Wish List for PFE Features

[John Donovan]

Hi Anette,
You're saying you want the user to be able to edit these delays within the application?    It would certainly be possible, though maybe in the Acquisition Options would be a better place?  Curious why you don't just want to modify these delays in the probewin.ini file until they work properly and leave it at that.  This would be for the delays *after* setting the column parameters? Correct?

May I ask why you want to change these delays in the application?    There is already a TDI incubation and/or carbon decontamination delay in Acquisition Options...
john

Hi John, I guess it actually does not make much sense to have them changeable on a sample per sample basis. How about having them on display though as in my dialog layout proposal but not editable. This would accomplish both that the user would be aware that there is a delay being applied upon changing conditions and allow that the values can be critically evaluated (whereas otherwise they would be "hidden" in the probewin.ini file).

What do you/everyone think?

If what is desired is seeing these delays then how about I just add some text to the status bar if it's adding a delay after setting these items?
john

[John Donovan]

Hi John, I guess it actually does not make much sense to have them changeable on a sample per sample basis. How about having them on display though as in my dialog layout proposal but not editable. This would accomplish both that the user would be aware that there is a delay being applied upon changing conditions and allow that the values can be critically evaluated (whereas otherwise they would be "hidden" in the probewin.ini file).

What do you/everyone think?

Hi Anette,
Please update PFE and open the Combined Conditions dialog.  It now displays the set condition delays and also displays a new message in the status bar as it's setting them.

Let me know if this works for you.
john

[glennpoirier]

Hi John,
This may be already implemented but I haven't found it.

Would it be possible to select samples by text in the Automate and Analyse windows? Say I have a run that has multiple sample setups and I want to select only the CPX to apply a setup. I'd like to be able to select samples based on their labels (i.e. I'd enter "CPX" in the search box ) and all the sample labels containing the string "CPX" would bet selected. A simple text entry would be great, but wildcards would be nice too (not to mention REGEX).

This comes about because I'm in Cambodia reprocessing some data and I don't have a mouse to use with my laptop, so multiple selection is a pain.
Cheers from balmy Banlung! (29 C at 9:30 high of 35! not that I'm complaining, 35 cm of snow at home!)

Glenn

[John Donovan]

Hi John,
This may be already implemented but I haven't found it.

Would it be possible to select samples by text in the Automate and Analyse windows? Say I have a run that has multiple sample setups and I want to select only the CPX to apply a setup. I'd like to be able to select samples based on their labels (i.e. I'd enter "CPX" in the search box ) and all the sample labels containing the string "CPX" would bet selected. A simple text entry would be great, but wildcards would be nice too (not to mention REGEX).

This comes about because I'm in Cambodia reprocessing some data and I don't have a mouse to use with my laptop, so multiple selection is a pain.
Cheers from balmy Banlung! (29 C at 9:30 high of 35! not that I'm complaining, 35 cm of snow at home!)

Glenn

Hi Glenn,
Well I'm going to complain! It's been raining in Oregon for the last 3 months- non-stop it seems sometimes.   

You can multiple select samples using the up/down cursor keys and the shift key, but for more flexibility than that, I suspect you'll have to break down and buy yourself a mouse.   
john

[Ben Buse]

Hi John,

Two ideas.

(1) Would it be difficult to have an option of converting a standard sample into an unknown - the example is I collected data as a standard and I want to use it for the blank correction (for which I need an unknown).

(2) When exported apfu (atoms per formula unit) I have to check the option in calculation options. It would be nice to be able to export the data to dat.file and excel without having to check the option in calculation options - for when checked in calculation options in Analyze window instead of seeing wt. % you see apfu.

Thanks

Ben

[Ben Buse]

Driving home I was thinking...

Regarding (2) I think it would be best if instead we had a check box saying "Display as formula units". So if you want to export you check calculate based on formula but don't check display as formula units. - Same as for display as oxides

What do you think, thanks

Ben

[John Donovan]

(1) Would it be difficult to have an option of converting a standard sample into an unknown - the example is I collected data as a standard and I want to use it for the blank correction (for which I need an unknown).

Hi Ben,
I understand this idea.  I've had it myself.  The issue is that generally we should collect our "blank" sample under the same conditions as our unknowns. Small differences in counting/beam current/etc. could lead to systematic biases in the "blank" calibration.  So for now I have resisted this request, but I'm willing to discuss it.

(2) When exported apfu (atoms per formula unit) I have to check the option in calculation options. It would be nice to be able to export the data to dat.file and excel without having to check the option in calculation options - for when checked in calculation options in Analyze window instead of seeing wt. % you see apfu.

This is easy.  Just right click the sample(s) in the Analyze! sample list and select the output type you want. usually I use the user specified output format...

Did you also know that by double-clicking any analysis line in the lower data grid in the Analyze! window you can can see *all* the data types for a single data point as shown here:

http://probesoftware.com/smf/index.php?topic=110.msg1141#msg1141

john

[Ben Buse]

Hi John,

Did you also know that by double-clicking any analysis line in the lower data grid in the Analyze! window you can can see *all* the data types for a single data point as shown here:

http://probesoftware.com/smf/index.php?topic=110.msg1141#msg1141

Thanks - thats good to know

This is easy.  Just right click the sample(s) in the Analyze! sample list and select the output type you want. usually I use the user specified output format...

