Hi Robin, I reached out to a contact within BST (Florian) and he arranged some engineers (Mary and Muhammad) to take a look into this anomaly with me. Based on their observations they believe this saturation may be caused by a number of reasons such as direct contamination, topics like sensor aging, heater profile settings, individual sensor history, specimen composition and concentration. However, these sensors/devkits were not that old (maybe 3-4 months old), and I was using the default profile. But they assured me this was something they have seen before and it may also just be the calibration of the sensor and it's sensitivity to something in the background air composition. I performed a few additional tests after that meeting to determine if this may have been saturation due to something in the background air composition. So I exposed the sensors to some smoke from incense, as well as coffee, and I still saw these saturated values. At that point I decided to replace the devkits in my setup with new ones as well as a new single sensor. It appears the saturated values still do appear occassionally, but not as much as with what I was seeing in the data I provided above. I suspect Muhammad may be right that with longer burn-in of the sensor it may be that it is more sensitive to something in the air composition of the environment I am testing in. Although, I also suspect that the substances I am testing with may also be causing this. I'll continue to observe this behavior and if it gets really bad as before I will follow-up here on the forum in case it might help others. For the time being we are allowing these saturated values to coexist in our datasets, but we will likely be pruning the data in areas where it is constantly appearing.
... View more