Gradient Curves on different controllers

[DeniseHeinz]

I am using an old HPLC method that states "use curve number 7 for Waters Gradient controller, Model 680." Then it says, “other instrumentation may require experimentation with gradient profiles.â€

[Uwe Neue]

Don't know about the Shimadzu, but the curve 7 is flatter (than linear) at the beginning and steeper at the end.

[Kostas Petritis]

I think they are referring to the eleven gradient curves (i.e. linear, 2 step, 4 concave, 4 convex). You can find a figure with all the different gradient curves in the link below in page 5.

http://classes.uleth.ca/200601/chem3410 ... graphy.pdf

This will give you a pretty good idea of what the curve 7 looks like...

[DeniseHeinz]

Thanks for the info. It doesn't completely help me here. My Shimadzu manual states the following equation for gradient curve profiles:
f(t/T): When a>0: ((e^(at/T))-1)/((e^a)-1)
When a=0: t/T [the linear example]
When a<0: (equation for the convex curves... n/a)
Where "a" is the curve value
T is the end time (??)
t is the time at a point in the run (??)


So, how does that compare to a "7"?

[skunked_once]

Denise,

Why not overlay a copy of the gradient image from the URL cited by Kostas on some graph paper and estimate a %B/time gradient at regular intervals? This may produce a close-enough approximation to test out on the Shimadzu instrument. Good luck!

[Uwe Neue]

Curve 7 is probably the curve where a = 1 (6 is linear).

[d.epstein]

Can anyone comment on the value of using non-linear gradient curves? Does it actually help? When should we use non-linear gradients?

[tom jupille]

Can anyone comment on the value of using non-linear gradient curves? Does it actually help? When should we use non-linear gradients?

When to use curved gradients? The short answer is "never". As this old thread demonstrates, curved gradients are virtually impossible to transfer drom one model instrument to another without significant rework. Anything you can do with a curve you can do with multiple linear segments.

Even multi-segment gradients have limited utility outside of a few special cases (e.g. initial isocratic holds, or steeper segments at the end to speed up late-eluters) because mixingand washout differences between instruments complicate transfer.