interface to slsDetector and slsReceiver classes, and detector-specific classes, like SlsDetector::Eiger which manage the particularities of each model.

First, the SlsDetector::Camera must be instantiated with the configuration file, and once the connection to the detector is established, a specific class is created depending on the detected type:

cam = SlsDetector.Camera(config_fname)
if cam.getType() == SlsDetector.Camera.EigerDet:
    eiger = SlsDetector.Eiger(cam)
    raise RuntimeError("Non-supported type: %s" % cam.getType())

hw_inter = SlsDetector.Interface(cam)
ct = Core.CtControl(hw_inter)

The raw images returned by the slsReceiver class might need to be reconstructed, like in the case of the PSI/Eiger detector. A LImA software reconstruction task must be then created from the LImA plugin and registered to the Core::CtControl layer:

if cam.getType() == SlsDetector.Camera.EigerDet:

corr = eiger.createCorrectionTask() ct.setReconstructionTask(corr)

Std capabilites

This plugin has been implemented in respect of the mandatory capabilites but with limitations according due to the detector specific features and with some programmer’s choices. We do not explain here the standard Lima capabilites but you can find in this section the useful information on the SlsDetector specfic features.

  • HwDetInfo


  • HwSync

The following trigger modes are currently implemented:

  • IntTrig

  • ExtTrigSingle

  • ExtTrigMult

  • ExtGate

The minimum latency_time and the max_frame_rate are automatically updated depending on the PixelDepth (4, 8, 16, 32), the ClockDiv (Full-, Half-, Quarter-, SuperSlow-Speed), and the ReadoutFlags (Parallel, Non-Parallel).

Optional capabilites

In addition to the standard capabilities, we make the choice to implement some optional capabilities in order to have an improved simulation.

  • HwShutter

Not implemented

  • HwRoi

Not implemented

  • HwBin

Not implemented


The main configuration will consist in providing the correct config file file to the slsDetector API. As mentioned before, the file is a list of commands accepted by sls_detector_put, and it should also work with the slsDetectorGui application.

Two important parameters define the image frame dimension:

  • PixelDepth:

    • 4 bit (not implemented yet)

    • 8 bit

    • 16 bit

    • 32 bit

  • RawMode:

If set to True, the image is exported to LiMA as given from the Receiver(s), without any software reconstruction.

How to use

The LimaCCDs Tango server provides a complete interface to the SlsDetector plugin so feel free to test.

For a quick test one can use Python, this a short code example to work with the PSI/Eiger detector:

from Lima import SlsDetector
from Lima import Core
import time
import sys

config_fname = sys.argv[1]

cam = SlsDetector.Camera(config_fname)
if cam.getType() != SlsDetector.Camera.EigerDet:
  raise RuntimeError("Non-supported type: %s" % cam.getType())

eiger = SlsDetector.Eiger(cam)
hw_inter = SlsDetector.Interface(cam)
ct = Core.CtControl(hw_inter)
corr = eiger.createCorrectionTask()

acq = ct.acquisition()

# setting new file parameters and autosaving mode
saving = ct.saving()

pars = saving.getParameters() = '/tmp'
pars.prefix = 'test_slsdetector_'
pars.suffix = '.edf'
pars.fileFormat = Core.CtSaving.EDF
pars.savingMode = Core.CtSaving.AutoFrame

# now ask for 0.2 sec. exposure and 10 frames


# wait for last image (#9) ready
lastimg = ct.getStatus().ImageCounters.LastImageReady
while lastimg != 9:
  lastimg = ct.getStatus().ImageCounters.LastImageReady

# read the first image
im0 = ct.ReadImage(0)

# cleanup in good order
import gc
del acq; gc.collect()
del ct; gc.collect()
del corr; gc.collect()
del eiger; gc.collect()
del hw_inter; gc.collect()
del cam; gc.collect()

A more complete Python script can be found under the camera/slsdetector/test directory.