...
Once the FCM was set up and the bottles were attached to the PVC and connected to each other with the appropriately calculated lengths of tube calculated by the FCM design tool, we measured the resulting flow rates manually to determine whether the calculations had been accuratetube lengths provided appropriate flow rates. (The intended flow rates were 100 mL/min for the clay and 5 mL/min for the alum when the tubes were both plugged into the middle hole on the inflow column.) We found that the clay flow rate was about 120 mL/min, and so we increased the tube length and measured continued to measure the actual flow rate by hand until we had a flow rate until it was as close as possible to 100 mL/min. The final tube length was 146 cm. The alum flow rate was a little bit slow with the design length of tubing was a little bit slow, so the tube was cut to 75 cm , where we measured the flow rate by hand to be 5 mL/min. for a flow rate of . The actual tube lengths for each of the 3 existing flow controllers are as follows:
FCM | Stock | Tube Length |
1 | clay | 149 cm |
| alum | 74.5 cm |
2 | clay | 150 cm |
| alum | 77 cm |
3 | clay | 124 cm |
| alum | 81 cm |
These results were interesting because they implied that in the case of the larger diameter tube, the head loss equation over-estimated the head loss in the tube, and in the case of the smaller diameter tube, the equation under-estimated the head loss in the tube. We also observed that it significantly impacted (slowed) the flow rates through the tubes to add bends by looping them around, which was done in order to avoid the inconvenience of very long tube lengths and save space. The added friction of the bends in the looped tubes caused the flow rates to slow, so that it was necessary for them to be adjusted accordingly in order to achieve the desired flow rates.
The modifications necessary after the construction of the flow controllers unfortunately make it difficult to have a concrete plan for their design in the future. We were not able to observe a distinct enough pattern in the inconsistencies to allow us to create a method for making standard adjustments to the design values without all of the manual flow measuring and adjustment that took place.