WeldingRod Bot: Exactly Constrained 3D Printer

A 3D printer based on the Cobblebot kit or OpenBuilds parts that is not over-constrained - smooth motion with minimal adjustment.

I'm printing now!

*** The name Cobblebot is the property of Cobblebot Inc. This work uses the kit of parts provided by them for a "Basic", but is radically different from their design. ***

I'm working on a modified build for the Cobblebot or OpenBuilds parts that will get it closer to an exactly constrained system (and much less sensitive to rail alignment). The target is a 15" cube build volume. Please note that there are both English and Metric versions of most of these drawings. If you get them plasma cut, use 1/8" steel; it cuts REALLY well. If you get them laser or waterjet cut, use whatever you want. They 3d print acceptably, but the end result is kind of floppy; I don't recommend it, but I did try it during my (many) rounds of testing.

Actual motion range:
X: 16" (which means that I can get BOTH heads to move over a 15.2" build area)
Y: 15.5"
Z: 17.5" with a little Y restriction on the front
16.75" for full Y travel

This weekend I assembled version AA of the Mad Scientist (whoops... Engineer) Bot. It is designed to be exactly constrained; no more than one thing controlling each degree of freedom (X, Y, Z, Theta, Phi, and Psi). With no actuators on it, I can grab the extruder carriage rail and wave it around (up and down, in and out) completely freely; WITH NO ADJUSTMENT! That's right, NO tweaking of rail alignment at all! Assembling the frame was amazingly fast!

I'm getting version AD cut today; it deals with mounting motors, belts and end stops. More bulletins to come!

The way it works is there are two Z assemblies, each controlled by a lead screw. Each of them has a long bearing on the front corner post that handles X, Y, Phi, and Psi. A pair of wheels runs on the left and right flat surfaces of the back corner post and controls Theta rotation, but NOTHING else. Z is controlled by the lead screw. The Y carriage has a long bearing on the left horizontal rail that controls X, Z, Theta, and Phi. The other end of the X rail has a pair of bearings running on the flats of the right hand horizontal rail that deals with Psi. I have set the Y drive up where it can either be driven from one side only (the side with the carriage), or driven on both sides using a long 5mm shaft across the front; I'll experiment to see which one I like better. Finally, the X motion is very similar to the stock extruder carrier, except that I moved the extruders really close together and provided a proper belt attachment. The extrusion volume should be close to a 15" cube.

Sorry, I didn't take any photos yet; too excited building!

I did a LOT of tapping the ends; this is a RADICALLY stronger/stiffer connection, if you are willing to do some drilling and tapping. The best choice for tapping the ends is a "gun tap" that pushes the chips ahead of it. McMaster http://www.mcmaster.com/#2703a73/=ymv413 is a good choice.

OK, I assembled rev E night before last (Rev F is going together tonight).

Big stuff:
Optimized all the plates (after initial fitup on AA)
Added motor, idler, and belt attachments
Added the ability to mount the hot ends UNDER the X carriage
Moved everything down to get the hot end to reach the bed
Created Z fittings with two different lowerings (to accomodate those of you that didn't buy 3/8" thick Aluminum beds)
Designed a bed
Designed an insulation holder for the bed

Lessons learned from Rev E:
Corner blocks DO NOT ensure your rails are rotated square to the other rails!
My planned end stop locations interfered with stuff
The belts are kind of short
Parts of the Y ran into the base
My X rail was hard to install
I didn't buy enough idler mounts
The power supply requires moving one of the bottom 20x60 beams over 20 mm

The AG revision is designed, now to get it cut...

Lots of rounds of cut/print, and try... I'm hoping revision AK is the magic one!

I got my Keenovo heater/temperature controller installed this weekend onto my 3/8" thick Aluminum build plate (I know, excessive).
Dang that thing heats fast! When I set it for 40C, I could feel the silicone heater warm up in 1 or 2 seconds! It got to temperature and overshot 2 degrees C in perhaps 10 seconds. The plate, of course, has a lot of catching up to do.

In my "mad scientist/engineer" mode, I couldn't resist adding instrumentation ;-) I have four miniature wire thermocouples set at center, 50, 100, and 150 mm radius and a separate thermistor. I insulated the base plate with Kapton tape, put down the sensors, added another layer of Kapton, then a layer of Copper tape over them for interference protection. I have also designed an insulation retainer for this plate. I promise that I will take some measurements on how even the heating is!

I test fit 3d printed AK version plates this morning; I think I have it now! Mwa-ha-ha-ha!

Ok, AK got the belts almost right, and got the X endstop to work BUT, the Y endstop hit the top frame and the X belt idler hit the top frame. Argh! On to AN and AM.

OK, Revision AP for the frame plates is the winner! The motions line up nicely, run smoothly without adjustment, the belts work, and the limit switches work! Time to start wiring! That, and make the magic stepper motor to shaft connector WITH a timing pulley on it.

The cable chains came out well! Cutting and extending EVERY single wire... Ugh. There is a photo of what the series Z motor harness needs to look like, in case you want to make it the way I did. You can check your Z wiring by shorting each phase and turning one screw to lower the gantry; the other side should lower too. It will NOT raise this way; not enough power output! The motor you are turning acts as a generator and drives the other one. I hope to plug everything in tonight.

Key item: I un-soldered one side of D1 on the RAMPS board to isolate the Mega from +12. I will use a regulator to feed the Mega with 9 Volts to bring down the power dissipation.

Another tip: my prox sensor produces an inverted output. Took a while to diagnose!

I am printing now; the first try on a 20 mm calibration cube looks good!

I printed a test piece with the far side Y belt disconnected to see if that makes any difference in the prints. It BARELY makes any difference. I will put the belt back on, since I already have it in place, but its optional now!