modularity vs integration in design

So, I like simple, straightforward designs to achieve a goal. A while ago I posted a thread to this end asking about BOMs, and whether it was better to use the 'ideal' part for each purpose at the expense of a complicated BOM, or standardize the parts you use for a simpler BOM. The simpler BOM makes sourcing and assembling easier, but can lead to come compromises in the design. Ultimately I chose to standardize, I'm using only one size bolt in my machine which is on the extreme end of standardization but did not come with too many compromises in my cases. I may be adding a smaller bolt size for mounting small things such as limit switches and the like without resorting to snap fits which can be troublesome.

Anyways, I ran into a bit of a dilemma when designing a part of my printer. I'm doing a GUS Simpson redesign which places the motors in the base, and I could integrate the drive pulley into the shoulder, or place it elsewhere as its own assembly on the base.

Integrating it comes with a couple of advantages. For starters, simpler and easier assembly since there is less parts. Fewer parts also means saving the cost of a few bearings and bolts. I guess it's also a bit cleaner and tidier. The disadvantages are that the shoulder needs to cantilever out further to accommodate the drive pulley which makes me worry about rigidity. It also means the design is less modular, if you want a larger drive pulley (to avoid the triangulation error from string walking) or want a larger drive gear for more mechanical advantage, you also need to replace the shoulder piece not just the drive pulley. Realistically if you want to enlarge the pulley to avoid triangulation error (if you're making the arms bigger), you may need to enlarge the shoulder anyway so your larger arm clears your print envelope.

I do have the third option of integrating them, and accepting the triangulation error from a smaller pulley in order to shrink the shoulder size.

Opinions? The initial dilema was over rigidity but it brought to my attention the modularity of the design which was one of my original goals as well as simplicity. I could implement all three and let the end user decide, I guess (I probably will, since I'll be wanting to change between them myself), but out of curiosity, what would you guys choose?

Do you have any images I'm struggling to visualise your setup. How would you define the shoulder?

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

I don't have a complete model of the printer, but I can post some stls a bit later of a rough draft of one of the designs for the shoulder (don't have anything for the other designs yet). If you google Gus simpson, the printer in basjng mine off, the shoulder is the part that connects the bottom of the arm to the base of the printer. I'll see if I can post an image highlighting the part a bit later if that doesn't clear it up, I'm away from a computer right now.

Good luck trying to use one standard screw, when you need 3mm for motors, 5mm or 8mm for pulleys(sometimes a bolt rather than custom shaft), 4mm when you cant use 5mm, 6mm when you want a little more than 5mm....different lengths.

Edited 1 time(s). Last edit at 11/04/2017 11:22PM by MechaBits.

I'll be using a sort of clamp to hold the motors in place, rather than the m3 holes in the motor body. If that doesn't work out I can add M3 screws for that and mounting the other small bits and pieces like limit switches. I'm not super sure what your other points were though, using a 8mm bolt in place of a 6mm bolt usually isn't a deal breaker, it might be overkill but it wont stop the design from working. As for the different lengths, I'm designing the parts around one length screw, and so far it seems to be working fine, sometimes it seems like an extra 5mm longer or shorter would be nice but never deal breaking. We'll see how it works out but so far I'm optimistic.

The original GUS Simpson from what I'm aware mostly used M8 hardware anyway, and with just some minor differences in length. I think I wont have too much trouble slimming down the BOM a little.

well if you set out to only use 8mm's and stick to that you will get what you want, 8mm's all round, i can see what you mean ie the gus does look like something that might get away with it, & benefit from it.

Yeah, and with the GUS Simpson, I'm basically forced to use the same hardware in a lot of places, since most (maybe even all, to be honest) of the bolts in my design go through a bearing and pulley which are all the same size. In some places I could change the size of the bearings and pulleys (a few of them need to be the same size to avoid triangulation error), but that just seems like a silly design choice in order to make the BOM more complicated.

Anyways, any ideas on the modularity? Personally I'm leaning over to having two designs, one that integrates the two but sacrifices avoiding triangulation error to avoid large cantilevers, and one that separates the two in order to do both.

Basically, I have three elements to choose from: avoiding introducing error, integrating the two parts together, avoiding large (flimsy) cantilevers. I don't know how much of an issue the cantilever is but looking at the design in OpenSCAD doesn't give me the warm fuzzies. I can have two of the three in any given design the way I see it, it's just a matter of which two do I choose.

The STL of the rough shoulder shape can be found in the link below, it's in the orientation that it would be when put on the machine. It isn't fully finished (there should be another hole for a bolt above the one with a nut trap for the actual arm itself to attach to, the one with the nut trap is for an idler pulley that redirects the string) because I keep jumping between designs but hopefully it gives people an idea of what it'll look like. The big gap that creates the yoke like thing is where the drive and anchoring pulley go, so there is a reason it's so wide.

[www.thingiverse.com]

Edited 1 time(s). Last edit at 11/05/2017 04:00PM by Trakyan.

I'll only add a comment about integration vs module designs from the first post. for proto typing or development where parts change, bolt sizes change, and placement of components in design is not 100% known, a modular approach is best. more parts means less time and waste when some part of the design is changing. I think this is why reprap designs seem to be in many pieces. it allows users to make mods easier, and feedback goes into the loop and best designs are modified into the end product. to me it only makes sense to create a product with integrated design and few parts if it is an older modeled part, or a common design that few changes will be pursued.

generally all around there is benefit to fewer types of fasteners. such as bulk cost advantage, lower inventory and tooling requirements. also easier for a new user to put product together.

Edited 1 time(s). Last edit at 11/07/2017 12:33AM by jamesdanielv.