By MikeS on Friday, July 9, 2004 - 01:08 pm:

I am hoping to get some opinions on this issue: When a Waters Alliance 2695 detects a bubble on compression, it goes into a recovery sequence whereby the piston stroke volume is increased and the stroke rate is reduced to maintain constant flow, to assist in bubble removal. As per the Waters manual: "During this process, pressure ripple as high as 40% may be observed." and "... shifting of retention times may be noticed due to a difference in mobile phase mixing...". It does not seem possible in a GMP environment to accept data that was generated for individual injections or for runs in which some injections have had this recovery sequence occur. We have seen severe retention time changes that we think are due to this issue. What do other users do? If you disable this feature, are you running afoul of other GMP issues ("Ahh, disabling error message functions, are you?") I do not see how data can be used when this alteration to your instrument conditions is being performed, especially when the manufacturer says to expect retention time shifts and increased ripple (our company's ripple spec is <2%). Hopefully I am wrong about this - Any ideas? Thanks

By A on Friday, July 9, 2004 - 02:36 pm:

Interesting Post....


1-A pressure ripple spec of 2% is pretty tight, and how (when) do you measure it (over the whole run or just a select portion). Does a brief spike (like during bubble recovery) averaged over the entire run put you out of spec? I have never used pressure ripple as a spec for a quality of a run, I am more interested in RT.
2-If you are constantly getting bubble errors, maybe something else is wrong (broken degasser, misbehaving check valve).
3- Severe retention time errors are seldom caused by this bubble correction (unless something else is broken).
4 - To enable or disable. If a pump chokes on a bubble, pressure ripple goes up, retention times shift regardless if this feature is enabled or not. With it on, the ripple and shifts are much less (flow remains constant, that's the whole point). Overall retention time performance is better with this feature on.

By MikeS on Wednesday, July 14, 2004 - 06:33 am:

To A:
1) Ripple is read/calculated from instrument display only as a post-maintenance check. Suitability of this check is not the issue here.
2) I do think this is check valve related, but regardless of the cause, my "GMP-ness" question remains
3) Retention times (minutes) for 3 consecutive runs for deg 1, active, and deg 2 respectively: 9.192,15.533,20.717;; 9.108,15.575,19.658;; 9.083,15.492,20.700 - The second injection experienced a bubble recovery prior to the "19.658" peak, and its retention time is way off from the others'.
4) I doubt my "19.658" peak would have come out even further from the others if the recovery hadn't happened - Flow would have to have dropped to 0 for more than a minute to cause this shift, which a brief "choke on a bubble" without subsequent recovery probably won't do.
To all 2695 users: Even if you ignore the fact that the system changes retention times when the bubble recovery happens, I'd still like to know if/how your GMP facilities deal with this issue. Do we simply not care when systems lose a minute of retention within the space of 5 minutes of run time due to this bubble recovery procedure?