|
The parametric CAD space sucks so much rear end right now. SolidWorks is good and powerful, but it's ridiculously expensive and its licenses are hostile to small scale or non-commercial users, and its CAM package sucks balls. I miss HSMWorks every day. Fusion has a great CAM package and it's free, and its modeling feature set is pretty decent, but it chugs worse than SolidWorks and it's got that insanely stupid cloud saving thing and AutoDesk is rear end in a top hat gradually clawing back features with DLC. FreeCAD, being developed by open source spergs, still can't figure out a basic property of parametric geometry that every other CAD program has had solved for years, meaning your model breaks all the time. OnShape only runs in a web browser. Inventor is just SolidWorks made by AutoDesk -- why? Creo, NX, SolidEdge, etc -- who? The only CAD software company today that I think genuinely has their users in mind, and cares about making a Good Product instead of extracting the most money and data for the least effort, is McNeel. But it's freeform surfacing, not a parametric kernel, and while Grasshopper can do some extremely powerful things it's still not a replacement for SolidWorks. If Rhino had a true parametric kernel mode I don't think I'd ever need to use anything else. Sigh
|
#
?
Aug 8, 2023 17:14
|
|
|
|
| # ? Jun 3, 2024 11:58 |
|
|
Sagebrush posted:FreeCAD, being developed by open source spergs, still can't figure out a basic property of parametric geometry that every other CAD program has had solved for years, meaning your model breaks all the time. Hey, the good news is that this is actually, really, no-poo poo going to be fixed soon. Not “go download this guy’s crashy fork of an old version” “””fixed”””, in the next non-bugfix release of mainline FreeCAD, which will be 1.0 because of it. Bad news is that I don’t think they’ve given an estimate for that release. But they don’t intend to work on any other significant features until it’s done, is my understanding. BTW, they just released 0.21, as a way to ship all the recent improvements while the dev team crunches out topo naming. Go get it if you’re cheap enough or enough of a grognard to use FreeCAD.
|
|
#
?
Aug 8, 2023 18:27
|
|
|
lol OnShape had a customer success story for the OceanGate billionaire murder sub https://www.onshape.com/en/resource-center/case-studies/oceangate unfortunately it has been memory holed
|
|
#
?
Aug 8, 2023 18:33
|
|
|
Even dangerous and fatal engineering projects need a reliable CAD system.
|
|
#
?
Aug 8, 2023 18:41
|
|
|
prob shouldn't use it to showcase your FEA tho
|
|
#
?
Aug 8, 2023 18:44
|
|
|
I can't use Solidworks because even at default resolution I can't loving select things properly, I did a class and my instructor was astonished at how much I would screw it up when everyone else was fine. Is there some kinda like "old person mode" to make the little dots of vertices and things just way bigger and more obvious? Or like for the magic the gathering phone game, if there's a stack of cards and you click on one it fans the cards out to make sure you are going for the right one, UI like that would be nice. Also I'm horrified to ask but is there any parametric CAD that can run in VR? Maybe if I could grab poo poo with my hands while streaming from PC to Quest or something it might be more intuitive.
|
|
#
?
Aug 8, 2023 19:00
|
|
|
I somehow doubt the simulation package for Onshape has the data for dodgy, expired carbon fibre. Though the finest software in the world is no match for a billionaire's ego & misguided self-confidence.
|
|
#
?
Aug 8, 2023 21:56
|
|
|
Just bump the safety factor to 1.5 to account for the degradation.
|
|
#
?
Aug 8, 2023 22:24
|
|
|
FEA is only as good as your inputs and interpretation of the results. That's a PICNIC (problem in chair, not in computer)
|
|
#
?
Aug 8, 2023 22:58
|
|
|
Zero VGS posted:I can't use Solidworks because even at default resolution I can't loving select things properly, I did a class and my instructor was astonished at how much I would screw it up when everyone else was fine. Is there some kinda like "old person mode" to make the little dots of vertices and things just way bigger and more obvious? Or like for the magic the gathering phone game, if there's a stack of cards and you click on one it fans the cards out to make sure you are going for the right one, UI like that would be nice. I'm pretty sure you can change your settings for graphic size. It's not a resolution thing. Like, the points can be sized up by pixel or millimeter. RE VR: Ugh..... CATIA does it pretty well, but CATIA is... Well it's a Dassault product.
|
|
#
?
Aug 8, 2023 23:01
|
|
|
NX got a VR button a few versions ago but I've been scared to ever touch it
|
|
#
?
Aug 8, 2023 23:06
|
|
|
Sagebrush posted:, SolidEdge, My coworker was just praising the free SolidEdge today. Has anybody here used it? I use SolidWorks professionally (daily) and OnShape at home (quarterly lol). I think being in a browser is an advantage as I do most of my onshape work on a chromebook on the couch or in the garage. I think I'll try solidedge when i get some time and maybe that freecad 1.0 when it comes out.
|
|
#
?
