filteron

Then the program automatically sets the correct number of filters in the beam to avoid the saturation of the detector. 

Note that in the case of very intense beams, like the direct beam, the maximum number of filters may not be enough to prevent the detector from saturating, and you will have to reduce the slits anyway. This is particularly true with aluminum filters, which have a low attenuation.

Another problem that we had is the following : if you stay on a Bragg peak and close the shutter, the system removes the filters. When you open again the shutter, the detector can be so saturated that the measurement of the dead-time will be completely wrong, and the electronics will not put ANY filters into the beam. This is an easy way to kill a detector (ours took 10-15 min to recover when this happened : it had seen so many counts that it was always seeing counts with the shutter closed). So if you close the shutter while you are on a Bragg peak, do a filteroff and a filterset 15 before you re-open the shutter : that’s safer.

Mode (3) is particularly useful when you do a L-scan on a Crystal Truncation Rod : you can measure both the low intensity part of the rod and the Bragg peaks in the same scan. It is also very useful for reflectivity scans. 

For alignments, mode (3) makes SPEC a bit slower but reduces the hassle of changing the attenuators by hand. Typically, a counting time of 0.1 s is too small for the attenuators to follow correctly. A safe counting time is 0.5 s.

Procedure for starting to use the filters under SPEC

- make sure everything is connected correctly (see below)

- put the 4 filter switches on the XIA card toward the left (red lights OFF, filters extracted)

- filteroff (in SPEC)

- unplug the attenuator card (the gold one) by extracting it by 1 cm from the rack

- unplug the upper 9-pin connector then replug it

- replug the attenuator card

- shclose (in SPEC) (closes the beam shutter)

- det_cal (in SPEC) This command should return values like :

00.084 4.5290.101

if it returns only zeros, try doing the “det_cal” command again and again.

- shopen (in SPEC) (opens the beam shutter)

- ct (in SPEC) (counts 1 second)

After this procedure, if you do a “filterset 1” in SPEC, the red light corresponding to filter 1 should come ON on the XIA filters card, and the “ct” command in SPEC should return “Attenuator = 1”.

For calculating Det_corr, one then needs to enter the attenuation factor into SPEC. 

For the automatic insertion of filters, one needs to setup the thresholds on the dead-time for inserting or removing the filters.

Here is the procedure to determine the thresholds (and also the attenuation factor) :

- put the diffractometer in a position where the detector is not saturated but the dead-time is large enough (typically 0.05). This should correspond to an intensity of about 55000 cts/s in EH1 (SIXCV) and about 30000 cts/s in EH2 (SIXC).

- filterset 0

- ct

this gives for example : detector = 31000 cts/s

dead-time = 0.045

- filterset 1

- ct

this gives for example : detector = 14900 cts/s

dead-time = 0.017

- filtersetup

this asks you first :

Threshold to insert filters (....) ?

Here you put a value slightly higher than the dead-time you measured with zero filters (for example 0.05).

Then it asks you :

Threshold to extract filters (....) ?

Here you put a value slightly smaller than the dead-time you measured with one filter (for example 0.015)

It then asks you :

Attenuation factor (...) ?

Here you put the ratio between the counts you measured with zero filter and with one filter (here 2.08 = 31000 / 14900)

and finally it asks for :

Serial line number in config (...) ?

Here you put : 

id3/ser032_1/15

if you are working in EH1 (SIXCV) and :

id3/ser032_1/13

if you are working in EH2 (SIXC). 

Note that you have to change the thresholds and attenuation factor each time you change the energy of the x-rays, the type of filters (moly, aluminum ...), or the type of detector (thresholds adjusted in EH1 may not work in EH2 because the detector is different).

Typical settings : at 17.14 KeV

- for SIXC (EH2), with molybdenum 0.2 mm filters :

filtersetup0.050.0152.71id3/ser032_1/13

- for SIXCV (EH1), with aluminum 0.5 mm filters :

filtersetup0.060.0252.04id3/ser032_1/15

After doing that if you type :

- filteron

and you go to an intense peak, SPEC should start inserting filters automatically as soon as the dead-time is larger than 0.05.

To go back to the manual mode you have to type :

- filteroff

What is the logic behind this procedure with the thresholds :

Imagine that you want the dead-time to stay under 0.06. So the electronics has to add one filter when the dead-time is above 0.06. But it also has to remove the filters when the dead-time becomes too low (that is : if by removing a filter one gets a dead-time that is smaller than 0.06, the system should remove at least one filter). So you would put the upper threshold at 0.06. The question is then : how much do I put for the lower threshold ? Basically, you have to estimate by how much you reduce the dead-time when you add a filter, being originally with a dead time of 0.06. So you go to a place where the dead-time is around 0.06, you add a filter, and you measure the new dead-time (for example 0.02). The first idea would then be to put 0.06 for the upper threshold and 0.02 for the lower. But if you put exactly those values, when you go to a point with a dead-time of 0.065, the system will add a filter, but then the new dead-time will be slightly above 0.02, so the system will remove again the filter it just added, and this will go on forever : the system will not be able to decide if it should insert or not the new filter. That’s why the distance between the two thresholds has to be slightly larger than the measured interval : in the case taken here, one would typically choose 0.065 for the upper threshold and 0.015 for the lower threshold.

What could be improved :

In the method given above, the attenuation factor is measured only for one foil (filter 1). One could object that the different foils could have slightly different attenuation factors, if their thickness varies a bit from foil to foil for example, or if a foil is damaged etc... It would then be more correct to measure separately the attenuation factors of the 4 filters. This is not so easy for filter number 4, which has typically an attenuation factor of 2⁸ = 256. 

Also, in the method given above, the attenuation factor is measured with a non negligible dead-time of the detector. One could object that the measurement is not accurate, since for accurate intensity measurements, the detector should always be in the linear region, with zero dead-time. There are two ways to get around the problem : First, one can do the two intensity measurements (with and without filter 1) in a region where the intensity is low enough to have almost zero dead-time. However, this requires to count longer to have a good statistics on the intensity measurement done with filter 1 in the beam. If one wants in addition to measure the 4 attenuation factors of the 4 filters separately, it is clear that there will be a big problem with filter 4 : if the detector is in the linear regime without filter, the intensity will be reduced to almost nothing when filter 4 is in the beam. The second method, to get more accurate measurements of the intensity, is to compensate for the dead-time in the measured intensity : for example, if one measures 30000 cts/s and a dead-time of 0.06, the real intensity is (1/ 0.94) times 30000 cts/s, i.e. 31915 cts/s. Of course, for doing this you need to be sure that the dead-time measured by the detector electronics is reliable : for heavy saturation of the detector, the dead-time is no more reliable.

The “blmenu” command :

Before using blmenu, make sure that no filter is in the beam. 

If you type the command :

blmenu (in SPEC)