I spent several months trying to fix the issue with bubbles on the lines causing inaccuracies with the doser. I tried several different styles of check valves, different tubing with same I.D. but larger O.D. thicker tube walls might be better, I changed the location of the doser lowering it to reduce static pressure in the lines. I opened up all of the pump heads and re-positioned the silicon tubing to be squarely on the pump rollers. I even swapped the worst line with different pump heads. Swapping the pump heads caused the largest change in performance, but nothing I did stopped the bubbles from getting into the lines, and these are not the small bubbles that come from outgas of the liquid, I'm talking bubbles in the lines that are several cm in length.
My conclusion is the the pump heads have inconsistent clearances, or put another way poor quality. In the end I'm still using this pump, because I like it and we're not doing nuclear physics here. I focused on the amount of consumed liquid being dosed over long periods of time, and calculated a correction factor for each pump head. I use this correction factor to adjust the amount of fluid that the pump "thinks" its dosing and it reasonably corrects the output. This is a bit of a pain, and I suspect similar to what the "Calibrate" function does in the application, except it does it for more than a single dose cycle and over longer time, and attempts to correct for air in the lines that accumulates in between individual doses.
So
@Savannah16 I'd swap heads around till you find the best two and then if necessary increase/decrease the dosing amount to get the true amount you want on average over time.
