Things I Learned Outsourcing My Parts (feat. @Xometry )

[Music] hello guys winston here pretty much from the day i got my pocket nc and with almost every single five axis project i've done i've been coming up with and using custom work holding on the v210 partly because there aren't many options out of the box you have a vice and an er40 call it but also because it's both a fundamental necessity and a way to dramatically improve the user experience the ability to securely fasten materials to a cnc's table is kind of underrated and it's as important as say being able to operate a cnc machine or generate tool paths you can't machine what you can't hold and historically at least my ambitions have always exceeded what the stock pocket nc could accommodate but that's not all creative workholding can enable it can be a force multiplier allowing you to machine multiple parts at a time as a hobby machinist and from a productivity standpoint that's probably the next best thing to having an automatic tool changer so for a long time now i've wanted to concoct my own ecosystem of work holding accessories for the pocket nc having watched folks like edcramer and deviant clockwork imagine and fabricate cnc accessories to meet their very specific needs i was honestly a little bit jealous their projects seemed like fun design challenges with very practical payoffs but ever since i dipped my toes into r d at carbide 3d and became a dad to a pair of huskies it was difficult for me to set aside enough time to see this idea through so when zometry reached out to me about collaborating on a project i thought this would be a perfect way to team up i'll go into more detail about the services they offer but long story short they took care of the machining allowing me to focus on designing and testing my parts and in this video i'll go through a lot of the things i learned through that process and what that experience was like the initial concept for my work holding ecosystem started off as a round plate that would bolt directly to the b-axis rotary table with counter-bored holes for low-profile m4 screws with a lot of my previous pocket nc projects this was how i did my work holding so it seemed like a fitting starting point when i was starting this design i wasn't sure if i wanted to bolt accessories on through the base plate or into the base plate in the end i decided that bolting through the base plate would be better than bolting into the base plate this makes it easier to design chunky accessories like a tombstone because you don't need to worry about how to stick bolts through it to reach the base plate to give that base plate some extra meat i extruded a boss on the base plate through the main bore in the rotary table installation and removal of accessories could be done from the bottom of the rotary table and ensuring consistent alignment of accessories would be done with holes for indexing pins recalling my prior experience maxing out the pocket nc's work envelope i decided to shave off two sides of that mounting plate this would allow the spindle to get closer to a work piece with short end mills as one of the major clearance limitations on the pocket nc is how close the collet nut can get to the bottom of your work piece and your fixture and only after modeling in all the key features of the base plate did i bother to sprinkle in chamfers this is one of those best practices for cam that i've figured out over the years because projected faces change when you already have a chamfer modeled in if you ever need to modify a feature you can end up breaking sketches and subsequent features same with fillets always make the part functional before you make it pretty from this foundation i can start planning out accessories the easiest one to conceive of is a tombstone not unlike the tombstone that my dfx podcast co-host eddie made but that wasn't where i would have the most fun the workholding concept that would actually drive the constraints of my design was a two-station vertical vice the idea was loosely based on the design of the pocket nc vice which to its credit is actually quite clever and efficient using modest pieces of aluminum bar stock some basic machining and a couple bits of hardware it allows you to hold wood and plastic stock pretty well the limitations of the pocket nc vise however are in its raw work holding strength if you really crank down on the vice there is an unbalanced torque created by a reaction force out of line with the clamping screw that tries to rip the dowel pins out of the aluminum body of the vise when you pair the modest power of the v210 spindle and common stock materials this vice is sufficient for probably about ninety percent of what most people will use it for but that last ten percent of use cases is where i have the most fun so if i could channel the mechanical simplicity of this vice but increase its strength and utility i would be golden the easiest way to counteract a torque is to create an equal and opposite torque conceptually if you had two pocket nc vices back to back and tighten them down equally there would be no problem this is effectively what i tried to do but with only one moving part and the vice mounted vertically i started this vise by creating a fixed block at the bottom and a floating block on top the floating block rides on six millimeter dowel pins and is clamped down by an m8 flat head screw jaws could be attached via m5 screws and aligned by eighth inch dowel pins now i'm going to pause right here because i know both the armchair engineers out there and the real engineers out there are probably thinking there are some questionable elements in the design and quite frankly i agree but sometimes you just need a physical prototype in hand to really understand what's going on and in this case a low precision 3d printed prototype wasn't going to be good enough for me to gauge if the mechanics