You are pretty much confirming my opinion that you know more about these than the guy who designed the one you are trying to build.
Some of my first projects were a box to house a PCB, A plastic foot to a stand mixer. and a rear cap for an SLR lens. And other things that don't matter if the tolerance is off by 0.5mm Then you work up to more complex stuff. The key is to develop CAD skills or you are at the mercy of random people who post to Thingverse and the quality there is all over the map. The key to printing machine parts is to never depend on precision. I'll place a screw adjustment on a shaft to shaft distance and then I can adjust the tightness of the gears after assembly. Precision holes are printed undersize then bored or reamed. One way is to treat printed parts as if they were raw castings. I'm starting another project soon, it will be an open-source robot that is modular, lego-like with wheels about 200mm diameter. We are designing this collaboratively with the idea that it by easy to print, even on low-end out of spec printers. We'll use 8mm drill rod axels and skate bearings. Long terms is a four-leg robot that is sized large enough to do normal stairs and have an eye-level high enough off the floor to navigate a normal house. Perhaps like a small dog. The key is being buildable by amateurs with limited tools and experience. The wheeled version goes first. The legged version needs 12 reduction drives. You can see my interest in reading about your problems with this. On Sat, Aug 22, 2020 at 9:26 PM Gene Heskett <ghesk...@shentel.net> wrote: > On Saturday 22 August 2020 19:25:25 Chris Albertson wrote: > > > Then go into the CAD software and fill the hex pocket and enlarge the > > holes to accept the brass inserts or make models threads. > > > > At some point to you have to decide that the person who designed this > > thing only made a first level prototype and more development is > > required. I'm following your story and I would have made that > > determination weeks ago. It is a good fist try, to prove it might work > > but they break quickly and seem to be very hard to print. My guess > > is that a few more design iterations are needed. To make them easier > > to print and stronger. > > > > I think there might be trade-offs reduction ratio and backlash and > > strength and "printability" > > > > I had a professor once who said you can always tell a new or amateur > > engineer because the stuff he designs only works if you use > > non-standard high precision parts. Better designs don't need any of > > that. Designing things that work and are printable is hard. Easier > > to design stuff in metal with 0.0001 level specs. This designer was > > tinking in metal but using plastic. > > > Tinking, could be thinking or tinkling. I'm in favor of calling it the > latter. > > "gear ratio" in the case of a harmonic drive, is a huge compromise, > tradeoff if you will, between the ratio and the amount of flex imparted > to the flexible member, lower ratios mean taller, wider "teeth" and more > flex. Higher ratios need smaller "teeth" and correspondingly less > flexing. And this one has about the maximum flex the plastic will allow > without rapid fatigue cracking and failure. Go for too small and you > are going to exceed the tolerance of your machine to lay up an accurate > representation of what the optimum design needs. And this design may be > right on the ragged edge. So the exact size of each of the three working > parts must meet the following criteria: > > 1. The max push needs to be far enough out of round that the "teeth" of > the flexgear spline must, by the shrinkage between the rollers, clear > the teeth of the internal spline at the mid-point between the rollers, > otherwise the teeth are forced to snap up and over each other as they're > forced to pass. The 3 rollers on each end of the rotating wave plate, > seem to force tooth contact over a wider section, in turn needing a more > precise pull-in tension to clear this hop. > > 2. The max push needs to drive the teeth to as close to full engagement > as it can, without trapping the teeth between the rollers and the > internal gear. Bottoming of the tooth engagement, trapping the flexible > spline against the internal gear is semi-desirable because that also > corresponds to the zero backlash condition, but that flexes the cup at > the base of the spline, forcing a quick breakoff at the bottom of the > spline. So there must be a teeny bit of backlash in the plastic > version. > > So you chose, by diddling the printers x and y scales, a size of wave > plate that allows those 2 conditions. And the tolerance would appear to > be about .1mm. Next you build the flexgear, again by diddling the > printers scale, such that the built cup, placed over the running