DIP2000 user manual

Previous: Alignment of the DIP2000 Next: Installation of software for the DIP2000 system

Calibration of the DIP2000

The calibration of the DIP2000 is done by measuring a fluorescent sample at an offset position of 30 degrees. For Cu radiation this sample is a thin film of amorphous iron, which generates Fe-radiation. From this image a response curve of Intensity versus plate radius can be calculated. This response curve is then used to make a calibration file.

The actual IP position during the exposure is like below (in the case of a left-hand side installation type, viewed from the sample position).

An image taken at this position looks as follows:

For non-Weissenberg type systems, where the IP offset position goes from 0 to 20 degrees, the IP will be partially shaded by the cover of the scanner box. This is not a problem, since there will be enough data points left which can be used for the calibration program.

The following programs and files are used in calibration:

Files:

Programs:


Calibration procedure

Do not stop here, since by having made this change, all image files written, will be uncorrected !!!

Making an initial calibration image

This will generate an image which is partially exposed and looks like the image at the top of this page.

The data on a circular line will be used for the calibration of the image plate. This line is a circle through the centre with the centre of the primary beam as the centre. If the beamstop was not positioned vertically, the shadow of the beamstop will partially cover the dashed line and consequently this part cannot be used for calibration. The line should also not be covered by shadows of other parts. If this cannot be avoided, an option exists to use only part of the data, but it is advised to put the beamstop such, that all data can be used.

Check if the image looks OK. For an image to be used for calibration, the exposure time to be used should be in the same order as the time used for normal exposure i.e. 20 to 30 minutes.

Making the calibration image

Make the long still exposure with the following parameters:

After the exposure and readout there will be an image written to disk. Click Display and the image will appear.

Zoom in to the centre of the plate by selecting the area with the left mouse button. Right click the Main Control button and select numeric dump. Write down the coordinates of the centre

For calculation of the calibration curve proceed as follows :

Calculating the calibration curve

Open a Unix winterm window and run the program calG to calculate the uniformity correction table.

%calG Fe_nocorr.ipf

The input of the program depends on the setup of the image and of the DIP system.

Distance between IP_change_axis and IP centre:

for DIP2020 series : 140
for DIP2030 series : 200

IP-offset : This is dependent on the position of the beamstop. We consider two cases : A right handed and a left handed system. Note that during the experiment the beamstop should have been put vertical. Check the image and select from the table below:

Pattern type
 IP offset Usable region
a
30
whole
b
-30
whole
c
-30
half -1
d
30
 half-1

Right handed system
Left handed system
(a)
(b)
(c)
(d)

Centre_x and centre_y of IP[mm]:
for DIP2020 series : 100 100
for DIP2030 series : 150 150

Radius of blind region at IP center : Area to be blacked out in the middle of the plate e.g. 1 mm
Fringe region : Small area (e.g. 0.5 mm) to prevent very low edge values to be taken into the scaling
Thickness of reference ring : Width of the ring of data used for calibration e.g. 3 - 5 mm
Scale factor : Use 1

An example of a log file of the program follows below:

no143:/disk114/xdip/cal% calG Fe30off_nocorr.ipf

*** G-factor calculation program ver.2.1 (Jan. 27 1997) ***
  (Dimension limits: 0.001[mm] < pixel_size < 1.0[mm])
(98.0[mm] < IP_radius < 500.0[mm])
Input file = "Fe30off_nocorr.ipf"
Output file = "DIP_CORRECT_PARM"
dtype= 321 (shift= 5, offset= 0)
pixel size= 0.1000[mm] IP radius= 150.0[mm]

Enter distance between IP_change_axis and IP center[mm]: 200
Enter IP offset(left rotation is positive)[degree]: -30
Enter center_x and center_y of IP[mm]: 150 150
Enter radius of blind region at IP center[mm]: .5
Enter fringe margin[mm]: .5
Enter thickness of reference ring[mm]: 5
Usable region on the ring(0:whole, 1:half_1, 2:half_2): 0
Enter scale factor: 1
distance=  200.0[mm]
offset=   -30.00[degree]
IP center=  150.0 150.0[mm]
blind R=   0.50[mm]
f. margin=  0.50[mm]
ring thick.= 5.00[mm]
using region=whole
scale factor=1.000
center and radius of reference ring[mm]= (123.205 250.000) 103.528

------------------------------------------------------
read IP data from 1431 to 3000
read IP data size = 3000 * 1570
number of processing pixels = from 1431 to 3000
number used pixels = 173598
number of zero intesity pixels = 0
------------------------------------------------------
Place the result file(DIP_CORRECT_PARM) in directory $DIPHOME/data as well as
DIP_CONV_PARM for subsequent MakeCorrectData processing.

Check the output of the program. Make sure the centre of the plate is correct. The position which is given as the centre of the plate should be the same as the centre in e.g. "numeric dump" of the IP display program or in "d-value etc.".
For DIP2020 systems, the pixel numbers should be multiplied by 0.08, for DIP2030 systems, the pixel numbers should be multiplied by 0.1. This should give the mm coordinates as given by the calG program. Also check that the number of zero intensity pixels is low.

Check the file DIP_CORRECT.temp. The last values of this file should be very close to 1. If this is not the case, rerun the program calG, and input the last value of DIP_CORRECT.temp as scale factor.

To make things a liitle cleare, the contents of the output file can be plotted.

Create a file with pixels and intensities as follows:

awk  '{print $1, $3}'  DIP_CORRECT.temp > calG.plt

This file calG.plt can be plotted as follows:

Display the flatfield image with the xpress display program. Open a zoom window in the intense part in the image.
Go to the Profile level in the display program.
Draw a line in the intense part. This will open up the XGraphV program.

By Right-clicking in the Graph window a selection menu will appear.
Select Open Text data and select the file calG.plt and press Load.
Now right click again in the Graph window and select Axes.

This will then show the response curve of the image plate.

It should look similar to the image below.

The file DIP_CORRECT_PARM must be placed in $DIPHOME/data.

Then you must run the program MakeCorrectData to generate a new file CorrectData. This file should reside in $DIPHOME/data. The previous file CorrectData (for no-correction) may be overwritten :

% cp DIP_CORRECT_PARM $DIPHOME/data
%MakeCorrectData -15 > $DIPHOME/data/CorrectData

When there are fine fluctuations in the plot of data in DIP_CORRECT.temp, the curve can be smoothed by specifying an option like -15 (this will tell the program to average 15 points around the centre)

The file CorrectData is the final result of the calibration.

Checking the calibration

Make a new still image of the iron scatterer, now corrected. The intensity should vary radially when changing the color scale of the image.

Also on this image the program calG can be run. The resulting curve of DIP_CORRECT.temp should be flat with fluctuations no bigger then a few precent.

(C) Nonius BV 1999, Latest change: 28 May 1999, A.J.Schierbeek