Perplexing gradient shift after maintenance

[Arne]

All,

I have encountered a really perplexing problem of my HPLC pump: I cleaned the proportioning valves and liquid displacement chambers of our Thermo Accela pump and swapped the proportioning valves around (A is now C and B is now D). The system is leak tight to 750 bar and delivers close to the requested flowrate.

However, when I run a gradient test compounds elute much earlier and the pressure trace is shifted to earlier times. It appears as though the pump is moving through the gradient much too quickly, i.e., it gets to 95% around 7 min instead of around 9 min.

Has anyone seen such behavior and do you have any suggestions as to how to address this problem?

I am planning on switching the proportioning valves back around tomorrow to their original configuration.

The black pressure trace is the new one and the red is the old one.
See this picture: http://i.imgur.com/XvhYhah.png

[bunnahabhain]

Can you add to the overlay the theoretical gradient? This would be helpful in detecting wether the old or the new pressure profile matches better.

[kubowicz.tomasz]

Hello

I'm not sure if your pump has this settings but if you swapped gradient valve channels you need to change parameters for pump (check if you have parameter called "primary channel" - it needs to be set properly).

Regards

Tomasz Kubowicz

[Arne]

Thank you for your help! Here is the theoretical gradient overlaid as %B.

Tomasz: There do not appear to be such parameters for the pump. I did not change the flow path just swapped proportioning valves. These valves are supposed to all be the same.



Bigger file: http://i.imgur.com/MPh4JGs.png

Thank you again!

[lmh]

(1) I believe that these pumps were designed with an assumption that the organic solvent would be on a line with a letter greater than the aqueous solvent (so if you're using A&B, B would be the organic; if you then change to use C&B, C should be the organic). Check this with a Thermo engineer though...
(2) Do you know that the pump's current behaviour is actually the incorrect behaviour? If, previously, pressures were low and peaks were late, is it possible it wasn't pumping the full flow rate later in the gradient before you cleaned it, and it is now delivering what it ought to?
(3) For a definitive idea of gradient test, I would suggest doing the 0.1% acetone in methanol versus methanol step-gradient method, making sure that you flush lines very thoroughly before starting. Good luck!

[Arne]

lmh,

thank you for your thoughtful reply.

(1) We always have the more strongly eluting (organic) solvent be the higher letter.

(2) yes, the current behavior is definitely incorrect.

(3) I just did something similar to what you suggested. I used MeOH and MeOH spiked with a UV active compound and tested delivery from each solvent line and ran step gradients. The result was that proportioning from "C" was incorrect. Cleaning the proportioning valve and flushing with MeOH solved the problem. The pump is operating normally.