Wanting to do Mendal for my Dissertation but can't think of any new ideas, any suggestions?

I am a 4th year Electronic student at Caledonian University. I was introduced to the Mendel by my cousin and think the 3D printer a great Idea for a project but can't think of any new improvements that haven’t already been done i.e. changing the bed to glass, using a glass nozzle, making a heated box for the Mendel to allow for faster printing, creating a bigger bed.

I have had one idea but not really sure if it will be feasible. 3D camera's are now coming out which would be a great addition to printer but I don’t think this idea would work.

Any suggestion would be very usefully, even if you know anyone who needs help with a current improvement to the Mendel please get in touch.

Something I've been wanting to do for a while is getting a height probe to measure the surface profile of the bed. Getting this extruder height just right is critical. Some people have made a probe for the makerbot bed (http://www.thingiverse.com/thing:4093), to set the bed height. Given that our bed is larger, and therefore more likely to have high and low spots. If we could get a surface profile, we move all three axes simultaneously to always ensure it’s just right.

Rhys: Wouldn't the objects come out slightly malformed this way? Imho, using a raft circumvents the problem in a better way, in the case where you can't make a level bed.

RepRap_GCALman: From the top of my head: I'd love to control my RepRap(s) via LAN/ethernet. And more ways to detect a hardware or software malfunction and stop the printer automatically would also be a great addition.

---

--
-Nudel
Blog with RepRap Comic

Yeah I like Nudel's idea of ways to stop the machine if there's a fault. Some users leave their machine unattended and go to work, but this kinda scares me, as i don't want to burn down my apartment.

I also would like visual feedback like you suggested, but with 2-4 webcams instead of expensive 3d cameras (at least expensive for now). Some projects i've seen require cutting the 3d model in half, printing, and then gluing the two halfs together. A ball or a skull are two examples i've seen. it would be great if you could just print the first half, flip it over and prop it up so the bottom is level, and then resume printing the second half.

On the 3D camera stuff. I assume you are aware of DAVID 3D scanner see [www.david-laserscanner.com]
Which uses a cheap laser and a web cam to produce a 3D object.

Always an improvement would be a step towards replicatable electronics. Like embedding wires in a part to get an electronic circuit. Or like printing conducting material.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Sure, I'd be happy to give you advice if you'd like to continue the Spoolhead project.

GCALman,

I have been thinking about 3D cameras both to scan a new object in, and to verify that the current piece is coming along as designed. A simpler, cheaper solution that might do well for your Dissertation in Electronics would be structured light approach to partial 3D. Have one or more LED lasers scanning the beam slanted at an angle attached to one side of the extruder, and a ccd camera on the other side observing the beam. Now Y position in the ccd will convert to item height or Z for the object, while X remains the distance perpendicular to motion, and multiple slices added together make the objects build the Y axis extent of the object. Cheap scanning lasers are available from Harbor Freight to use as cutting guides for power tools. Cheap 'board' ccd cameras can purchased from supercircuits.com. An optical filter to select only light the color of the laser will help eliminate stray light and make the laser signal clearer. You will likely need a more powerful computer to process the ccd images, so this would be a great time to experiment with the 32bit microcontrollers, or a very cheap single board PC.

You do not need a full Mendel to test the electronics and software to do this, just a fixed track above an object where the camera and scanning laser beam can slide slowly over the object. You will need stepper motors or an optical scale to measure distance moved. Experiment with the best angle for the laser and the camera to get the fastest, most accurate reconstruction. Tweak the software to the Z axis perfectly vertical instead of slanted like the laser. The initial data format may be a simple two-dimensional array of z-axis heights.

Adding an interface between the 3D image building CPU and the Mendel main control microcontroller and/or an ethernet interface for network monitoring of the build process would also be good. I think this should easily complete a Dissertation.

Mike

How long have you got? In my my part of the world a dissertation is normally 1/4 workload for a year. If so, you don't want to tackle anything too ambitious but you could help me with what I probably won't have time (or know-how) to do myself in my PhD.

I am currently working on 3d scanning from ordinary handheld digital camera photos with the intention that the photos eventually come from a computer controlled toolhead camera. If you could figure out a toolhead and controller that could mount a tool (webcam etc.) and not just move x,y,z directions but also change the tool orientation (pitch/yaw/roll) it would be beneficial for me at least.

i.e. 6 degrees of freedom rather than the standard 3 of the current mendel.