Your right! It works! SORRY I thought when I last tried it it didn't work unless "calculate on formula basis is checked" -but I'm glad I'm wrong.

Select user specified output format... by either right click or from the output menu. Tell it to output formulas -



It outputs formula even when "calculate on formula basis" is not checked. Default cations =8, no end member when not configured.

As configured here



Thanks

Ben

[John Donovan]

Your right! It works! SORRY I thought when I last tried it it didn't work unless "calculate on formula basis is checked" -but I'm glad I'm wrong.

Select user specified output format... by either right click or from the output menu. Tell it to output formulas -



It outputs formula even when "calculate on formula basis" is not checked. Default cations =8, no end member when not configured.

I do try to be helpful once in a while...   
john

[Karsten Goemann]

Regarding the using/converting a standard to an unknown to be used as blank, my take would be that it is already possible to change the acquisition conditions between unknowns. So nothing's really stopping me from picking an unknown as a blank that has been acquired with different conditions, right?

So the limitation to only being allowed to use unknowns as blanks isn't really that much "safer", unless there would be some sanity check to ensure the conditions were the same. And with the latter it should then also be possible to use standards as blanks. Having said that I often acquire standards and unknowns with the same conditions anyway.

[John Donovan]

Regarding the using/converting a standard to an unknown to be used as blank, my take would be that it is already possible to change the acquisition conditions between unknowns. So nothing's really stopping me from picking an unknown as a blank that has been acquired with different conditions, right?

So the limitation to only being allowed to use unknowns as blanks isn't really that much "safer", unless there would be some sanity check to ensure the conditions were the same. And with the latter it should then also be possible to use standards as blanks. Having said that I often acquire standards and unknowns with the same conditions anyway.

Hi Karsten,
That's a good point.  I could add code to check for similar conditions.   So check for matching column conditions and element setup including acquisition order and count times?

This brings up the interesting question of shouldn't our standards always be run at similar conditions to our unknowns?    My feeling is that we should, at least if the standards are compositionally similar to the unknowns. But for trace elements we can of course use almost any standard and instead just concentrate on the background measurement accuracy. As you have demonstrated in several presentations, this is because the background measurement method(s) of our unknown always dominates the accuracy (and precision) of the trace elements in our samples, as opposed to the accuracy of the matrix corrections.  For those new to thinking about these issues, please consider that the accuracy of the matrix correction of Ti Ka in SiO2 is probably better than 2% relative or so, but the background intensity measurement accuracy can be hundreds of percents in error, due to any number of background artifacts at trace levels.

So for example, I usually count say 40 sec on my TiO2 standard when measuring Ti ka in quartz and at 30 nA, but count say 640 sec and 200 nA on my unknown quartz samples *and again* also 640 sec and 200 nA on the blank standard (e.g., synthetic SiO2).  So I do need to add a check for similar conditions at least.
john

[Karsten Goemann]

Yes, for traces I usually adjust the beam current between standard and unknown.

Would that check for "similar conditions" also have to include the background correction, i.e. off-peak positions, correction type etc? In the case of blanks in trace element analysis even a slight difference in the background correction will have a big effect.

[John Donovan]

On the other hand, is it too much to ask the user to measure a few points on their SiO2 standard using the unknown conditions?

I mean, even if one weren't intending to actually utilize the blank correction, for example you'd still want to see if you can measure zero Ti in the blank standard... and SiO2 won't be a standard anyway since you'd probably wouldn't be measuring Si, and instead be specifying an SiO2 matrix by difference...
john

[Karsten Goemann]

Sure, that's fine.

I guess the point I'm trying to make is just that while I understand the reasoning behind the limitation that a blank standard has to be an unknown I still question it. I can change conditions anytime between samples. So even when I use an unknown as blank I still need to make sure that the conditions were the same. And it is easy to run an standard at the same conditions as an unknown and vice versa anyway. So why not run a standard as a standard? ;-) But it is perfectly usable the way it is.

[John Donovan]

I hear you, but I guess I just don't see the point of providing this additional flexibility if it "encourages" the user to not do the right thing. 

I think the "limitation" of not allowing a standard to be used for the blank correction is essentially insignificant under normal circumstances (we just need to run an extra unknown sample which so happens to be one's blank std using the same sample setup as one's other unknowns that we are automating). This ensures that the blank standard tends to be run at the same conditions as the unknowns, which in our example of trace element standards using different conditions as described above, would *not* be ensured...

Yes, one *could* create a different unknown sample setup from the other unknowns, and then use it to run the blank standard differently, but that would require *extra* effort.  The way it is now I think is more "natural" for having the blank std and unknowns all run using the same conditions, because they need to be the same sample type...

The only situation I can think of where allowing a std sample to be utilized as a blank sample would be worthwhile, would be a situation in which one forgot to run the blank std as an unknown.   But even in that case it's very likely, for example, that either the SiO2 standard was not run at all (when analyzing Ti in quartz, why would one run an SiO2 std?), or that the standards were run at quite different conditions than the unknowns, for example, using 30 nA on the standards and 200 nA on the unknowns...

Sorry to be so pedantic about all this.
john