Aug 9, 2023 01:29
|
|
|
Spaghett posted:FEA is only as good as your inputs and interpretation of the results. That's a PICNIC (problem in chair, not in computer) Yeah I don't think this can be emphasized enough when it comes to FEA, the hardest part is setting up your model accurately and knowing the expected result well enough to interpret what the simulation tells you and correlate it with reality. I'm not a simulation engineer but I work with a very talented group of them and they're all pretty clear when we're discussing results that they should be used in a comparative manner between simulations and not as a predictor of absolute performance. Most of their effort goes into correlating simulation results with our mechanical lab testing results to improve material model and understand the relative performance of the simulated parts vs our real physical parts. Ryanrs if you're interested in learning FEA to learn FEA thats awesome but to be frank a static loading of a simple beam section like your drill press wheel problem is probably going to be easier to do pen and paper than set up a simulation, and you'd want to have an approximate understanding of the stresses from a simple analysis to sanity check your simulation results anyway. Not trying to shoot you down just giving you some info that might save you big headaches down the line when your thing that works in simulation doesnt work at all irl. Sagebrush posted:
Lol NX is a massive player in the giant corp huge project CAD space and is ludicrously powerful if you need to combine parametric modeling, class A surfacing, simulation, and CAM all in a single package and need it to just work. Doesn't have as much visibility as some things like solidworks but seems to be really common in aerospace and automotive in my experience working with engineers in those fields. meowmeowmeowmeow fucked around with this message at 03:55 on Aug 9, 2023 |
|
#
?
Aug 9, 2023 03:52
|
|
|
The one time I used FEA professionally, I had a known good part and made it lighter and stiffer by optimizing ribs and material removal. After trying several things, the broad conclusion was that adding material around the outer perimeter gave the most stiffness/kg so that's where I put the metal. (The actual design was somewhat complex, so it wasn't quite so obviously simple at the outset.) I'm thinking of applying FEA to the drill press problem because it's a very simple problem, so good for learning. But the software cost makes that idea stupid, so never mind. e: I think most Apple stuff is done in NX.
|
|
#
?
Aug 9, 2023 07:06
|
|
|
Thrilling update: I hit the VR button in NX and it told me to gently caress off because we don't have the license for it
|
|
#
?
Aug 9, 2023 12:59
|
|
|
Ah cool, sorry for assuming you hadn't used it before.
|
|
#
?
Aug 9, 2023 16:35
|
|
|
I'm an FEA guy so if you want to chat, hit me up in the PMs and we can arrange a call. I'm always game to have fun with our lil elements
|
|
#
?
Aug 9, 2023 19:29
|
|
|
Not sure about mechanical stuff, but when we do complex numerical analyses for dirt and rocks it is important that you know about what the answer should be before you start. Otherwise you may as well be running a random number generator. Just because you can make cool-looking contour plots from the output doesn't mean you have a good answer.
|
|
#
?
Aug 9, 2023 20:20
|
|
|
ryanrs posted:lol OnShape had a customer success story for the OceanGate billionaire murder sub  Published before Onshape even had simulation so I’m gonna guess that they did….none, even if they could get it right. Re: NX, it’s so good that the Parasolid kernel that drives it is what SOLIDWORKS and Onshape are based on! There is a FREE version for students that, sadly, does not include CAM posting. You can program, but not post. There’s lots of tutorials but I haven’t dropped in for more than a sketch or two. https://blogs.sw.siemens.com/nx-design/nx-student-edition-free-download/ Seems like the biggest cause for folks to not use NX is because it is intensely restrictive. If you want to share files with Onshape and SOLIDWORKS users, you can use the .x_t file export, it’ll read perfectly in those others (and Plasticity and Shapr3D and…). Zero VGS posted:I can't use Solidworks because even at default resolution I can't loving select things properly, I did a class and my instructor was astonished at how much I would screw it up when everyone else was fine. Is there some kinda like "old person mode" to make the little dots of vertices and things just way bigger and more obvious? Or like for the magic the gathering phone game, if there's a stack of cards and you click on one it fans the cards out to make sure you are going for the right one, UI like that would be nice. SOLIDWORKS is unbelievably fucky with resolution and display scaling. Are you running a high DPI screen (i.e. beyond 1080p)? Check if your display scaling is more than 100%, and maybe check this article out: https://www.goengineer.com/blog/solidworks-display-scaling-solutions And for VR, maybe Gravity Sketch? As mentioned, there are Dassault options that are 
|
|
#
?
