Empower 3 calibration curve equation coefficients in CFs

Hello!
So when Empower 3 makes a calibration curve (first order linear in my case) it also makes an equation, that it can display on report methods under calibration plot. As i understand it, in Emp3 this equation is expressed as y=B*x+A, where B and A are Polynomial Coefficients.

I would like to get these coefficients into a custom field for other calculations, however, a custom field with the formula:  A  or B (Peak, real, result set only, all or nothing) returns these values incrementally, meaning, the value is different for each injection of my calibrated stadards. I can see the correct coeficcients in Review->Results->Peak calibration curve table, however my custom field reaches this correct value only at the last injection of my calibration.

Since the number of lines and injections in my calibration can vary, i need a reliable way of always getting the correct equation coefficients in my custom field, is there a way to do so?

P.S. these sample sets only consist of the calibration itself, there are no samples form which i could try to pull the coefficients

Best Answers

  • Accepted Answer
    What other calculations do you need the A and B in? Is it the final value of A/B or the incremental value? Can you give an example of the further calculation required so I can have an idea of how you want to specify the A and B? Thanks.
  • Accepted Answer
    Hi, i need to find the final A and B value for the entire calibration curve. What i'm trying to calculate is to solve X when Y=0.

    Basically it's for calibrating our autosamplers, 3 injections for each volume (1, 3, 5, 10 μL and so on) and then check the linearity and deviation from a calibration graph of area VS injection volume. However our HPLC's have different injection ranges, some go up to 50μL, some to 200μL, so i'm having trouble getting the proper A and B values every time.

    For now i found a workaround, each set of 3 injections gets an unique, consecutive label, then i run a summary function through those labels to find the max injection volume. From there i can identify the last, largest injection set and pull the A and B values from the third injection of that set. A big, lame line of boolean checks, but since the injection volume increments and the number of injections in each set are constant, it seems to work fine, even though it seems like a really inefficient way of doing it, but at least it's easy to get my results in a way that's easy to use in a reoport method.

    Perhaps there's some sort of summary function, that could find the final A and B within a label/consecutive labels?
  • Accepted Answer

    Thanks for that. Yes, a few Bool equations seem to do the job but are messy. In general, the more Bool functions in an equation, the more inefficient the output. Since the number of injections is always 3, and you want to pull the final A/B from each set of 3, you could do a peak, real CF result set only with formula of SAME.%.(A*(EQ(Injection,3))) or SAME.3.(A) if its quicker. That worked perfect for me when I tested it on a sample set with 5 sets of standards with 3 injections each- you can label each set whatever you want, as its covered by SAME. Just don't repeat the same label as that could affect the output. So I called my set S1 up to S5 and that formula returned the correct A value for Injection 3 of each set. Repeat the formula for B, that's 2 custom fields so far.

    A final custom field could be used for the overall final A/B values for the last injection of the last set. Give your label for this a unique title like Final. Then a peak, real cf as above with formula Final.3.(A) and Final.3.(B). This will return the final value of the calibration curve for A and B.

Answers

  • >A final custom field could be used for the overall final A/B values for the last injection of the last set. Give your label for this a unique title like Final. Then a peak, real cf as above with formula Final.3.(A) and Final.3.(B). This will return the final value of the calibration curve for A and B.

    That's an excellent idea! I'll set up my labels as
    FL1
    FL2
    ...
    FLz <- and there's my A/B!

    Thank you so much for the helpful response!
Sign In or Register to comment.