Currently I am gathering data to make viscosity-temperature curves on a Brookfield LVT. I keep running into a problem of inconsistent viscosity readings from my LVT. I generally work from 70-100 degF and take data points within this range. The problem is that as I increase the temperature there is not always a distinct trend in the data i.e. viscosity is increases with temperature. Sometimes I will increase the temperature by 5 degF and I will get a value higher in viscosity. This shouldn't be the case as liquids have an inverse viscosity-temperature relationship. Attached are some example graphs of the data I collected. The LVT gets calibrated weekly with Brookfield calibration fluid. The products are highly non-Newtonian (shear thinning, thixotropic). Could this be a problem with shear history of the fluid or the complex non-Newtonian nature? I have thought of solvent evaporation as an explanation but I am not convinced this is the case, yet. As I have faith in the reliability of the viscometers/thermometer calibration.
I multiple Brookfield RVTs and LVTs available as well as a R/S-CPS+ rheometer.