On the topic, maybe a rigid material like we currently use (glass or aluminum) that could be vibrated or tapped to loosen the part. I’m thinking if the tapping is done in the center of the plate so that it flexes slightly, and that you do it a few thousand times might be enough stress to free the part.
Or maybe it would just break the glass.
For the record, the belt used by MBI’s ABP was PET, not PTFE. Also, the people who had the most success with the ABP replaced the PET belt with titanium foil covered with kapton tape.
Personally, I think that a swappable rigid surface would be best. These surfaces could be reusable and held in a hopper, so that when a print finishes, the platform is ejected and allowed to cool, while another is fed in like the page feed in a paper printer. The user would just have to come along periodically to remove the parts from the ejected platforms and load them back into the hopper. If you have a large enough hopper (and a reliable enough printer, and an appropriate place for the full trays to be dropped, etc.), you could start a print queue and leave for the weekend, letting the printer continue until you get back on monday.
Any plan that relies on a mechanical device without feedback mechanisms to remove prints of various and unknown shapes and sizes from a platform without changing the shape of the platform (like the way the belt peels itself away from the print when it goes over the rollers) is doomed to the type of failure that could potentially damage the machine.
I sometimes use thin carbon fiber covered in Kapton. Bend it and the print pops off. What about a track that curves to force an automated bend in the thin carbon fiber, hopefully popping of the part then retracting back to be flat??
Brook
Not sure about carbon fiber, but composite material is a good idea. I’ve gone back several times to the idea of a belt made of rigid slats so that a belt would flex in one direction (held straight by the tension of the belt), but could not bend in the opposing direction.
Those who remember trying to use the ABP remember that the warping of a print would pull the sides of the belt up, and the movement of the platform would cause the belt to twist under the inertia and nozzle-drag of tall prints (though this latter issue would be mitigated by not moving the platform in the direction parallel to the axis of the belt’s rollers).
Making these slats out of something more rigid than carbon fiber (though perhaps woven with carbon fiber, kevlar, or steel wires running perpendicular to it) and infusing the whole thing in a hard but flexible resin might produce the ideal belt for such a system.
Is there any memory materials that could be used so the plate is tight when hot and slackens when cool. Aiding removal.
@Nigel_Dickinson in addition to trying to go back to its previous shape when heated, memory wire shrinks axially when cooled. Different wires are rated by the force of this contraction.
Another thought I just had, it would be a lot easier to keep a belt from lifting or twisting if you weren’t trying to keep it flat. It would make the kinematics trickier than a straight-up cartesian system, but if you were printing on a curved surface like a cylinder, it would be much easier to keep the belt from deforming. If your “platform” was shaped like a large D-shaft and your printer could print on the part of the belt that is stretched over the round side, being bent in one direction would make it resist bending in the other direction, and advancing the belt would peel the part away when it moved over the flat side (probably on the bottom, over a bin).
I do huge amounts of automated printing - leaving printers running 24/7 or over a weekend in the E3D industrial unit. All ABS printed fan-ducts for the E3Dv5 are printed on an automated system where the machine prints 4 ducts at once, detaches them from the bed, and sweeps them off the build surface down into a box where they are collected. The printer then starts the next print.
I’ve posted about it on the forums here:
http://forums.reprap.org/read.php?1,216570
And you can find a whole slew of videos about it here:
http://www.youtube.com/user/Sanjay1337/videos
@Sanjay_Mortimer I knew I had seen somebody in the forums do this, just couldn’t remember who 
Nice, that’s pretty much how I imagined it. though having the sweeper on the head I didn’t think of, I kind of had a ‘oh… duh.’ moment when I saw it.
I print straight on clean glass and can remove the parts really easily once the glass is cool.
finding a way to quickly clean the glass after the print would be good.
Yeah, if you could add a zamboni type attachment to the head, that sprayed a solvent and scraped the crud off, without removing the kapton, and without making a complete mess… Crap, i’d like that for my kitchen sink, microwave, bathroom sink…
actually, i meant cool the glass rather than clean.
Just to be clear- I’m printing abs onto glass with hairspray. No kapton. Exactly the same surface for PLA. The surface remains unharmed even after many many prints. I actually find that it’s the manual removal process involving scrapers, dirty fingers etc that degrades the surface, not the printing.
Ah, nice. I used a glue stick once on kapton. Let’s just say the kapton was stuck to the part long after the part was removed from the bed.
This was PLA with no heat. I just left the kapton on the bed since it was already there, and I wanted to test out the glue stick
Bumping the prints off can be made to work reliably for things like stretchy bracelets and the e3d fan ducts that are flexible and have an attachment to the platform that is basically a thin line. When these types of objects are hit by the bumper, they flex and peel themselves away from the platform. The same might even work for things like single-wall spiral vase prints because the base is thin enough to flex, but it would not work nearly as reliably for larger, less flexible objects with a wide, rigid base.
It works perfectly with PLA with large and rigid objects. The hairspray I use sticks like the proverbial when hot, but when it hits 35~c the parts detach completely. They need zero force to detach, the head just sweeps the parts off the bed. It’s the same with ABS - the parts detach entirely at around 60c. However when printing ABS the bond to the bed isn’t strong enough for warp prone objects. I need to experiment with larger objects in abs and brims.