of this vice design were valid deeply flawed or somewhere in between i needed to be able to put enough force into an aluminum prototype to see if this vise would bind up or check if my fasteners were sized appropriately for the forces generated during machining but for the past 18 months or so i've just been too backlogged with other projects to give this project the time it deserved so when zamatry reached out to me and offered to collaborate on a project it dramatically accelerated my timeline for things could i have made these parts sure but the majority of these parts need at least a fourth axis to do right which would mean a lot of custom fixturing and machining time on the pocket nc zometry let me focus on the idea while they handled the manufacturing this being the first time i've outsourced a significant portion of a project i learned a lot which is to say some mistakes were made but i'll go over that in a little bit here's how my prototyping experience went using zometry's site i uploaded step files for my vice zometry offers a lot of different manufacturing processes in this case i wanted the parts made via cnc milling out of 6061 aluminum and with tolerances of five thou which is the default i also had to upload a drawing providing details on the tapped holes since there were threads in these parts throughout this process i could see how different variables affected the price in real time and i don't just mean with the various drop downs and options if i caught a mistake in my parts i could make changes to the 3d models and re-upload them and see the price change when i was ready to commit to a design i could just click to submit the order and know that my parts would arrive in the given time frame i don't know if any of you guys have worked with a local shop before but playing phone tag or email tag with a vendor just to get a quote makes you really appreciate this now my experience using zometry was generally positive however anytime you outsource a process you need to be cognizant of your tolerances where things can go wrong and have plans in place for remediation of things that don't go according to plan when i was designing my part i gave extra margin for error and features that would be difficult for me to rework for example the way that the lower jaw of my vice seats in the base plate can accommodate up to a quarter millimeter of error in their diameters these ended up fitting with plenty of slop and for the next prototype or in a production scenario i would probably tighten these nominal dimensions to allow only a 0.15 millimeter gap however the holes from my locating pins were features i knew might give me issues in a production environment if you spec a 0.127 inch hole for a slip fit you buy a 0.127 inch drill bit or reamer but this isn't production it's prototyping and what you think is the right way to do things as a machinist is not always how something will happen in the prototyping stage the prototyping shop isn't going to stop their machines to add a custom drill to their limited tool carousel your holes will most likely get interpolated in this case some of the holes in my prototype parts were technically in spec but undersized for an eighth inch dowel pin greg paulson zometry's director of applications engineering gave me a few tips and said that if i had added drill specific annotations to my drawings like tip angle and better tolerance the holes that i might have been able to force the prototyping shop into drilling versus interpolating this is definitely something i'll do next time but at least holes are much easier to rework than most features i can just use my own drill bits to open up any holes that are undersized i chase the undersize holes in my part with a stubby eighth inch drill bit and just the slightest bit of wiggle of my drill so that my dowel pins would fit nicely the prototyping process isn't only about managing tolerances and expectations though sometimes it's about dealing with your own mistakes or bad assumptions or maybe it's a bad assumption that the prototyping shop made in my engineering career my cad drawing skills have been put against the full gamut of expectations i've made really crappy cad drawings as a student annoyingly tedious drawings working for the navy and now i make what i consider to be adequate though not exceptional drawings for carbide 3d the standards to which i make a drawing generally reflects how closely i'll be working with a vendor at carbide 3d i'm usually in regular communication with the person running my parts so i can quickly address any questions or provide clarification and having spent the past few years in this mindset i could admit that i was not as rigorous in my drawings provided to zometry as i should have been there were for example a series of holes in my vise parts that i assumed the programmer would understand were symmetrical they did not one side of my vice was tapped the other side was not i should have shown a projected view on both sides or specifically called out symmetry in the drawing notes fortunately that is an easy thing for me to remedy the takeaway from here is that you really ought to brush up on drawing skills and not take shortcuts or make assumptions in your drawings when you're using a service like zometry your lines of communication with the shop making your parts is very limited do not allow any ambiguity in how a machinist is going to interpret your parts in the end though i got a couple prototype faces together with a not unreasonable amount of manual rework at this stage some of the shortcomings of my design started to manifest when i first conceptualized the vice i knew that binding would potentially be an issue especially if i was loading stock that wasn't prepared identically if one side of the vice hits stock first it could create a torque across the vise and cause the bearings to lock up with friction