Aug 10, 2023 00:16
|
|
|
ryanrs posted:Next step, what's the design criteria? It turns out scenario (1) isn't quite so simple. Besides CoG and balance points, there are tradeoffs re. floor clearance of the back edge and how far back you need to tilt the drill press back to move it (even if well balanced, if you need to tip it back 45 deg, it gets hard to maneuver just because the bounding box is wider). Simplified diagram of the drill press:  The drill press balances on its back edge at 78.0 degrees. The drill press balances at 90 deg on a fulcrum 224 mm from the back edge. This puts the CoG 224 mm back, and 1054 mm above the floor. In the diagram, the drill press is balanced on its back edge, with the caster just touching the ground. When tipped back a further 8 deg, the CoG becomes balanced over the caster, and the drill press ground clearance is 20.46 mm. This is the unstable configuration described in (2). Normal travel position will 25-30 deg off vertical, with 50+ mm of floor clearance for running over cords, etc. A more upright travel position would give better maneuvering, but requires the caster to be placed closer to the back edge of the base to prevent (2) instability, which reduces floor clearance. A possible design: 
|
|
#
?
Aug 10, 2023 01:18
|
|
|
But how does it handle on the freeway? Can it overtake a two trailer semi at 80+? What if you put it on a sick pump track
|
|
#
?
Aug 10, 2023 01:36
|
|
|
FEA Questions  The drill press weighs 150 lbs. The bracket is attached to the cast iron base with two 3/8" bolts. Will the sheet metal around the big bolts yield in use? When rolling over bumps like cords and pebbles? When rolled down the stairs one step at a time? I sketched out the design in 10ga steel (3.43 mm). Should it be thicker or thinner? Maybe more compliant ribs will reduce peak stress on the bolt area during the stair test? The stair test may not be even be feasible with just those two bolts. Certainly letting the drill press free-fall off a step will bend or break things. How do I characterize the forces from "kinda gently guiding it down, but without lifting the full weight"? Not that I contemplate rolling my drill press down the stairs. But I think it is a realistic scenario to worry about if you were selling these things to the public. So, how easy or difficult is it to answer these kinds of questions with FEA? Are these even the right questions to be asking?
|
|
#
?
Aug 10, 2023 02:36
|
|
|
I think a lot of simulation suites now come with a pre-baked drop test scenario which would be helpful in this case, but its a dynamic contact analysis and I'm not sure they'd love it being a multi-body simulation and having to worry about all the bolted joints. Characterizing impact forces (this is from vague memories so may not be right) is generally starting from potential energy = mgh to get how much energy the thing has from falling, then doing mgh= 1/2*m*v^2 : v = sqrt(2gh) to get velocity at impact. If its a perfectly elastic collision (ball bearing on hardened steel plate) it'll rebound with a velocity equal to its initial velocity, if its a in-elastic collision (ball into sand) it wont bounce at all. You estimate this (can usually go with 0.7 - 1 and be pleasantly surprised when its less) and now you have your total change in velocity, which you can throw back in the kinetic energy equation to get impact energy. For the reaction force for the impact, work = force*distance, so now estimate the amount of distance the thing will travel during the impact (basically how much it'll flex in impact, prob not much), then you can estimate average force for the impact and then maybe double it for peak force? This peak force can then go into a beam bending equation to look at stress and how it compares to your material stress limits, but won't help understand distortion of the plate around the bolts. Thinking about it as I write all this out you could probably just look at beam bending strain energy vs impact energy and get the amount of deflection required to store your impact energy and see if that exceeds your yield stress or fatigue limit but your beam is so short a lot of the assumptions in beam bending models may not apply. Personally I'd put a big old weld washer under the bolt head, increase the vertical rib thickness to 1/4", and maybe put a rib along the outside edge to keep it from taco-ing on impact and call it good? If you were interested in selling them, I'd make a test article, a dummy load similar in weight and position to your drill press so you're not banging it up in testing and go beat the poo poo out of it and see what happens.  Hope I didn't over explain things you already know with this, but in general these are (I think) the right questions to have but while this is something FEA could help with optimizing a design you're still going to have to build and beat on a test article to know where you're at.
|
|
#
?
Aug 10, 2023 05:13
|
|
|
No, I don't know any of this! I like the idea of shoehorning the drill press into Solidwork's iphone drop simulator, ha ha. 'Anti-taco Ribs' sounds like a club for barbecue-supremacists. But I can add one almost for free by adding another 90 deg bend and a 20 mm flange. Poking at the bracket in isolation using Solidworks SimulationXpress (crippled FEA), it looks like the vert ribs see very little stress, and it's all concentrated where the rib meets the top of the tube. But I doubt SimulationXpress knows about buckling. Current thicknesses: flange 0.135" ribs 0.119" tube 1.5" OD x 0.120" weight: 4.0 lbs (excl casters, bolts) Are you scaling everything up to 0.250" or just the ribs? And if we're going with the heavy duty option, we can change the mounting to a chunk of steel sandwiched between the column and the base.  Just make the entire thing out of 0.25" AR400, no ribs, and let it flex to eat the impacts, heh. meowmeowmeowmeow posted:If you were interested in selling them, I'd make a test article, a dummy load similar in weight and position to your drill press so you're not banging it up in testing and go beat the poo poo out of it and see what happens. A drill press crash dummy! Alas, this project is not a great candidate for destructive testing. I would like to keep the cost of the casters less than the cost of the entire drill press ($360 on craigslist). And if the bracket does get bent up, it won't even affect the operation of the drill press. However, if I beef up the bracket a bunch and break the cast iron base, that will be quite annoying. ryanrs fucked around with this message at 09:34 on Aug 10, 2023 |
|
#
?
