Creating New Instrument Method

<p><span>Hello People,</span></p><p></p><p><span>I have couple of questoin regarding creating new Instrument Method.</span></p><p></p><p><span>When creating the gradient in Binary solvent manager section, what does <strong>curve value 6</strong> refers to?</span></p><p></p><p><span>What is the function of <strong>Filter time constant</strong> under TUV detector section?</span></p><p></p><p><span>Can somebody reply to this question ASAP please?</span></p><p></p><p><span>It will be appreciated.</span></p><p></p><p><span>Many thanks</span></p>


  • lizh


    Waters's gradient curves have been in use for many versions of pumps "6" the default is a linear and you can have convex and concurved too. We have found that they can be used to tease apart two peaks in a ballistic gradient, but 6 is used. I and 11 are also useful "1" is go to the conditions immediately at the beginning of the time segment and "11" at the end, and is often used for shutdown conditions. The online help and tool tip should give you a graphical representation which illusteates the curves best.

    The filter time constant is important. It is used on the detectors - in other vendors the required data rate selects the filter time constant for you, but while that might be convenient it means that you are given a high filter time constant at high data rates. That may not be appropriate. For example noise on a UV detector is generally a combination of the factors. However, high background absorbance produces a higher noise and so if you have an absorbing mobile phase and are monitoring at low UV wavelengths, such as for TFA then the noise (baseline) will be high. THe same separation without TFA the baseline noise is lower. Yet if you chromatography is fast or at UPLC rates, when the filter and data rate is combined the peaks will require a higher data rate, and will be given the same filtering value. Yet the non-TFA method do es not require the same level of filtering. Too much filtering distorts the peak shape, usually decreasing peak height which adversely affects S/N and thus LOD's. This is why for Waters have the two values decoupled, but have meaningful defaults.

    Try injecting a representative sample and pick a data rate that will give you 20 points across your smallest peak of interest. Try the default filter and one lower and use the data rate that give the best S/N.