there is something in mechanics called the binding ratio take the ratio of the distance of an applied force and the length of a bearing if that ratio is greater than 2 you're likely to get binding or a stick slip condition that's a very simplified guideline and there are multiple factors that could increase or decrease that ratio if for example your vice parts were machined on two different cnc's with slightly different tolerances and one of your dowel pins isn't perfectly coaxial with a bearing it could just be out by a couple tenths that could reduce the binding ratio due to increased friction if there were a bit of slop in the holes for the dowel pins and they didn't set perfectly parallel that would again increase friction and reduce the margins of that critical binding ratio in my design i had vice draws that were located 17.5 millimeters away from the center line of my dowel pins my bronze bearings were 10 millimeters long a ratio of 1.75 was cutting things just a little too close that was my bat here's the complication though when i was shopping on mcmaster for bronze sleeve bearings i made the mistake of buying a couple of the cheapest six millimeter bearings i could find this was a prototype after all but those bearings were nine millimeters in diameter all of the longer bearings on mcmaster that i could have subbed in were 10 millimeters in diameter if i re-machine the holes in my vise i would likely introduce an even greater alignment offset between the dowel pins and the bearings so for the purposes of this prototype i would be stuck with shorter bearings that put me really close to that binding ratio and you can see the effect here another thing that someone online commented is that they thought the countersunk flathead screw was a bad choice a flathead screw isn't exactly a high precision piece of hardware so there's a chance that the countersunk taper of the screw head might have some eccentricity to it and apply lateral forces to the moving jaw but these forces would not really be out of line with the bearings and are not a consideration with respect to the binding ratio so it turned out to not really matter regardless of all those factors i did get a prototype vice moving pretty smoothly the only thing that was left to get this thing operational was to install some jaws and instead of having zombray do that i thought it would be a good exercise for a nomad there's nothing too special here just using double sided tape for the first stop with an onion skin so i didn't get any tape gumming up my cutter then making a quick and dirty fixture so i could profile the edges and clean up that onion skin and add some groups to one side of the vice jaws this grooving operation is just a trace tool path using a thread mill and some radial and axial offsetting in these settings and i figured that while i was using the nomad i might as well prepare some bite-sized pieces of stock for the pocket and seat you'll recall from before about the discussion about the binding ratio this vice works best when the stock dimensions are within a couple thou of each other [Music] when it came time for final integration things actually went pretty well the way i envisioned the mounting plate sitting partly inside of the b table worked out pretty well and the way i bolted through it left open the possibility of being able to load a vise offline and swap it onto my machine when ready with normal benchtop cnc tolerances the stock blocks that i made on the nomad could very easily be clamped in my two-station vise without much risk of binding so all in all i think this prototyping experiment did produce something that was actually useful and i'll likely be using this vice or a derivative of it in the future but although this vice prototype ended up being quite functional and practical and the experience of having it manufactured taught me a lot about the importance of tolerancing and having good drawings there are elements of this design that were not fully validated the thing about designing a product when you have knowledge of machining is that you'll likely end up creating constraints that optimize for manufacturability in a very specific way earlier i said that the way a prototype shop will make your parts can differ from how you would actually make your parts this doesn't just apply to whether a hole should be interpolated or drilled it can extend to the most fundamental of assumptions how a part will be made and what kind of machine will make it when i was creating a vice in the back of my head i was considering how someone could produce these components in a cost effective way the obvious way to make something like this is in a fourth axis if you got real fancy you could even tab it off or maybe avoid the need for a secondary operation to lop off any remaining stock but even so this is not really an efficient way of making small relatively simple parts like this you would have a cnc operator babysitting a milling machine swapping in a block of material every few minutes as someone who only runs milling machines one fact i don't appreciate enough is how miltern can be a significant force multiplier if you had a lathe with a bar feeder live tooling and a sub spindle you could make hundreds of identical parts lights out this is how carbide 3d's tease nuts are made for example if you optimize the part to fit within a certain diameter bar you can make tens of thousands with no human intervention so in my head even though i had no intention of mass producing this design i designed the vice in such a way that a properly equipped shop could automate its production with a sufficiently large lathe symmetry provides competitive prototyping rates but those prices are paid to shops that can't amortize setup costs over hundreds or thousands of parts until you take ownership of the production process the true pricing efficiencies and speed you think