Problem with the chip and/or device resulting in very poor sensorgrams

Hi,
thanks a lot for your quick reply, but I think I've described the problem not quite well. So, we've performed 4 SPR experiments on different days and after the third/fourth experiment, it goes wrong. We always measured up to 12 samples (five concentrations each) in two replicates. But these replicates are "technical replicates", so the instrument is injecting the sample twice from the same vial. Maybe something is happening in the vial during storage, since each experiment is always running for about 20 h. But still, the problem has only arisen for a few months. Before that, it always worked.

We used regeneration conditions, because the Fc gamma receptor is captured by an anti-His antibody. So, after each binding cyle, the receptor needs to be removed.

If this is the case, how can we check non-specific binding? We also observed that the problem often starts on flow cell 2 and then followed by the other active cells.

Non-specific binidng is best recognised from the raw data traces.
Overlay all channels and look to the response of the curves.
When the reference is higher than the active, something goes wrong and subtracting the reference wil give lower/negative curves.
Arnoud