As for the structured light scanning idea above, its good and has been successfully done for a few years but I'm trying to cut out as much of the hardware as possible (plus lasers aren't good for the eyes) and solve the problems in software instead. So the only additional hardware would be the toolhead.

Although, looking back at your post, the toolhead itself may be more in the Mechanical rather than Electrical engineering realm :-(

We need an automated build platform for the mendel. Makerbot's blog has some stuff, as does the wiki. It's essentially a conveyor belt that we print on, with a heated bed underneath. Heated bed stops the warping (a big-ticket item on something with big print area, i.e the repraps) and when we're done printing, the conveyor just cycles until the part rolls right off the end. It lets us make thing after thing after...well, you get the point

Nothing to stop you printing one of the printable versions of the Makerbot design on Thingiverse and attaching it to a Mendel.

---

[www.hydraraptor.blogspot.com]

I was thinking last night of trying to hook the mendel up with a data acqusition system. all the user to control of the mendal from a distance. but will have to speak to the lecturer, i'll post a wikki page of what I do thanks for all you suggestion they have been create, I should know by the end of the week what I will be doing

Edited 1 time(s). Last edit at 10/05/2010 07:23AM by RepRap_GCALman.

Long time listener, first time caller...

I work at a consumer products company that produces a lot of plastic products, so here are some features that I think would be cool that I think would allow the RepRaps to standout vs. industrial injection molding and bring the concept closer to being able to print a whole machine.

1-Multi-material or multi-color printing. I understand that some of the newer generation electronics allow for a 2nd print head, but I haven't seen it executed. I think proving the concept would be cool.
2-Insertion or integration of a pre-manufactured part into a RepRep print, ie, encasing something in plastic. This could be done by pausing extruder, having it move out of the way, and then restarting where it left off. Additionally, if the part could be inserted by the RepRap, that would be even better (additional X,Y,Z motion control). This would be great for framing, working with bearings, adding support or strength (like rebar in concrete), etc.

I'm sure I could come up some some other ideas if you like these.

1: [ifeelbeta.de]
2: [www.thingiverse.com] (one example) Automation would be great though.

:-)

---

--
-Nudel
Blog with RepRap Comic

OK, how about 3 or more print heads then. Rainbow parts would be cool.

Here are a couple of ideas for your consideration:

1. (Analyze the key issues for) and design a granule-based extruder.

1.a Devise a practical method of measuring the flow rate of melted plastic out of an extruder's tip
IMHO, this is a key thing needed to make a granule extruder work well for printing. For void-free rod stock, one can reasonably estimate the output flow, based on the speed at which the rod enters. However, for granules (with voids between), it is unclear how one could measure -- and then control-- the volumetric flow of plastic.

2. (Analyze the key issues for) and design a system to convert (recycle) PET containers (e.g. 2-liter soda bottles) into granules or 3 mm rod stock.

2.a Design an extruder capable of extruding PET rod or granules (needs high temperatures.)

====

Please take these with a grain of salt, and note that others have tried to tackle 1 and 2, but with mixed success.
(So don't take on something you cannot get at least partial results for in the time you have available.)

---

Larry Pfeffer,

My blog about building repstrap Cerberus:
[repstrap-cerberus.blogspot.com]

Some more ideas if you have not already decided.

1: Figure out adding a microcontroller and memory to the reprap so it can print without been constantly attached to a computer.

2: Design a multi filament extruder to extrude seamlessly different filaments. (Different colours or types)

3: Design a extruder to extrude wax for model printing(cover the wax in silicone then when set melt wax out)

4: Have a go at improving the resolution of the reprap.

5: Printing with metal? (Here's a hint, look at an ark welder. A single small electrically generated point of intense head just enough to melt the tip of a coil of wire)

Any of these any good for you?

You could also research the various 3D graphic file formats to find something better than STL that can handle multiple material types, or colors, or handle adding embedded wires for making conductive circuits, or specifically angled fiberglass threads to reinforce the plastic. There are many existing file formats, some open, some proprietary but well know, some closed. If you examined many of them and produced a matrix of desirable qualities like: open vs closed, common between many programs, handles curves, arcs, spirals, handles different materials in partial or complete contact, handles embedded (pre-made) objects, handles embedded wiring, handles fiberglass along stress lines, cost of programs that can create files in that format, how many other programs can import that file format, how easily complex shapes can be shared in that file format and then re-edited by someone else, etc.

I believe that this kind of a trade study in file formats, programs, and compatibilities and capabilities should count well for a project.

Mike