USP Signal to Noise Calculations Using Empower

[Chrom Inquisitor]

In my lab we have run into an issue with the proper calculation of signal to noise using Empower. Unfortunately were are using Empower I as Empower II has not been validated in our lab as of yet. Empower I seems to have limited noise calculation options and I would like to inquire about the "correctness" of the calcuation as compared to the most recent change to the USP (S/N = 2h/hn).

We have traditionally used baseline noise as the means of determining noise but as this is a gradient which produces a sloping baseline the baseline noise function seems to be overestimating the noise. We would also use a custom field to complete the signal to noise calculation and also convert the units - (2*height)/baseline noise/1000

Is there a better way to conduct this calculation within the constraints of Empower I?

[krickos]

Hi

Regarding your calculation, I do not get the "1000" factor you have added, then formula below is straight from USP and harmonised with Ph Eur it seems.

The unit of H and h (or hn in USP) is the same, like picoAmpere or milliVolt...So the expresion is unitless

Appearantly, the measurement can be tricky in gradient methods, but have done your initial trials on a solution corresponding to your LOQ or monograph requirement?? A too concentrated test solution may be the source of your problem as width at half peak height increases and leads to longer measurement range for hn.

Ph Eur used to have a factor 20, now it is five like in USP for measuring distance of hn, so if it is a diluted narrow peak you should be fine, but then again there are always a risk with pharmacopiea methods when they implement stuff retrospectively.

Last, it seems like the S/N calculation addition in USP is for information, not a general requirement (was under discussion earlier though as Ph Eur have) such as precision/RSD, so if you have a S/N requirement in your monograph, how did you do before?

The signal-to-noise ratio (S/N) is a useful system suitability parameter. The S/N is calculated as follows:
S/N = 2h/hn
in which h is the height of the peak corresponding to the component concerned; and hn is the difference between the largest and smallest noise values observed over a distance equal to at least five times the width at the half-height of the peak and, if possible, situated equally around the peak of interest