Focus drift during line acquisition

[Malcolm Roberts]

Folks,
Well...this may seem screamingly obvious to some, but as someone says the only stupid question is that unasked. I set up a couple of traverses yesterday across a couple of large grains involving 40 spot analyses with a focused beam equidistant between the start and stop positions. Both the start and stop were in focus. However, I drifted in and out of the lab to watch things and noticed that focus deteriorated as the traverse progressed and was certainly well out of focus at the end despite the confirmation at the start. I even refocussed at the start of the next point if I happened to be there at the right time. The second profile was even worse starting and ending well out of focus. Data was crap for the most part to say the least.
How can I prevent this from happening? I am aware of limitations of flat surfaces during sample prep...
Cheers,
Malc.

[John Donovan]

Have you tried the different Auto-focus options available from the Automate! window?

http://probesoftware.com/smf/index.php?topic=112.msg419#msg419

john

[Malcolm Roberts]

Hi John,
I have noticed an "auto-focus" button in the software and wondered whether it would be the solution to my ills...... I will have a look at your link.....
Thanks....

[Malcolm Roberts]

Hi John,
This is not entirely the panacea. Another question begs an answer and that is after storing a start position for a new unknown line traverse, and commencement of that automated acquisition, the instrument does not set the Z for the focus as it was initially digitised, rather randomly goes to some z value more or less close to but not in focus as the original digitised point. I noticed along the last traverse that the focus drifted out and certainly did not finish at the value originally input for the stop point. Any wise words on why this is happening?
Cheers,
Malc.

[John Donovan]

Hi Malcolm,
Just go to the Acquire! Acquisition Options dialog and uncheck the "Use Random Z Positions" check box!     Ha, Ha!  Just kiddin'!   

Seriously, it sounds like a stage reproducibility problem.  If it is a mechanical stage problem, you should see the same behavior with the JEOL software.

[Malcolm Roberts]

Haha.... good idea. One thing I am going to do when I get a chance is to check for crystal alignment by running some on peak counts as a function of z per crystal per spectrometer. If there's anything tweaking to do this should show up nicely........

[John Donovan]

The Startwin application has a nice feature for this alignment procedure under Peak/Scan Option button...

[Malcolm Roberts]

That's new to me. If I recall, in PfW under startwin, I got the data and plotted it up in grapher following manual acquisition. Maybe this was the old 733 though.......
Meanwhile, I have recalibrated the JEOL stage and will see what happens....... watch this space, but I am not very optimistic........... It may be engineer time......
Cheers,
Malc.

[Malcolm Roberts]

Problem seems to have evaporated after a full system reboot.......