Splitting analytical lines off into a new 'sample'; utility of 'grain' concept?

[orlandin]

Hello, all!

I was pleased to discover the 'grain increment' functionality in Digitize last night, thinking that would be a slick way of using the same Sample-level attributes but having PFE calculate homogeneity and T-test values on a per-grain basis of the same material. However, it does not seem like the grain designation appears at all in the Analyze window. Is it possible to treat lines grouped by grain #, without giving them a sample-level distinction? Alternatively, how can I peel lines out of a Sample into a new Sample?

Best,

Phil

[Probeman]

Interesting.  The "grain number" feature in the Digitize! window was intended for exactly the opposite purpose! 

That is, if one had a number of phases to digitize, it would normally be expected that one would digitize them as separate samples, e.g., Un 1 Feldspar Ca-rich zone, Un 2 Feldspar K-rich zone, etc. That way one could apply different mineral calculations to each phase.

But if for some reason one wanted to average a number of different grains of similar composition, but separated by epoxy in the mount, the "grain number" feature forces the software to blank the beam at acquisition time in between the grains, even though they are all within the same digitized position sample.

In fact the "grain number" isn't even saved to the acquired sample, it is merely an attribute of the position database for acquisition.  For example, a standard mount where it is desired to average a number of grains for a standard to obtain a more accurate composition (i.e., the grains are somewhat variable in composition.  And the grains are separated by epoxy, e.g., Smithsonian standards).

But if you actually have different compositions, just digitize them as separate position samples and then they can be treated with different calculation options from the Analyze! window.

But another trick that I often use to apply different post processing options to the same dataset is to make multiple copies of the MDB file, and turn on/off the different options, e.g., MAN vs, off-peak etc. This is how I treated the data from the Ti-V interference run I did recently acquired using off-peak (and also turning on MAN in post processing) for LiF, PET and aggregated LiF and PET datasets):



You might try doing something similar except disabling/enabling different lines in each copy of your database...

[orlandin]

Ah ha, thank you! I had never really thought through the subtleties of what to consider a Sample in PFE until now, and this makes sense. This concept of Sample is also reflected in the very handy 'automate image acquisition' that is a per-sample basis that I just discovered, and I'll be sure to explain the idea using both features to users so that they can plan their data collection most efficiently. Thank you for explaining!