Mirroring my post from here. I haven't found any development on an injection-moulding machine on the OSE wiki so far.

I've been wanting to build or get hold of an injection moulding machine since I've found that a couple of my major design projects (modular rotary hydroponics, and 'wind-lens' turbine prototyping platform) could be most cheaply produced by moulding waste plastic into cast or milled aluminium forms, where I have put many replaceable repeating segments in those designs.

Having seen this kickstarter for a bench-top injection moulding machine, and seeing that the same group apparently offered a $500-600 version that consisted merely of a cylinder that you mount in a drill press along with a power supply. Unfortunately, neither of these systems is Open Hardware, so I don't know exactly what's going on with the internals.

Both their first and second versions seem to be nothing more complex in operation than concentric hopper-piston-cylinder-nozzle, as seen working here and here.
However, I learned little about the operation of these machines at uni, where they were treated pretty much as a black box with an auger in it that conveniently produced high temp&pressure plastic goo for a mould that we would be expected to design, but never expecting us to need the finer points of actually designing the machine itself.

I understand from looking at Fig.1 of this patent for example that some commercial systems may incorporate a sliding valve at the nozzle, to prevent molten plastic from spilling out before it is needed to do so (what we in the reprap project use 'retraction' for on filament extrusion systems), and LNS' systems don't even seem to incorporate this, however I'm not completely sure what goes on at the other end of a piston system, specifically what happens to the air gaps between plastic pellets/nurdles poured into the hopper. Do the pellets all melt down into a pool before being injected, or do some remain in the cylinder floating on top as the piston shoves them down? Is a large air gap between the piston and the top of the melting plastic acceptable, and do air bubbles reliably rise out of the plastic so that they are not injected in such a system?

I can build or get access to a lever/press quite trivially, but the piston/cylinder is where I'm stuck here. I want to mould some parts a with volume approaching or even exceeding 100ml, but high precision is not crucial here, so I don't mind the system using relatively low pressure and parts suffering from minor shrinkage.
Would it be reasonable to get a 2-inch steel pipe, weld a plate onto one end with a threaded-fitting brass/alu nozzle, and use a 1cm thick cast or turned alu disc at the other end as a piston? Would a rubber (silicone?) o-ring be needed on the 'piston' (for a better seal)? Let me know if I need to sketch that for clarity.

Edit: someone on RepRap IRC named Jonas found this bit of copyleft documentation on someone's home IM build. Instead of any kind of pipe, they drilled a hole through one large block of aluminium.