Delta printer build vs "buy"

[soofle616]

So i've decided to try my hand at building a delta printer. For a lot of reasons not really worth going into at the moment I want to design my own but my experience with 3d printing in general is limited to the last month and change that I've spent with my ff creator pro. Part of me thinks, screw it, I'm a reasonably intelligent guy with an engineering background and I have the interwebz to refer to if I get stuck, design my own and make it how I want right from the start. The other part of me is a bit more reserved and thinks that I might be better off downloading parts from an established design and simplifying my first build to replicating what others have already done and then work on designing my own from there. So there is my debate. Buy/print the parts to build a clone of what's already out there and just copy/paste everything I need to make it work, or design my own and get it done the way I want the first time while accepting the higher probability of failure? Those of you that have built a delta, how difficult is it really? Was yours a "kit" or did you design it yourself? Based on the experience you gathered from your build what would you do differently if you were to do it again?

For the hell of it here are some of the goals for the one I plan to design myself:

12" triangular heated build platform (no apparent reason to limit myself to a cylindrical build area)
12" MINIMUM build height
lead screw motion
80/20 extrusions (much easier for me to get than the openbeam stuff)
active print cooling (switched)
enclosed build area with active cooling (switched)
power supply and other electronics housed in base for a cleaner finished look and easier transport
provisions for 2nd extruder (will not be part of the initial build)

[curious aardvark]

one thing I'd add to a self design is a flexible drive shaft to the print head. That way you can keep the motor on the frame but have a direct drive feed on the head that's really lightweight. Dual extruder maybe.

I saw this done at the tct show.
It was one of these guys who'd built it: http://www.wedo3dprinting.com/

Lets you have use 1.75 filament. Gives you the control of a direct drive with the lighweight design of a bowden setup.

I'm sure if you email them they'll tell you where to get the driveshaft.

[soofle616]

one thing I'd add to a self design is a flexible drive shaft to the print head. That way you can keep the motor on the frame but have a direct drive feed on the head that's really lightweight. Dual extruder maybe.

I saw this done at the tct show.
It was one of these guys who'd built it: http://www.wedo3dprinting.com/

Lets you have use 1.75 filament. Gives you the control of a direct drive with the lighweight design of a bowden setup.

I'm sure if you email them they'll tell you where to get the driveshaft.

Didn't see anything on their site about that but it's an interesting idea for sure. Definitely worth exploring further, thanks for the idea.

[curious aardvark]

Not on their site - but one of the guys in the company had built a printer with a flexible drive shaft and they had it with them at the tct show in october.

[RAMTechRob]

The SeeMeCNC RoStock kit is very nice. I bought the Assembled Delta Orion and I love it.

[soofle616]

Well my design is well underway. I'm on effectively my 4th or 5th iteration. All my earlier attempts would have worked but were, frankly, ugly as hell. This one seems to be going well and I'm pleased, if not necessarily ecstatic about the aesthetic quality of it. A quick teaser rendering of the chassis is attached, obviously I have a lot of work left to do.

I've done a bit of searching but haven't found much useful info, can anyone tell me or point me in the direction of how to calculate the length of the arms and/or the angle the arms should be at relative to vertical when all three carriages are at the same Z? I assume that an absolute minimum arm length would be whatever is necessary to reach from a given carriage's pivot points to the associated effector pivot points with the arm horizontal to the build plate and the print nozzle at the extreme edge of the build platform. That would result in an angle of about 45° when the carriages are all at the same position. Is there any reason to make the arms longer than that other than to improve movement speed? Things I've found so far:

1) longer arms = more error
2) longer arms = less print height for a given machine height
3) accuracy improves as the arms approach horizontal because the carriage moves farther for less effector movement
4) shorter arms = reduced movement speed as more carriage movement is required for a given amount of effector movement

[richardphat]

Longer arms does not mean more errors, it's the fact of control on tolerance and the use of uneven length of arms. Any well picked arms will work just fine.

Speaking of error caused by the arms, it's probably a fraction compared to the end results when you print. Shrinking is the killer and the error can be as big as 10 times your accuracy of positionning.

