I now have three of my peristaltic pumps out in operation — two are pumping cyanoacrylate in a factory in New Hampshire, the other is pumping beer wort in my office as part of a continuous brewing project my coworker has set up.
To pump the cyanoacrylate I eventually had to swap over to polyurethane tubing. The silicon tubing (which is generally what you get if you order peristaltic tubing) was getting degraded by the cyanoacrylate, which was kind of crazy. If I left glue in the tubing in the pump overnight, when I returned in the morning the outside of the tubing would be glued to the roller bearing. The only way I can explain this is that some glue particles were diffusing through the silicon tubing. WHAT
The polyurethane tubing is much stiffer, which required me to use a larger motor. This was a pretty simple design change in terms of changing the mounting holes and increasing the bearing radius, however it introduced more problems on the occlusion front. There are two good ways of solving this problem and one bad one — for the pumps in New Hampshire I went with the bad solution, but I will be making up to four more of these for other projects and hope to implement one or both of the good solutions for those.
The bad solution: Allow the top acrylic piece to flex at the inflection point when both rollers are in contact with the tubing at the same time. Given the design the side parts of the top acrylic piece can flex quite a bit if they are not fixed down. My original design fixed them with screws, removing these screws allowed them to flex. This didn’t really deal with the inaccuracies lasercutting this piece introduced — I still had to experiment quite a bit to find the best dimension. Too much flex and I would lose the seal, too little and the motor would stall.
One good solution: Machine the base instead of making a lasercut stack up. This should deal with a lot of the dimensional inaccuracies I was getting. With a higher tolerance piece the occlusion should be constant at all times and the inflection point should not require that much more torque from the motor. We recently got a mill at work, which will make this a feasible option.
Another good solution: Use three roller bearings. This removes the inflection point altogether. The reason I did not do this is because I did not have enough digital pins on my microcontroller to sense the position of the roller bearings. Every roller bearing introduces another point in a rotation when no glue is pumped, so I didn’t want to introduce more of these. In future designs I will definitely make space in my electronics for one sensor per pump. A hall effect sensor should work well — mount the sensor on the pump base and put magnets on each arm above the roller bearing.
I did some other, small redesigns to the pump which make it easier to mount the motor. (The base is also designed as one piece for milling.) I’ve included here some screenshots. Eventually I will add in the sensor and magnet placement, and change it to three arms, and then I’ll post it on Thingiverse.