Page 1 of 1

Can this be done??

Posted: Thu Sep 14, 2017 4:07 pm
by EmpowersBane
I need Empower to perform a calculation on 2 working standards:

Standard 1- 1 Injection, contains Caffeine Label A
Standard 2- 1 Injection, contains Caffeine. Label B

I need to perform a standard compliance on the Caffeine area by doing (and I know there's some debate, but its written in our SOP so I'm tied to it):

[Area of Caf in Inj 1- Area of Caf in Inj2]/[Area of Caf in Inj1+Area of Caf in Inj2/2]*100.

No matter which way I construct a CF, I cant get it to work, I tried
(A.1.(Area)-B.1.(Area)/(A.1.(Area)+B.1.(Area)/2)*100 but it doesn't calculate the second part correctly, even after I move around a few things.

Any ideas , I would be grateful?

Re: Can this be done??

Posted: Thu Sep 14, 2017 4:13 pm
by DR
Assuming that these injections are sequential, how close is Empower's "Deviation" value for B (bring up the cal window in review)?

Re: Can this be done??

Posted: Thu Sep 14, 2017 6:13 pm
by EmpowersBane
Thanks for the reply, I will check this out, we don't usually use the cal curve in day to day use, are you saying there is a "Deviation" value calculated for the Std B? Can I obtain this value in the cal screen from a list of options?

Yes, we would run the injections sequentially.

Re: Can this be done??

Posted: Fri Sep 15, 2017 6:33 am
by ydna1977
Hi

I think this should work too

You may want to check the rules of the syntax for intersamples CF

All or nothing enabled or disabled may help

Re: Can this be done??

Posted: Fri Sep 15, 2017 4:40 pm
by DR
EmpowersBane wrote:
...Can I obtain this value in the cal screen from a list of options?


There's a column for it when you open up the CAL window and choose a peak name. I'm not sure whether it can be extracted and dropped into a report, search fields under CAL objects in the report editor.

Good luck,
Dan

Re: Can this be done??

Posted: Fri Sep 15, 2017 4:59 pm
by EmpowersBane
Thanks very much for the replies, both of you. Very helpful. :P

Re: Can this be done??

Posted: Thu Sep 21, 2017 1:44 pm
by ZnBr2
I may be very late to the game; however, for a few of the samples that we use an external calibration for I use the Empower3's default "Calibration Report" and the tables there indicate the %Deviation for each constituent for each standard used in the curve.

You may have to go back to your installation software, look for a disk with "Default Projects" to find this report.

Re: Can this be done??

Posted: Thu Sep 21, 2017 6:01 pm
by EmpowersBane
Thanks ZnBr2, but for that to work wouldn't you need to enter the concentrations of the standard into the component editor in order to obtain a calibration curve? Typically when we run standards to calculate compliance we only use areas as values so we never enter a component editor.
If they WERE entered, I could just check my pair of standards (Std soln A and B) and then check the % deviation result, is that what you mean?

Re: Can this be done??

Posted: Thu Sep 21, 2017 8:06 pm
by ZnBr2
EmpowersBane,
Simple answer, yes, I think you understand where I was going with this - you would need to build the curve. So maybe this is not the route you want to take, I was trying to follow in the last post by Dan. As I mentioned, late to the party (o_-)

Looking back at OP:
SOP Calc:
Your attempt:
(A.1.(Area)-B.1.(Area)/(A.1.(Area)+B.1.(Area)/2)*100

You're saying that the second half isn't calculating, do you mean:
A.1.(Area)+B.1.(Area)/2 ?
What I read here is:
Divide B.1.(Area) by 2
Then
Add this value to
A.1.(Area)

Is that what you intended to do? I suspect you want the calculated average of the areas between "A" and "B" in the denominator, this is a relative percent difference calculation - yes?

So lets start with that change and add an explicit grouping of the result:
(I've learned the hard way to explicitly group when working with databases, programing languages, etc... )

( ( A.1.(Area)+B.1.(Area) ) / 2 )

Explicitly group your first numerator:
( A.1.(Area)-B.1.(Area) )

Put these together along with 100 to get percent:
( A.1.(Area)-B.1.(Area) ) / ( ( A.1.(Area)+B.1.(Area) ) / 2 ) * 100

If you would like to give that a whirl (I've added a bit of spacing in there so that the eye doesn't get lost so a direct cut and paste may not work) and please let me know if we're on the right track.

I do not think you should need to have the "Summarize Custom Fields" selected in your sample set with this calculation; however, you might take a look at my thread here (this is the final post)
http://www.chromforum.org/viewtopic.php?f=4&t=46831#p225394
As you can see in that thread, this simple thing gave me months of heartburn.
(@_@)
-Z

Re: Can this be done??

Posted: Fri Sep 22, 2017 5:36 pm
by EmpowersBane
Thanks very much for your reply ZnBr2, much appreciated!

On the standard curve point, as a company we don't tend to calculate our standard compliances that way. We work off our weight adjusted areas and do that calc (Wt adjusted A-Wt Adjusted B)/(Mean of Wt Adjust A + Wt Adjust B)*100 and our result must pass a spec, usually NGT 2.0%.I know some people do it different but that's the way our SOPS are written.

Would the standard curve still work if I put in the concentrations of A and B in the comp editor? its only 1 injection each (A and B) so would the % deviation still apply?

Thanks for your tip on how to group the calculations. Yes, that's exactly what was happening with my custom field, it was doing the 1st part right but not the second part. I actually had to make 3 CFs to get it to work- one for the difference one for mean and one for the calculation of one over the other x 100.

When I'm logged in next I'm going to edit the formula to get it in one, as you suggest. I'm confident it will work. Thanks again.