repeatability of desktop 3d printer

In my redesign of the gus simpson ive rerouted the cables along the outside to make stringing easier and let it print flat without support. However, this combined with the addition of a couple of pulleys where the shoulder meets the arm have resulted in ridiculously wide shoulders with rather narrow (flimsy) arms.

I cant seem to find a solution to this with my current design and have started to consider splitting the gear arm in two halves and sandwhiching the pulleys and stuff between them. Nesting the parts like this would fix the issue but this idea is something i've avoided from the start for two reasons. First it adds more parts and slightly complicates assembly, but this is seeming like less of an issue now. Second, and more importantly, I'm worried about how identical the two gear halves will be when printed and any misalignment that could occur because they are seperate parts.

I'm just wanting opinions from people as to whether they think the arm halves would be sufficiently identical and if there would be significant misalignment. Thanks in advance.

If you've already made the leap of faith that printed gears are going to be OK driving a precision mechanism, they'll be fine whether you split them or not.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

That's like saying just because you nicked your finger you might as well cut it off. The printed gears have been proven to work, the question is how much slop can you introduce before it becomes too much.

Quote
Trakyan
That's like saying just because you nicked your finger you might as well cut it off. The printed gears have been proven to work, the question is how much slop can you introduce before it becomes too much.

Yes, you are right, printed gears work, wooden gears too BTW.

Now how about making just a prototype and see how they "work". (Not in wood, you could really cut your fingers)

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

Quote
Trakyan
That's like saying just because you nicked your finger you might as well cut it off. The printed gears have been proven to work, the question is how much slop can you introduce before it becomes too much.

I think your question is more like asking if it hurts more to cut your finger half way through or to cut it completely off.

Edited 1 time(s). Last edit at 01/04/2018 07:31AM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

The original design worked, and was printed. I'm still working on my design, I keep getting bogged down in small features, once it's done I will print it and demonstrate it.

Quote
Trakyan

"bogged down in small features" .

the devil is in the details !

As I always say, nothing like a prototype.

One learns more from mistakes than from success.
That way I learned a lot, ... mainly from other's mistakes. Saves time and frustration.

I saw the "Gus Simpson" video, yes it moves around a "head" but I don't call that "working".

A nice exercise true but as a practical 3D printer, no way.

So fine if you prove us wrong.

Edited 2 time(s). Last edit at 01/06/2018 04:05AM by MKSA.

---

"A comical prototype doesn't mean a dumb idea is possible" (Thunderf00t)

[www.youtube.com]

It prints, if what you're implying is that it doesn't. There are several other people who have made videos of the original design printing, and another person's revision.

I'm fully intend to build a prototype, and will post pictures/videos/files when I get one working. I need a custom cut frame and bed for my prototype though so I'll be building a CNC to cut it first. I'm still working on completing the design of the printer itself so I need to wait for that too.

I don't know if it convinces you to print the Simpson, but I just made a dozend parts for a set of printable Delta magnet arms.
They are all the same, including a small gap in the brim, which I can't explain.
What I'm trying to say is, if you have a working gcode and a reliable printer, the gears will mesh. But getting the perfect gcode will cost you some wasted filament. Also make sure, you buy plenty of filament from the same batch, so you don't have to re-tune the printer in between.

Another remark: Why do the gears have to be printed? Can't you use GT2 belt or similar, clamped to the arms, instead?

My issue isn't with the gears meshing, it was whether a gear split in half and then assembled together with some bolts and spacers would have good enough alignment.

Belt isn't really made to mesh with itself like a gear, but even if it was this doesn't address the alignment issue. Anyways, since I'm relying on the printer to print a round backing for the belt, I might as well rely on it to print the teeth as well so they're integrated into the profile and have the correct tooth shape.

I'll be designing two versions, one with a single piece gear (and just dealing with the need for longer hardware), and one with a split piece gear and see in practice which I prefer or which works better (if there is even a difference).

Maybe using a herringbone gear and splitting it down the middle would be a good way to do it. Or better still, use the split to make a backlash free gear by allowing the two halves to slide against each other with a spring between them.

Edited 1 time(s). Last edit at 01/07/2018 08:59PM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

It will be a herringbone gear, and backlash wont be an issue as the gears are just for registration as the arms roll along each other sort of like cams, the gears themselves dont drive or transfer any motion. They basically serve as rolling cams and the gear teeth are just there to make sure they move symetrically. Besides, the gears are lashed together with a tensioned block and tackle which should eliminate any backlash in the teeth.

I saw the videos and the herringbone gears seemed very huge to me. Probably because the printers, which made them weren't able to print smaller teeth reliable?
That's the main reason for my GT2 belt idea. They are much smaller and thereby the arms could be smaller too.
And they mesh very well...


Edited 1 time(s). Last edit at 01/08/2018 02:50AM by o_lampe.

The size of the gears isn't related to the size of the teeth, I think the tooth size in the original was chosen so the teeth weren't super small as to cause printing issues. I think the tooth size is arbitrary and just limited to how small you can print them, on my version, you can specify a number of teeth you want and it'll calculate the size from there. The size of the gear is related to how much mechanical advantage or 'gearing' the actuator has, a larger gear gives more mechanical advantage and so a higher resolution and force. I'm not sure what effect the size of the tooth would have (positive or negative) but it's definitely worth investigating, I'll print gears with different sized teeth to test out. I have a feeling it'll have an effect on rigidity of the actuator (the number of teeth in mesh, smaller teeth, more mesh. A larger helix angle should increase the number of teeth in mesh, something else for me to investigate.) and possibly the wear on the mechanism.

In terms of how well they mesh, I meant the tooth profile and backlash wise. I believe they are based off plain circles which don't usually make a good gear, that and the relatively flexible rubber teeth could impact rigidity. Overall I'll be sticking to printed gear teeth for simplicity and reprap-iness. The machine is supposed to be super low cost as well and while belts aren't exactly expensive these days cost is one of the reasons for using spectra line.