Previous: makefries: make an undistortion calibration file
Next: caldark: Calibrate for shutter closed effects

makedetalign: measure detector alignment

The "makedetalign" program can measure the detector alignment, and creates a calibration file that allows other collect programs and the evalccd software to take the detector alignment into account.

Prerequisites

The program needs to be able to perform distortion correction, i.e. it needs a file "distorpol.vic" either in the current directory, or in the calibration directory. See the makedistor program for more information.

Usage

The program is invoked by the command line:

makedetalign [options] <imagefilenames>

or (in an empty directory):

makedetalign [options] make

Different image file types are recognized from their names. The images specified should be a recent series of images for the ammonium bitartrate test crystal representing a 360 degree phi scan (either 180x2 degrees or 360x1 degrees will do fine; 90x4 degrees for Cu radiation is fine too).

If run with the "make" argument, the program will start by collecting the frames it needs for the calibration. By default on a MoKa KappaCCD it will collect 180 single images of 2 degrees wide, at dx=30 mm, each for 10 seconds. For CuKa the measurement will consist of 90 images of 4 degrees each measured for 3 seconds. These values can be changed by the options "frametime=##", "frameangle=##" and "dx=##".

The analysis will start by reading all files, locating diffraction spots. After that, partial diffraction spots in adjacent frames will be joined. Finally a series of quartets (left and right impacts of each reflection and its Friedel opposite) will be collected. Based on the positions of each individual reflection in the quartets, the calibration is performed and written to the file "detalign.vic" in the current directory. To be used by other programs, this file can be copied to the CCD_DATAHOME directory on your system.

Problem solving

Dx misset

The calibration performed is slightly dependent on the calibration of the detector to crystal distance. Ideally, "dx" should be calibrated within something like 1-2 mm before running makedetalign. The problem is that the caldx program requires a detector alignment to run. If dx is too far off, you will get a warning message or a fatal error message telling you that there are not enough quadruplets. There are two partial work-arounds:

You can run "makedetalign dx=xxx" where xxx is an approximation of the dx value where the measurement really has been performed (as opposed to the faulty value in the KappaCCD headers). If you succeed to correct the problem this way, this will result in a detalign.vic file that corrects for this error. If you correct the missetting of dx later, you will need to edit detalign.vic, replacing the "detzerox" value by 0.0
You can try to do a rough dx calibration before running makedetalign using "caldx rough". See the caldx manual for more help. This can only work if you do it before collecting the makedetalign data.

Detector mounted off-center

There is another reason that could cause that no quadruplets are found at all (or less than 10): the detector is mounted seriously off-center.

If you suspect that is what happens, do the following:

Create a file "detalign.vic" in the current directory, containing:

     DETZEROX 0.0
     DETZEROY 0.0
     DETZEROZ 0.0
     DETROTX  0.0
     DETROTY  0.0
     DETROTZ  0.0

Run "ndisp s01f001.kcd", and position the mouse cursor to the center of the beam stop. Now read the values "mmx=" and "mmy=" at the bottom of the screen.
Edit the "detalign.vic" file, and put the "mmx=" value you found at DETZEROY, and the "mmy=" value at DETZEROZ. Please note that the XYZ of the world coordinate system do not correspond to the XY of the detector coordinate system.
Run "makedetalign" again.

Please note that if the detector is mounted off-center more than approximately 1.5 mm, the anti-collision software and strategy calculations will produce sub-optimal results.

Example output

% makedetalign 
NOTE: Using local file 'detalign.vic' for detector alignment
NOTE: Pixel variance: 486 listed
s01f001.kcd;  23 peaks found, 46 trials, 3.07 seconds
s01f002.kcd;  41 peaks found, 63 trials, 3.31 seconds
[...cut...]
5230 raw peaks located
NOTE: Joining partials around frame 10
NOTE: Joining partials around frame 20
[...cut...]
NOTE: Joining partials around frame 170
NOTE: Joining partials around frame 180
Joined spots together to 2757 reasonably strong reflections
NOTE: Using local file 'detalign.vic' for detector alignment
=============== REFINE CYCLE 1 ======================
NOTE: Quad number 100 found at phi=-162.0
NOTE: Quad number 200 found at phi=-140.0
NOTE: Quad number 300 found at phi=-107.1
NOTE: Quad number 400 found at phi=-80.0
NOTE: Quad number 500 found at phi=-42.0
561 Quadruplets located
NOTE: Preliminary refinement
Average= 0.000+/- 0.001,  0.000+/- 0.001
detzeroy,z now= 0.807738623226 0.871068500284
NOTE: Refinement
Average= 0.000+/- 0.001,  0.000+/- 0.001
detzeroy,z now= 0.807738708435 0.871068509455
Calibration data are stored in 'detalign.vic' in the current directory.
=============== REFINE CYCLE 2 ======================
NOTE: Quad number 100 found at phi=-162.0
NOTE: Quad number 200 found at phi=-140.0
NOTE: Quad number 300 found at phi=-107.1
NOTE: Quad number 400 found at phi=-80.0
NOTE: Quad number 500 found at phi=-42.0
561 Quadruplets located
NOTE: Preliminary refinement
Average= 0.000+/- 0.001,  0.000+/- 0.001
detzeroy,z now= 0.807738673192 0.871068458031
NOTE: Refinement
Average= 0.000+/- 0.001,  0.000+/- 0.001
detzeroy,z now= 0.807738698476 0.871068467108
Calibration data are stored in 'detalign.vic' in the current directory.

NOTE: Successful run. Program is happy with the results

Previous: makefries: make an undistortion calibration file
Next: caldark: Calibrate for shutter closed effects