I'm printing on my trusty LulzBot TAZ-6 ... it's gotten to the point that I can trust it to print multiple objects for super-long prints ( 78 hours is my record so far ) without fear of failure. I've only put two mods on it -- a heated build chamber to stabilize my ABS prints and prevent warping, and filament detector by Andrew Tunell that tracks if the filament is being fed and pauses the print if the filament has not moved at all in 8 seconds.

( Andrew Tunell's wonderful detector is : Filament Monitor ... I think of it as a must-have for ANY printer )

 

First print of a vertex complete -- print time 10 hours 8 minutes, 300 micron, 6 perimeters, 25% infill. PLA, using Gray PolyMax [ edited, corrected ].

That's nearly a full half-kilo of PLA...

 

Looks good, but you probably should be using PET-G or ABS to print the structural parts like this.

 

Second vertex complete! Reduced the perimeters to 4 from 6, same parameters ... saved me about 50g of material w/o compromising strength. 8 hours 16 minutes to print.

Elmo, why do you suggest ABS or PETG? They do have more 'flex' than PLA ... and certainly PETG has a higher temp of melting .. but PolyMAX PLA is the toughest PLA I know of, and the particular chemistry of that PLA gives it properties you may not expect out of polylactones...

[Edited to add] PolyMAX is, as you can see by my prints, prone to "spider web" stringing. Even printing slow-and-low-temp it does require a little post-cleaning. The prints do come out nice looking.

Polymaker | PolyMax™ PLA

 

Any parts that are going to take heavy flexing and mechanical stresses, I'll likely print in eSun PetG or MakerGeeks PetG. Anything that needs high temp or friction resistance, I'll print in PC-Max ( polycarb ) or Taulman nylon.

 

Okay .. all the towers are now printed, as well as the endstops... next step is get the carriages printed.

I have the old-style split V wheels from OpenBuilds already ( from my old WolfStock build ) --- will they fit on the Carriage V3 for large wheels?

 

I need some help re-calculating the correct arm lengths. My sides are 500mm and the height of the sides is 1000mm ( I didn't feel like cutting the rods, and they come from OpenBuilds precisely cut to within a high tolerance ).

Generally, I know people have been making rods to match the length of the sides and that's "good enough" .. but what would the optimum rod length be for this printer, if I want to be able to print on a 300mm diameter heated bed?

 

Take a look at the attached openscad calculator in this message & see if that will give you the answers you need.
Delta-Six

 

Thank you David! And for your message that had some of the values pre-calculated. 517mm sounds about right. I'm starting this build with a 300mm diameter borosilicate glass plate w/ SeeMeeCNC heater -- eventually I intend to increase that to the theoretical max of ~370mm.

 

Has anyone modeled OpenSCAD (or just an STL file) vertexes with 4040 verticals?

 

I made a set using 4x EMT1/2" (18.4mmOD) together which is just a little less than that. I will take a look at that today & see if I can easily change that.

 

I played with this design today and was an interesting puzzle. Here is a STL & openscad file for the Bottom Motor Vertex using 2040 horizontals & 4040 verticals. Width wise, I started out with the same width for 2040 verticals but increasing the depth for the extra 20mm. Because of the bolts that needed to go in the ends of the 2040 horizontals, I had to increase the width for what 2080 verticals would be. I included an image showing a cross section to show the problem on a narrower vertex. The 4040 inner cutout is sized for 40.5mm. If you decide to use this, I would do a test print of a short cube with 4040 cutout to make sure it fits before printing the full print.

BTW, what size Delta are planning with these huge 4040 verticals?

 

Does anyone on the forum know what size bearings Jerry used for the extruder tension/pressure arm? There are two of them, and they are really small. It looks like they use a 3mm ID, but I'm at a loss to figure out which ones to order.

Also - David - GOOD GOLLY those 4040 Vertex towers are monsters. As long as you've done that to them, however ... might as well finish the set. It will require a double on the bottom for stiffness. ( vertex-tall ) Do you have that part SCAD-ed out?

 

Of course -- these 4040 verticals give me an idea... but the tallest V-slot 40x40 sold by OpenBeam are 1500mm...

A 1500 mm tall, 500mm dia heated bed using a 110V silicone heating pad bonded to a circular cut borosilicate glass... 1 -> 1.5 mm extruder using a 24V e3d V6 + Volcano .... this has possibilities for printing Very Large Things.

 

You asked for it. These are untested & I do not plan to print these monsters. Here is both the single & double height using 2040 horizontals & 4040 verticals in STL & openscad format. I found one discrepancy in the motor cut hole, so I changed that on the single height one also. Let me know if you see something I missed in the design.

 

Awesome, thanks David! The description from Martin above almost perfectly describes what I've been working on - 1500mm tall, 500mm rods, 24v e3d volcano, duet wifi, etc. I've been working on it little by little based off of justin edward's BigDelta design, which is really close to the original Delta Six design. But I've not been too happy with some alignment issues, and I've tracked it down to the pretty bad job I did when I printed my vertexes the first time (not all of them were printed using the same settings, and so I think it's throwing my calibraiton out). So if I'm going to go to the trouble of reprinting the vertexes then I wanted to see if there was a bit more updated design that took into account all the hard work over the last year. I'll print a one tonight and see how it fits in the next day or so.