wave > gear, can be seen to be flexing about 1.5 times the height of the teeth. > > Then you measure the teeth at the maximum push, and make the internal > ring, about .15mm bigger to preclude the entrapment between the wave > gear and the stationary internal gear, and when you then drop that ring > over the rest of it, you see, on both sides, that they do clear enough > to do this hop-over, without catching and causing a click. If that ring > gear doesn't fit the body's you've already made, then you adjust the > wave bearing size by say .2mm and start over. But you can't go very far > without exceeding the amount of slop in the 4mm bolt holes that put it > all together. So the size tolerance for each of those 3 parts is > around .1mm. > > In plastic? Keep your rabbits foot VERY handy. Oh, and the flexgear and > the output shaft/flange should be a press fit in the 35mm bore of the > main bearing, so in the case of the flexgear, you can only adjust it by > about .1mm undersized at the projection into the bore of that main > bearing where they are bolted, up to what will still be drawn together > by those 6 bolts without making the bearing too far out of round and > lumpy turning. > > In short, this was NOT a good first project for a new printer that was > way the hell out of calibration OOTB. I'm not convinced it was a readily > achieveable project. At $10/hr for my time, I should have bought 3 new > small ones on ebay and been ahead of this game. > > > On Sat, Aug 22, 2020 at 3:42 PM Gene Heskett <ghesk...@shentel.net> > wrote: > > > On Saturday 22 August 2020 16:30:56 Chris Albertson wrote: > > > > The solution is not to use captive hex nuts. > > > > > > Thats what this thing is designed for, if indeed it has pockets for > > > the nuts. The m4's generally have a pocket, but the m3's in the wave > > > bearing carrier are plumb bare. > > > > > > > You can simply print M4 > > > > size threads inside a hole and screw the cap screw into the > > > > threaded hole. This works if you don't go through so many > > > > assembly/disassembly cycles. The best is to use these heat-set > > > > brass thread inserts. Then your screw goes into brass threads. > > > > > > > > In all cases, blue (medium) Locktite works well. > > > > > > And of coarse I have red and green. :( > > > > > > Got the second one running, fairly quietly, but I do need to split > > > it and molycoat the wear areas. 3rd one has a slightly smaller > > > flexgear about 5% built, hopefully it will work with one of the ring > > > gears I have a pile of at various sizes, flow looking good at 100%, > > > even on the brim. Running at 210 and 55 once past the startup layer, > > > I read someplace where strength was enhanced with less cooling under > > > the nozzle, more time to weld itself I guess, so its running that > > > fan at 55%. And I've resliced with some elephants foot comp ( > > > 0.5mm) enabled. Thats why I had to take the first white one to the > > > lathe and shave that off. That was not a problem with the black, but > > > that stuff is more brittle than the white. And I'm outta m4x10 ss > > > cap screws and m4 hex nuts, miss-counted badly. Monsterbolts and my > > > banks MC is my friend. :) > > > > > > > amazon.com/Threaded-Heat-Set-Inserts... > > > > <https://www.amazon.com/Threaded-Heat-Set-Inserts-connecting-injec > > > >tion > > > > /dp/B07BH5X252/ref=sr_1_18?dchild=1&keywords=m4+brass+insert+3d+pr > > > >int&q id=1598128018&s=industrial&sr=1-18> > > > > > > That would need the source .stl's modified, and I don't know enough > > > about freecad to attempt that. So I'm stuck using the designers > > > fasteners. > > > > > > Thanks Chris. > > > > > > Cheers, Gene Heskett > > > -- > > > "There are four boxes to be used in defense of liberty: > > > soap, ballot, jury, and ammo. Please use in that order." > > > -Ed Howdershelt (Author) > > > If we desire respect for the law, we must first make the law > > > respectable. - Louis D. Brandeis > > > Genes Web page <http://geneslinuxbox.net:6309/gene> > > > > > > > > > _______________________________________________ > > > Emc-users mailing list > > > Emc-users@lists.sourceforge.net > > > https://lists.sourceforge.net/lists/listinfo/emc-users > > > Cheers, Gene Heskett > -- > "There are four boxes to be used in defense of liberty: > soap, ballot, jury, and ammo. Please use in that order." > -Ed Howdershelt (Author) > If we desire respect for the law, we must first make the law respectable. > - Louis D. Brandeis > Genes Web page <http://geneslinuxbox.net:6309/gene> > > > _______________________________________________ > Emc-users mailing list > Emc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/emc-users > -- Chris Albertson Redondo Beach, California _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users