Design Philosophies

Hey guys, I'm currently designing a printer and would like some opinions on a matter.

Is it best to minimize the number of (different) parts, or to use different parts where they are most suitable.

For instance, designing a new printer I could make all the bolts M3, since that is the only 'nessecary' size, since that's what the motors are mounted with. Idlers, bearings etc all come in sizes to match this.
In this case, if you used m3 for everything you'd end up with some fairly weak joints, use it on an idler and that idler bolt could get bent since the bolt is quite thing. Or you could choose an 'optimal' part for each joint/task, and thus end up with more (different/unique) parts. The issue here is that it means you need more different types of parts on hand to repair it if it breaks, rather than a 'one part fits all', and assembly gets more complicated as you have to search for a different bolt/bearing/nut for everything.

As someone who's put together a few printers, do you have a preference? If something breaks, think about it not just from the perspective of a person who has a drawer for every bolt under the sun, but also the average consumer who may not have many if any spares available, and would need to source the parts for repair. During assembly, how many types of bolts to sort through for assembly is too many?

So far, I'd need at most M3, M4, and M8. I can cut this down to M3 and M8 (at some added $$ expense), or purely M3 (at the expense of relying on M3 bolts for the joints, I'm afraid they might end up bent from printing forces). Does three bolt types (and three nut types to go with it) sound reasonable to people? Also, if someone could chip in on the matter of M3s getting bent, do you imagine they'd get bent if used for a pivot joint (with a bearing, of course)? Think scara arms.

Less parts. Coming from someone who has designed and sourced his own custom printer, shipping is a bitch. Take this with more than a grain of salt.

MathewHall has a great point there in shipping: do the shopping, and do the math for the shipping, even if you can get all parts from the same place, are you likely to keep spares for all the types of parts?

As for screws and bolts, there likely is a way you can use multiple weaker/smaller ones in the place of a larger one, or, you could overengineer for strength by using all larger ones.

Of course, more and larger parts has an energy and possibly speed price while running: engineering is the budgeting of reality and not everything can be optimized, so this is an optimization for costs (theoretically) and keeping a minimum number of spare parts, or at least of shipping for new parts.

I'm all for similar vitamins. I'm thinking about modifying a printer to use 8mm Z axis guide rods instead of 6mm... but it's just one more thing to buy, vs a bunch of 6mm rod stock.

I REALLY like the last printer I built because it used M3 screws exclusively. For someone who does not live in the metric part of the world, it vastly simplified purchasing hardware.

For the OPs case, I'd have to err on the side of picking the right screw for the job though... Minimize when possible though, i.e. M6 or whatever for SCARA arms if needed, but M3 everywhere else if possible.

I've actually managed to get it down to just M5x60 (the bolt diameter and length is up to you, that's just the one I chose). I might add some m3 screws for motor mounting but I'll try avoid that. Side note, it'll support imperial hardware, you just need to convert it to metric dimensions .

It seems like the effort to constrain to a single bolt size will only make things much more complicated. Anything you buy has its own native hardware set.

For example, 20 series uses m5 bolts. The bolt depth in the extrusion itself is 6mm, but we don't want to actually have bolt bottoming out. With that said, anything in the 3-4mm thickness range can be mounted with 8mm long bolts. This covers pretty much every type of panel as well as corner brackets and a few other bits. From there up to 1/4" (6.35mm), you can use 10mm long bolts. This covers everything that needs to be mounted to the frame, such as printed motor brackets, shaft clamps, pulley blocks, etc. There you go, you're entire assembly is now made from m5x8 and m5x10 bolts. From there, everything else would be relatively limited specialty, such as m3x8 (or m3x10 depending) stepper bolts, a few longer m3 for clamps, and a few longer m5 for pulleys. That's your entire build.

different extrusion has it's own native size. For our purpose, this goes up to m8 with the 40 series extrusion. My point it, stop trying to design in a cheap way, and design the way the design itself wants to be designed. That will give the most uniform part selection in the end while catering to exactly what your main components need. It will also make assembly much easier as everything will be naturally made for each other. You won't have to fiddle around with aligning undersized bolts.

Moving on from nuts and bolts, if you start with a basic approach of symmetry, then your unique parts will be somewhat cut in half. You may have to mirror some parts, but they aren't actually a different design. For a core xy, you're designing around pulleys that happen to be the same all around the printer. Why make one a 16t and the other a 36t just because it's cheaper? Just do it uniformly and everything becomes simpler and less costly in the end. The same goes for a delta. Each arm is identical, which only triples the complexity if you choose to stray from the materials natural hardware sizes.

If you start to stray from what wants to naturally happen, then that's your problem. You are now over or cheap engineering it in ways that will provide zero benefit.

Edited 1 time(s). Last edit at 05/11/2018 10:54PM by boredom.is.me.