To clarify, water both boils and freezes in vacuum. While most other similar liquids do the same, some other fluids (including polymers) have different reactions.
Printable View
Water boils, this boiling causes the temperature to drop, that dropping can cause the water to freeze. The freezing is not caused by the vacuum and whether it freezes or not depends on the temperature it started out at and how much energy it can absorb while in the vacuum.
I would think you would need something with a great deal of structural rigidity to survive the vibration associated with takeoff. I think a scaled-down delta would be too fragile. Given the amount of time you have available to actually print at zero G, maximising build size and height is pretty much irrelevant, as you won't have time to print anything of any size.
I would look into building a scaled-down reprap, using RAMPS controlled by Octopi running on an Arduino. I would use 20mm aluminium extrusion to build the frame as it should give you reasonable structural rigidity. I'd avoid using linear rails as they may shake themselves loose, but go with bearings and smooth rods for the axes. some of the plastic parts would need to be adapted to the extrusion but this is relatively straightforward, plus the use of extrusion cuts down on the plastic part count as points of failure.
To cut down on the power requirements, I would choose a material that had a relatively low melting point and did not require a heated bed. PLA would seem to be the obvious choice, but I'd go further and create some custom filament out of polymorph. It has a lower melting point and has successfully been used in 3D printing before (to create a prototype conductive filament combined with lamp black, from memory). Polymorph should be relatively available in pellet form; you could reasonably use a filastruder to create your own filament, or ask someone on here to do it for you.
The steppers will obviously have to be scaled down, plus I'd go for a bowden extruder instead of one mounted on the carriage, as there's less mass to move and therefore less of a power drain. The huxley uses nema 14 steppers, which work quite well.
If you have 1 amp hour available, you should reasonably get a few minutes of printing with the above setup.
Edit: using Octopi is daft, as I doubt you'll have a wifi connection to the rocket at apogee. I would consider customising the Marlin firmware to start printing from SD given a specific signal, which should be simple enough to do with Marlin running on a RAMPS board.
Edit again: you'd need to start heating the hotend before you hit zero g to maximise printing time.
Make the cylinder like this and use a slotted linear bearing for the sliders with small roller bearings to stop it rotating. It would stand anything you threw at it in terms of G force and vibration.
http://3dprintboard.com/attachment.p...4&d=1415056876
Is that printing window the ONLY time the printer can be active? I have the mechanics of Kerbal Space Program stuck in my head where the main engines create an enormous amount of surplus electricity. As 3dkarma suggested, preheating any elements that require it would be a major bonus to printing time, seeing as you only have 3 minutes. If you can use power sources during launch to do this without using up any of your power budget, it would be very useful.
And while the pressure inside the vessel will be about 1 atm, what about the temp? Will be able to be maintained at an optimum temp, or it will it fluctuate as the flight progresses?
I do wonder if you could apply a large static charge to the printer bed to make the extruded material stick. The charge should propagate up the layers as the material is laid down.
I like that idea, Mjolinor, my concern, though, would be an accidental grounding to something critical, like the printer control board, or even the parachute deployment system. ZAP! wiiiizzzzzzzz SPLOOSH!
But it would stop those pesky space aliens from sticking their fingers where they shouldn't.
At the very least it would keep us from removing the part before the bed cooled enough.
I think i'll just stick to trying to hover one in orbit on a hexacopter and print, it all sounds too complicated for me :p