Questions to be answered for production refinement

Here I am going to put the list of things I would like to test and concerns I have about the existing process. I will edit this when time permits. It is a work in progress.

The goal is just basic research where we can conclusively discard approaches that have no value. I also want to make a production system that can replicate the production parameters with consistency.

Concerns

The time it takes to mix the 2 main ingredients via the peristaltic pump. The mixing from the second input reservoir is not instant and may take considerable time. Therefore the precipitation reaction will start with a very dilute solution that has more time to work as the rest is pumped in. This really muddies what is going on and makes it impossible to narrow it down to one variable. The ratio of the 2 inputs varies from 0:1 to 1:1 through-out the process final mixing process. The original solution has considerable more time to precipitate.
It would be useful to just dump the 2nd solution all at once into the mixing container but we have yet to find a simple and elegant way to do this. The peristaltic pumps are fairly slow as they are powered by stepper motors. So this makes the mixing process rather slow. We could try other pumps but the cost and complexity starts to rapidly go up. The RAMPs board has plenty of capacity for stepper motors but the amount of DC motors that can be powered is limited. (Atlthough I’m still working that out).
At some point we may also run up against a limit on what can be done by the gcode. At that point the project’s complexity will go up significantly so I’m doing everything to avoid that. For now the ingredients can be mixed instantaneously by manual intervention. Then the result can be tested and a determination can be made.

Parameters to test -

  • Specific pigment particle size - By mixing and letting the water partially settle - we are able to take samples of the various sizes of pigment particles. It seems safe to theorize that the higher particles in an agitated solution (where pigment is still settling) will have smaller the pigment particles at the top. So one could create a pigment with the smallest or largest of the pigment particles to see how it impacts performance.

  • Mixing it all together at once while heavily stirring. What is the result of just dumping the 2 solutions together at once?

  • Mixing it all together at once without even doing the timed mixing process.

We could try mixing both ingredients at once while it is being mixed. This might make as much sense as a convoluted mixing process.

Another thing to do is to try the various mixing timings. Although nighthawkinlight’s original blender recipe is dependent on the timing - this may not do much. There are 3 distinct timings in the original recipe - each aimed at creating a different size particle that fit together more tightly. If the tests in the previous post show that the elaborate mixing process is needed - then we will vary the parameters within that process to see if we can eek out a small gain in cooling power. I haven’t really thought out a test-plan as the other tests need to be done first.

Given the pumps are relatively small and do not pump fast, it means that the mixing of the 2 solutions happens over a period of time. This means that the initial particles start to precipitate immediately and presumably keep growing. So with each mixing process there will be likely be a wide variation in particle sizes. This is roughly what nighthawkinlight was trying to accomplish with his multi-staged blender recipe. The difference is he mixes the 2 solutions instantaneously while we have to rely on the slower peristaltic pumps.

ps - sorry about the back and forth with the accounts. I’m doing admin work setting this up and don’t pay attention to which account is logged in. - Brian B

I am not sure if I mentioned this elsewhere but I am skeptical about the 3 different and distinct mix phases that vary in the time the mixer is on. Our final product was very similar to the effectiveness of what nighthawkinlight but since the 2nd ingredient is not instantly dumped into the mix container - the pigment may have a varying size simply because the nature of the timed release/mixing aspect.

So this would be another thing to test. The speed of the pump. Our first iteration has a pump that goes as fast as it can but it might be better to not go that speed and make the stirrer stay on a constant amount or until the end.

So this assumption really should be tested before much else is done. It might be that we need to find a new way to mix in the second ingredient that is faster but it would complicate the design.

The first thing to test is creating the pigment at room temperature. Mixing in the ingredients is exothermic. NightHawkinLight’s original recipe has you lower the temperature before mixing. This takes considerable extra effort and equipment. When we’ve done it before, we have used the freezer bags meant to cool coolers. We have also looked at the peltier coolers often seen in PC cooling. Both of these will be problematic. If we can simply mix the ingredients and let it cool to room temperature, maybe the cooling aspect is not needed?

This would be a large simplification to the process. If not, we will likely recommend the icebags that we can reuse in the freezer. Although not idea - it is a simple way to do it that will work consistently. The other issue is that the mixing process takes 4 hours so the temperature is drift back to room temperature after the initial icing.

Here is another thing I am unsure about. The gcode generator can generate the gcode based on a variety of things.

Do we let someone specify the mixing counter size and then calculate backwards from that?

Do we specify the amounts in the input buckets?

It seems like it would be better to have the script figure out a semi-optimal amount of mix-cycles along with the corresponding pump times.

The way it current is, it doesn’t make the most sense and can easily lead to overshooting in amounts.

-edit - This is currently being redone to support 3 usecases depending on your goal.

I have made a contact with someone who has experience with this and has mentioned that the calcium carbonate can be had at a far cheaper price. One of the first thing I’m going to do is just throw the ingredients together and see how it impacts the efficiency, It would be a shame if all the mixer development I made is wasted - but it would be better for the long term if the paint doesn’t need these elaborate stages. It might also be that the difference in efficiency does not justify the extra involved process.