Obviously a lengthy read, only really of interest to those thinking about using RP or wanting an alternative to sleeping pills.
I had no personal experience in this field when I decided to use RP for a project, other than what I had read from a few simple Google searches. I found a couple of local service providers, not wanting to go international postal, I decided that these would have to do.
Their print area was not spectacular, so I chose the largest of the two which was 14cm x 14cm x 20cm.
I had questions that were bothering me; Firstly I had seen 3D prints on the web of assembled parts printed out as an assembly, also the traditional ball in a ball thing. This made no sense to me, as the parts would have to start printing in mid air. Also, if I gave a positive fish model, how could it possibly print, as the contact with the ground plate is minimal.
I submitted a positive fish model as a test, to see what came back. Unfortunately they took so long that I jumped the gun and submitted my mold model, which was a mistake on my part. Still, plenty was learned from the two models.
The bluegill came back looking a right mess. The model that I provided was split into two halves and glued together after printing. The join line was along the lateral line. There was a quite heavy seam and a mess of glue. Had I known that this was the method of printing, then I would
have printed two halves and carefully glued them together with more precision myself and made a far better job. The only scruffy parts of the print were the rear of the dorsal fin, which is an overhang and the rear of the ventral fin for the same reason. To understand these issues, deeper research was required.
I was right, the process cannot print on fresh air, so what the print software does, is print from the base, a rough support structure, to support the model for printing. This structure is then manually trimmed away and you are at the mercy of the guy doing the trimming. It was not too bad, but I could have done a better job myself.
When the mold parts came back, it was plainly obvious that holes in anything but the vertical were not possible with any accuracy. The 20mm diameter piston hole came back as an elipse with a 1mm error. I had to significantly trim the piston down to get a free fit, but the 1mm gap on the long axis of the elipse meant that it leaked like a sieve and there was not enough pressure to fill the mold.
I designed the mold with spherical locators and depressions, I also included undersize holes for bolting. This was a good decision as the locators did not fit, being too tight. I drilled out the holes and then removed the locators completely.
The next problem was that the mating faces were far from flat, with a significant warp, despite the fact that I included a lattice work behind the face, to add stiffness and reduce the likelyhood of warping. The plastic was however soft enough to pull together, but a close seal around the cavities was not possible with the edge bolting and leaky gaps occurred and combined with the inneficient piston, prevented the mold from filling.
The surface finish was far from smooth, but this was no surprise given the layering method for depositing the plastic. This I was prepared for and willing to accept.
I have just coated all the mating surfaces of one side with a contact adhehive, in an attempt to lay down a rubber gasket. I am hoping that this might allow me to get a few complete pours for water testing.
When printing the mold halves, the guy again split the model to enable printing. Unfortunately right through the mold cavity, making the mold
useless for anything other than a rough prototype.
Other problems occurred, but these were with my model, with the fins being too thin.
Rapid prototype is not cheap, the bluegill positive and two mold halves cost me just short of $100. This is not too bad considering that carving a master to this standard would take me days and ultimately not achievable for someone with my carving skills. Printing a positive master is a good plan, as the rough surface finish can be fixed with a thin coat of epoxy and a turning wheel.
Conclusions - first impressions would be that this was a total waste of time and effort. But time is never wasted when exploring new techniques.
In fact, once you understand the limitations of the RP process, it is possible to design something that will work.
My plan now is to fix the computer model and make an RP mold master for pouring plaster mold halves. The mold halves will be symmetrical, so only one master needs to be printed. Once the plaster molds are poured, dried and drilled, they will be skimmed flat on a large sheet of emery paper. This technique will allow me to produce as many molds as I require for production, plus replacements in the future.
The built in piston idea cannot work with RP, so I am going back to vacu-venting as I believe that I can make this work. I also think the vacuum method will allow me to achieve the fine detail and a four cavity mold.
So, more work time and money to be put at risk, but who's counting. This is what I do - experiment.
The image shows the mold halves and the piston. The right hand piston was printed in the vertical and shows how the software built up structure to support the model. The piston was eventually split and printed horizontal.
Also included a pic of the bluegill positive, showing the split and messy glue.
TU RP mold 3281.jpg 157.57KB 5 downloads
TU bluegill positive 3282.jpg 101.54KB 5 downloads
Edited by Vodkaman, 14 August 2015 - 03:27 AM.