Hi,
As far as I observed is the batch to batch variation between CM5 chips low. We use research grade chips. For the best reproducibility use the certified sensor chips.
In general the sensorgram is higly dependent on the used ligand and analyte. So when you find totally different sensorgrams first check the stability of your samples and reagent before you doubt the sensor chip or machine.
The same holds for the lifetime of a immobilized sensor chip. Every cycle of analyte binding and regeneration will change the surface and affect the lifetime. A positive control will help you monitor the validity of the surface. In addition the baseline will give you some information about the state of the surface.
I'am not aware of special CM5 sensor chips.
Arnoud