Here's my tip, if you design a delta, make sure it's overconstrained when it comes to the carriage running up and down on the tower. As well as the arms attaching to the hotend device.

[soofle616]

Longer arms does not mean more errors, it's the fact of control on tolerance and the use of uneven length of arms. Any well picked arms will work just fine.

Speaking of error caused by the arms, it's probably a fraction compared to the end results when you print. Shrinking is the killer and the error can be as big as 10 times your accuracy of positionning.

Here's my tip, if you design a delta, make sure it's overconstrained when it comes to the carriage running up and down on the tower. As well as the arms attaching to the hotend device.

Maybe error was not the correct term, mathematically longer arms would result in less resolution because for a given amount of vertical movement, there would be more horizontal movement with longer arms vs. shorter. Basically the math tells me that the closer to horizontal the arms are, the more vertical motion is required to obtain the desired horizontal motion. More vertical motion = more rotation of the stepper motor which means more steps are being used which means higher resolution because I don't have to rely on microstepping which isn't necessarily precise. As for controlling the length of the arms relative to each other, that shouldn't be too much of an issue.

As for your tip on over constraining, I'm not sure what you mean. To me, over constrained means failure. Telling two things they are the same length and then telling one that it's 1 inch while the other is 2. That would be an example of an over constrained condition at least as far as I understand the term. As far as ensuring "pure" vertical movement and solid fixing of the arms to the carriages and effector, that's being dealt with. Rather than the usual slide bearings that I see a lot of in homebrew printers I will be going industrial on mine. I can't afford preloaded ball bearings so I'm going with high precision plain bearings. Based on what I've seen of them in use in my shop they are as precise, suffer very little wear even in extremely abrasive environments, have minimally more resistance to sliding, and carry the same or greater axial and cantilever loads as ball bearings do. Not to mention I can get 4 of them for the cost of a single ball bearing. There will be 4 in each carriage riding on 2 rods. For the moment I've spec'd out 3/8" rods as a compromise between rigidity and excessive size but I have the option to increase that and/or go with open bearings and a fully supported rod if I decide it's needed.

Movement will be controlled by acme lead screws with doubled up nuts in an adjustable housing so I can set preload on them and eliminate backlash. I will be using heim joints to connect on both ends but I'll be using aluminum rod ends so they will be stiffer than the usual plastic bits. I thought about going industrial with the joints as well but even the smallest ones are rated for such extreme loads that I'm concerned they will introduce excessive resistance to motion on the system. As for the rods themselves, I haven't decided on a material or source yet. I would love to go with carbon fiber for the low weight, extreme rigidity, and cool factor, but I have to see if I can find stock material in the length and diameter that I need. If not I'll go with aluminum.

[richardphat]

Oohh ,if you talk about the resolution, again, should you be worrying that much? I don't think so, because you're likely going to use NEMA 17 stepper motors(1.8deg) and you can do microsteps. Even that, depending on if you use belt or string then it may or may not matters. But keep in mind that your nozzle will likely cripple all movement and you will have srhinking that outcomes when speaking of error/tolerance.

When I was mentionning overconstrained, I did not mean on a 3D cad perspective, but what you mentionned about the number of rods for each carriage. Because having myself with 2 Delta self assembled, I see flaw where the hot-end mechanism wobble.

I designed my laser burner printer based on 3/8th rod I found from inkjet printers, the linear bearing cost for this typical dimension cost way too much and I have to scrap the whole design. I decided to downgrade to 8mm shaft and use LM8UU bearing.
The cost for the linear bearing drop from 15 to 2-3$/unit.

[JohnA136]

I like your design so far. We built a couple of Kossel Deltas, we printed all of our own plastic parts and sourced out the hardware to various vendors. Both have heated build platforms and both print very well. We have pony printed PLA on these machines because they are open framed and we already have six different closed frame printers for ABS.

Of course, the downside on the Kossel is the Power Supplies do not fit under the machine. We also have our Ramps Boards mounted on the outside of the frames so we can have active cooling although I have seen some Deltas with the boards mounted under the Build Platform