Grounded Delta (Name in Progress)

So, a while ago I posted about a grounded (inverted) delta project of mine, promising stupidly optimistic completion dates and seemingly giving up on the project. For those of you unaware, it is a redesign of the original GUS Simpson reprap with some basic improvements and some design changes.

Fear not! I didn't give up, only procrastinated (and life got in the way). The project is nearing completion, I'm just finishing off the shoulder, which took a surprisingly long time to get the geometry right to accommodate the new stringing and figuring out how to arrange everything on the base of the printer. After that, I just need to knock together an effector which shouldn't take more than half an hour (famous last words). In this finalizing stage, I'd like to ask if anyone is interested in building one and helping me debug and improve the design in the coming year? Any features you want to see?

If interest builds around the printer I'll put together assembly guides and FAQs. I'll also be posting some STLs of the parts that are ready, and eventually the OpenSCAD code once all the parts are modeled (and the clutter is cleaned up.... yeeesh....). There are 3 parts of this printer that aren't off the shelf/3d printable and will need to be machined (frame, oddly shaped print bed and an optional removable print surface). I'll be putting together a CNC in the coming months to machine these for myself and would be happy to machine a set for someone else. Worst comes to worst, the frame could be 3d printed and the bed either replaced by an off the shelf shape (no automatic steps/mm calibration :( ).

So far, it looks cool if I do say so myself.

EDIT: Also, suggestions for the name are very welcome. So far I've gone through several names. Spyder, because the way it prints reminds me of how a spider spins a web and spelling it correctly would be boring. DeltaGear, because its a delta driven with gears. Geared Up Delta, as a nod to the GUS (Geared Up Simpson).

Quote:

---

EDIT: Also, suggestions for the name are very welcome.

---

Pictures of the printer with some nice prints it did on the bed would go a long way to generating some interest. Is your printer running Marlin?

The printer is still in the design phase, hence no pictures. You can check out pictures of the original GUS and a revision done a while ago by jfriessen to get an idea of what it'll look like. I'm doing a redesign to move the motors off the arms and have them stationary in the base, and some other minor tweaks. I should be putting together the mill I need to make the frame and bed in the next two months or so, once that's done I can build the printer itself.

The firmware is TBD. No firmware (as far as I'm aware) supports the GUS Simpson kinematics, and I have an auto calibration routine in mind to help compensate for any inaccuracies from the printed parts used for the printer which I'll probably need to code myself. The controller will be 8 bit for cost reasons and I'm tossing up between Marlin and Repetier. Marlin seems to support a lot more new features, but I hear the code base has gotten a bit less beginner friendly due to all the different people who have contributed. I hear repetier has a cleaner code base. That aside, I've also heard some people have had better luck (less stutter/pauses) with repetier on deltas than with marlin. I've also heard some say there is no difference, but I haven't heard people say they've had better luck with marlin so I'm inclined to think there is something to it.

A slight update, I'm running into a design hiccup of trying to get a thick enough arm using bolts of reasonable length (i.e. not 150mm long) and have considered two options. Either splitting the gear in half, which will allow me to nest some parts to make the assembly more compact, or print upright. I've been avoiding these two outcomes because the first one introduces more opportunities for misalignment when the gear arm is in two pieces, and the second results in teeth that have a stepped surface rather than a smooth contour. I thought the stepped surface on the gears contributed to the original gus's inaccuracy but now I'm not sure.

Any thoughts on the viability?

none whatsoever - it's all greek to me :-)

But you could call it 'Hamish'
No reason, other than I like the name 'Hamish' lol

Haha, I'll add hamish to the list. I've managed to get the bolt sizes down to something reasonable (ish). I think the first model will have a print volume around 100mm diameter and 100mm tall. The scad file will be completely parametric of course so you could change that to whatever but that's where I'll start. I've decided to add an option for how to define the machine, instead of giving it a bed diameter for the base of what is roughly a triangular base pyramid, I'm adding support for the user to define a cylinder that they want to be able to print. I figure it'll be easier to visualize and people will be more used to this from linear deltas (plus it means I don't need to find a way to incorporate the funky pyramidal print envelope into a slicer).

https://www.thingiverse.com/thing:2552984

Ain't she purrdy? Finally finished the shoulder. Getting that thing right so everything lines up, fits and doesn't get in the way of anything else was surprisingly hard. The gear arm STL on there is old, and the design has changed a bit. I'm hoping to have the finished files out in the next month or two.

looks good - but why the low poly stl ?
I know it doesn't make any practical difference - just looks better with more roundness :-)

Because I'm on a netbook and it has a heart attack and a half when it has to preview, let alone render, high resolution models :P

The resolution can be easily adjusted in the file from a "res" parameter I've created that controls the length of the segments that make up the polygons. I believe it's set to 3mm in the render.