you can achieve will remain theoretical side note on pricing if you find something inexplicable in zometry's pricing you can request a human review sometimes there are features in your part that their machine learning algorithm just deals with in unusual ways my tombstone attachment provided a good example of this for weight reduction purposes i had a three-quarter inch hole board through the bottom of the tombstone using the zometry plug-in for fusion 360 i tested several permutations of the tombstone and noticed that the depth of the central hole would significantly change the estimated price of that block and that's because zometry's ai pricing algorithm was looking at the part in a vacuum thinking on average a part that looks like this would cost x it doesn't necessarily think about your part in specific or relative terms or on a per operation basis if you are thinking about making this part on a lathe you're just shoving a tool down the center line of the part drilling a hole a few millimeters deeper should add mere seconds to the overall cycle time on a lathe and yet here it drove up the cost significantly in this case the pricing algorithm wasn't considering the tombstone to be a lathe part he was thinking it was either a three setup fourth axis part or a two setup fifth axis part and again you can never assume a prototyping shop has the perfect drill bit for a job that means you price it as though a three quarter inch hole needs to be interpolated so as the ratio of depth to diameter increases so does price i don't think 14 dollars to get another 20 millimeters out of that bore was intuitive which is why i reached out to greg at zometry and he did some digging into how my step file was interpreted by their pricing engine in a scenario like this i would probably ask for human eyeballs to take a look at my part or if i was making a lot of similar parts with only a few differences but the price was changing dramatically i would also try to get a human in the loop but this all goes back to the fact that prototyping is just a very different beast than production and as someone who's used to macgyvering solutions to problems that weren't worth my time having someone else make my parts was an educational experience for me so to close out this video which is probably becoming a rambling narrative at this point let me summarize the key things i learned number one assume worst case on your tolerances if your parts can be off by plus or minus five thou any mechanical interfaces need to allow for at least twice that amount of slop or misalignment you might end up trying to jam a cylinder that's five thou oversized into a hole that's five thou undersized you can tighten your tolerances later once you go into production two drawings aren't an inconvenience they are gospel assume that a programmer or machinist will interpret your uploaded step file as creatively as possible to make their lives easier annotate hole diameters minimum threaded lengths whether a hole is blind or through add thorough notes in your drawings add specific tolerances where necessary for inspection purposes you have one chance to communicate the critical aspects of a part or face tedious rework three understand that while you will end up with a part you won't end up with a process that's something you will need to develop for yourself when you're ready for production if you just need a one-off part that's no big deal if you're looking to start a business venture be aware of all the work that will still lay ahead of you after the prototyping phase if you're not careful you'll end up like john grimsmo these lessons about prototyping are all in addition to everything i've learned about developing these work holding concepts overall i'm really pleased with how this design and manufacturing exercise turned out i think there's an element of something genuinely useful in this modular workholding experiment but i still got my butt handed to me on some pretty fundamental engineering concepts like the binding ratio i need to get device draws much closer to the center line of the vise for usability's sake i also need to add some markings or graduations to the vice draws to make it easier to center my stock in it by eye and it would also be nice to have a narrower version so i can get tools closer to the stock when coming at it from the positive and negative x-axis directions these are good reasons that keep me motivated to keep trying and learning new things though and i look forward to further refining this design i want to thank zometry for getting my parts made for this project and bringing my slightly crazy idea to life zometry offers a wide range of cutting-edge manufacturing processes at competitive prices and it's seriously easy to get your parts quoted through either the website or their add-ins for fusion inventor and solidworks i also want to give a shout out to greg who was able to share some of the backend data to give me some insight into how my parts were being analyzed by zamatry's quoting algorithm and what variables were driving price and machine ability greg and i also had a very enjoyable chat about the story and technologies behind zometry on my podcast the digital fabrication experiment i highly recommend giving it a listen if you want to know more about how their platform works and how jobs are distributed among a network of shops all over the world i want to thank you all very much for watching and i'll be back in 2022 with more cnc content and diy nonsense [Music] you

Channel: Winston Moy

Views: 2,314

Keywords: Shapeoko, CNC, Workshop, DIY, Maker, Carbide Motion, Carbide 3D, Milling, Projects, How To, Do It Yourself (Hobby), Woodworking, Carbide Create, Fusion360, Shapeoko 3, CAM, CAD, Tutorial, Workflow, 3D, Nomad 883, Nomad CNC, PocketNC, Manufacturing, Xometry

Id: LQP3AunnYQQ

Length: 21min 58sec (1318 seconds)

Published: Fri Dec 17 2021