P.S. Justin's github is at GitHub - justinledwards/BigDelta: BigDelta 3D Printer reprapish, fyi.

EDIT: Oh wow, the double height one is ridiculous, haha. I will not use that. Should be stiff enough w/the 4040s it's not required.

 

That is an interesting build. They have some nice vertex STL files there. From looking at their 4040 vertex, maybe I could have made it narrower, but you have make sure those end bolts are installed before putting the verticals in & you would not be able to get them out without removing the verticals. I also chopped the inside vertical v-slots out like the other versions in this thread. Good luck on your build. I might build a bigger delta this winter after I learn a little more from playing with my Folgertech 2020 Delta.

 

I've had a lot of issues getting something this large calibrated right. Mostly squaring issues due to my assembly, I think. I've got a 425mm mic6 plate with a PEI surface and a 400mm heater glued to it that I want to use as a build plate once I get it back together.

 

I found that if I print vertexes on a badly calibrated delta, they tend to have a slight "twist" in them that is not evident until you get to the 1 meter mark. To build my 20x40 - 1000mm tall x 500 mm wide delta, I chose to print the vertexes on a well calibrated LulzBot TAZ 6. We will find out tomorrow or thursday how well I managed to calibrate the print, and how square the final build is ... it's always a challenge.

However, the ABSOLUTE NUMBER ONE cause I've found of Delta machines that won't converge to a solution, is that the length of the rods is not absolutely the same. It's super imperative, especially on these large machines, that the rods be very precise. I build and glue mine together using a jig that's bolted onto a spare piece of 1000mm OpenBuilds V-slot rail I can't use to build a machine (has a dent on it). The template was inspired by one I saw on Thingiverse:

Rod assembly jig by ichibey

The number two reason? Inconsistent lengths of the sides. OpenBeam does a great job of cutting their extrusions with extreme precision -- my 500mm lengths were 500.05, 500.08, and 500.03 mm respectively when they arrived. That's "close enough" to make any errors in length my problem.

 

Yep. Visually it looks great, yet there's a pretty good effector tilt from one side of the plate to the other (like 3mm+). Which maybe I could dial out in software, but it's bad enough I want to fix the hardware side instead. I'm printing on a Lulzbot mini. I think the issue really was I printed the vertexes from different filament types and experimented with different infill and speeds while printing... which made them non uniform enough that over 1.5m it got pretty bad.

6 Vertices is >1kg of filament though, may end up doing 3 in one roll and 3 in another to match.

 

Those big Double Vertexes I used on my bot are 375 grams each ... I couldn't quite get three out of a 1KG filament roll. Took ~10 hours to print each one too.

If I decide to build an even bigger bot after based on these 40x40 extrusions ... I'm going to have to start asking ProtoPasta to send me PLA+ in 3kg spools!

 

3 x 10 x 4mm Deep Groove Ball Bearings 623Z

 

Thank you DiggerJ!

Confirmed ... I scavenged some 623Z's from my old CupCake Lowrider XY build ( I replaced the 623 bearing version with a UU8 linear bearing mechanism a long time ago, and kept the 623 bearings. )

=====

Since I've decided to go for the Compact Effector -- what have people used for an auto-leveling probe? I'm now at the stage where I need to pick some kind of auto-leveling mechanism for the build. With a build this large, I have to reject the FSR-style leveling simply because the heated bed will destroy that kind of leveling probe ... FSR's melt.

=====

 

I am using an inductive proximity probe on my Delta which I am happy with. If you are using a glass bed, here is an interesting method to use an inductive probe with that. Inductive proximity z-probe adapter. Works with glass cover. by paul_delta

 

Yeah, I'm going to be using a glass bed, so those inductive probes are going to prove to be too much of a PITA to use for auto-leveling the effector and calibrating the printer. I've been looking into them, and the assemblies are pretty big.

I'm starting to look at things like the BLtouch, or perhaps copying what LulzBot did and putting the level points all around the perimeter of the glass as clips and completing a circuit.

 

Use a 8mm approach inductive sensor instead of the common 4mm.

 

That would require that I use an aluminum build plate right underneath the borosilicate, and then put a heater under all of that. I'm planning on using the SeeMeeCNC heater directly bonded to the borosilicate plate, mounted into a milled out piece of FR4 ( same stuff circuit boards are made of ). I don't think inductive sensing will be the right approach for this build.

 

Did you look at the video on that thingiverse thing? You do not need a metal bed with his method. He uses about a 2mm OD metal nail that touches the bed & a spring moves the nail head close to the sensor to engage the proximity. My inductive sensor is only 1.23oz including the wire. More weight & trouble to hook up probably than the bltouch, but a lot less expensive.

 

Yeah, looked at the video and the design. I have some doubts as to the repeatability and accuracy of the setup. Assuming I use one of those 4mm inductive sensors, I'm not sure if I could get a 0.005mm repeatability with the combination of that nail head, pin, and spring. I'd want to have a precision milled CNC pin and sleeve for that setup, and a way to make sure it consistently calibrates.

Well, $38 vs ~$20 or so. Plus, I have to go find some very small/light spring from a set of springs in Home Depot, and a nail to fit, and...

 

Look at dc42's solution. I've got one, in limited testing it works well. Mini height sensor board