It's really coming along. Just ordered the non printed parts (I'm probably a bit short on cash now but Uni will be starting soon so I won't have much time...). They should arrive within the next month or two, then I'll start a build thread. The last part left to design is the base/frame (and a hotend mount, but that WILL be easy, I hope), that should be fairly straightforward and easy (famous last words, each time I said that it's been wrong).

I'll upload an STL of the effector pieces soon. The effector/shoulder offsets ended up a bit bigger than I'd like because of some collisions I hadn't anticipated when I started, but that's that. The machine looks a bit clunky and disproportionate because things like the bearings, nuts and pulleys take up a lot more of the machine at smaller scales. If you build a 250mm+ machine it starts to take on some nicer proportions.

For anyone curious, V0.0 will be a 50mm radius, 100mm height cylinder build volume. Of course, due to the nature of the printer it can actually print bigger but no slicer supports this shape of build volume so I made it use some more relatable dimensions and shapes.

By the way, don't use the STLs on thingiverse to build one of these, they were all generated with different settings so I can't guarantee it will make a usable machine. I can almost guarantee it wont.

what electronics setup are you using ?

Control board is a Mks Sbase, chosen purely for the reason it has software settable driver current and was the cheapest board I could find with this feature. The SD card based text file config is another bonus for making tuning and setup easier.

The motors are the cheapest NEMA 17s I could find off aliexpress, hotend is an e3d v6 clone from the same place. No endstops because that's part of why I needed software settable driver current (the other part is no fiddling with trimpots). If I'm missing something please tell me because I forgot to order it then :P. PSU if you're wondering is one of the cheap LED PSUs off aliexpress, 120 watts which is plenty for what I'm using and could probably handle a small heated bed. My solidoodle 2 has a 100W power brick and runs a 150x150 heated bed.

I'm not using a heated bed because I couldn't find 100mm diameter heatbeds, the smallest I found was 160/170 or so. Actually, I found one on ebay but it was ridiculously expensive and I'm pushing my budget as it is at the moment. I'm also building a CNC at the same time to mill the bed for this printer and just cause I've wanted to build one for ages for projects.

Out of curiosity, is anyone other than me interested in building this printer? If so I'll work on tidying up the source files for release and doing a BOM and assembly instructions, if not I wont waste my time. For those curious a complete scratch build of this printer would cost about $150 USD + some plastic of your choice. This is with some "bells and whistles" like a power switch for the PSU, the Mks Sbase (not the cheapest board around) and excessive quantities (like ordering 50-100 bolts/nuts when you need like, 20-30) of some things.

If you cannibalize an old printer this goes down to like $30. The bulk of the cost here comes from the controller board and motors, which are probably half the cost if not more. The PSU also adds a fair bit to the cost. Other than the standard components of a 3D printer (hotend, control board, motors, extruder gear, PSU) this printer only uses nuts, bolts, washers, bearings and some fishing line. All of which are the same size to boot. For the record, the cost is more or less the same regardless of the size of printer you are building. If you do need to use larger hardware, the bump in cost is absolutely minimal, and M8 hardware will be suitable for anything 'desktop' sized. The only real difference in the cost will come from the filament you're using, but you'll only need 1 or 2 rolls depending on what size you're going for. If a 100x100 cylinder printer costs me $150, I expect a 300x300 to cost less than $200.

Update, V0.0 will have a 95x95 cylinder build volume. Due to unforeseen chamfering (which is totally necessary, I won't take any arguments) the base/frame of a 100x100mm printer is too large for the 150x150mm bed of my solidoodle 2 by about 3mm.

I could break it up into parts and glue them, but I'm too lazy. I could use my prusa clone but it has no heated bed, I'm down to my last spool of PLA, and someone recently gave me a but load of ABS. I hate printing with the stuff to be honest, it gives me a raw throat when I'm around it even with the windows open but I can't say no to free.

The chamfering was to break some sharp edges created by cutting away some of the base (to save print time and plastic), and made it look really awesome in my opinion compared to what I was originally planning.

I'll upload the effector parts and base sometime soonish, might be in a bit over a week as I'm going down south with my partner to visit some of her relatives for a week.

Holy mother of a fillet. You never realize how bloody hard it is to fillet between a box with an arbitrary translation and rotation, and the cutout of a circle, until you try it. I spent literally an entire week trying to get this to work. All for one measly, cosmetic fillet. Alas it looked out of place un-filleted with all the other fillets I spent too much time on (none of them quite this long though).

Anyways, base is DONE. Last time it was "done" I forgot to add motor mounts. The absolute last part to design (I swear), is a hotend mount for the effector. Should be easy, famous last words.

Scratch that, still have the drive gear/pulley and motor gear to do as well. Those should be quick.

The design is done. I'm not entirely happy with it, mostly because I'm trying to make a small printer with big parts and it isn't working out.

As of yet it is untested, but if anyone wants to try, I'll generate a set of stls for them. I stress this is at your own risk. While I've gone through it a lot of times to try and account for everything, it's no substitute for having actually tested it.