New calibration and consumables, percentage problem

[tleek]

Hello all. So I was having some recent problems that I posted about. We got a tech in. New consumables, longer column, good method, good curve. Everything was changed. My standard has 3 components. In our actual samples, one analyte is usually pretty high, and the other two are almost always less than 1% by weight of the sample. However, the lower analytes usually have a broad range. I am consistently getting 0.24 to 0.31 percent by weight of one low analyte, and 0.14 to 0.19 percent by weight of the other low analyte. Any ideas for what could be causing this? I added a low calibrator of 0.025 mg/ml, and i have good linearity and correlation, but the amount given after analysis of this calibrator is off by 20% of what it should be. I need to tweak this. Perhaps better pipette skills. Could this be a cause? Thanks

[tleek]

Silly question. In chemstation how do I go about deleting a level that I added to the calibration table?

[Steve Reimer]

To delete a level in version D, E, or F: Select menu: Initial Calibration/Update Levels/Delete level

As far as your calibration question goes, without knowing how wide your calibration range is it isn't possible to say whether a 20% error on the lowest standard is unreasonable. For some of my applications 30% or even 50% is acceptable.

[Peter Apps]

. However, the lower analytes usually have a broad range. I am consistently getting 0.24 to 0.31 percent by weight of one low analyte, and 0.14 to 0.19 percent by weight of the other low analyte. Any ideas for what could be causing this? I added a low calibrator of 0.025 mg/ml, and i have good linearity and correlation, but the amount given after analysis of this calibrator is off by 20% of what it should be. I need to tweak this. Perhaps better pipette skills. Could this be a cause? Thanks

Whenever results from samples give unexpected results the most likely explanation, and the one that has to be eliminated before you consider anything else, is that the results are correct and the samples have an unexpected composition.

If your lowest analyte percent is 0.14% why are you calibrating five orders of magnitude below that ?

Peter

[tleek]

The recent results are as low as 0.14% by weight yes. But from previous samples when we were up and running perfectly a couple months ago, we can get quantities that are much lower. The previous tech that was in this lab had methods that had the lowest calibrator at 0.025 mg/ml, so I wanted to do the same. Just having trouble.