Aug 10, 2023 05:58
|
|
|
Instead of making this thing stupidly strong, maybe I should give it a suspension to absorb shocks? What if I added polyurethane bushings? 
ryanrs fucked around with this message at 03:54 on Aug 11, 2023 |
|
#
?
Aug 11, 2023 03:42
|
|
|
Yeah I could see that working, add deformation to reduce peak forces the bracket sees. To be honest I'm a washed engineer who does product development/project management now so while I remember the basics from school I'm well out of practice so take everything with a grain of salt. I'm also not super used to working with metal when designing stuff and its usually aluminum, so my recommendations might be very off. I think if you bump up the ribs and/or increase the height or thickness of the little tube bits you'll be fine. The sim probably knows about buckling but the stress being concentrated at the top of the tube/rib join makes sense. I drew up a simple analysis of how I'd think about the problem:  You've basically got a cantilever beam loaded in transverse shear, so your bending moment will be highest at the fixed side (aka your bolts). With stress from bending being inversely proportional to the moment of area and moment of area being (b*h^3)/12 for a rectangle, the shorter effective height of the rib is reducing the ability for the design to carry stress in the highest stressed location of the part. Throw in some stress concentration from the sharp join in the geometry and makes sense you've got a peak right there. Throwing some weld beads into your model and re-analyzing might be interesting to see if anything changes. But generally i think you're gonna be fine and if it bends a bit it'll bend a bit and it wont be catastrophic like you said. I'd stay bolted through the base, two 3/8" bolts are gonna give you like 10k lbs tensile strength easy, the hard part is gonna be spreading the load out from the bolts well but I think you're on the right track.
|
|
#
?
Aug 11, 2023 05:09
|
|
|
The ribs are fine; the problem is your contraption is just one giant lever that's going to shred the bolts. You need to extend the plate forward more, and probably the ribs along with it. Edit: 0.25 plate. They can bend the side flanges to minimize welding. Force on the bolts will be only slightly higher than the weight of whole thing. 
TheLastManStanding fucked around with this message at 06:09 on Aug 11, 2023 |
|
#
?
Aug 11, 2023 05:28
|
|
|
Yeah that's a great point
|
|
#
?
Aug 11, 2023 05:57
|
|
|
The bracket doesn't sit flat against the base. There are two 32 mm dia raised pads with bolt holes. It's not great. But the mechanical advantage on those bolts is half the pad diameter over the length of the lever. 16 mm / 206 mm = 13x (is this an ok first-order approximation?) Clamp load for a single 3/8" grade 5 bolt is 4950 lbs. Divide by 13 = 380 lbs. Figures for various grade 5 and grade 8 bolts: 3/8" gr5: 380 lbs; gr8: 536 7/16" 522; 738 1/2" 698; 980 Try to drop the drill press on both wheels at the same time, so the load is spread across both bolts, ha ha. The holes in the cast iron base are 12.1 mm. So a 7/16 bolt will fit, and a 1/2" with a little drilling. The problem is on the bottom of the casting. I don't have a pic, but there's thin ribs near the holes. A 3/8" clip washer will easily fit. A 7/16" clip washer might not, or I might need to fabricate a very clipped washer. For 1/2", I don't think the nut will clear the casting rib. TheLastManStanding, you make a good case for trying to shoehorn a 7/16" bolt in there. The strength increase over 3/8 is quite a lot. e: re. thickness, the skid plates on my van are 3/16". 1/4" is very thick steel. Peak stresses may call for something that thick near the bolts, but not for the rest of the bracket. The fat weld washer idea sounded good. ryanrs fucked around with this message at 07:05 on Aug 11, 2023 |
|
#
?
Aug 11, 2023 06:21
|
|
|
Unless you're taking your drill press on the pump track, I don't think you have any risk of tacoing this thing. The load path goes directly from the bolt hole to the caster. Ok the load goes to one side of the caster, whatever. The far side of the caster is supported by the stringer that partly boxes in the weird ~120 degree angle you've got going on there that makes it easier to tilt back and adds a significant amount of strength For 150 lb jobby I think you'd be fine with 14ga. If you wanted absolute overkill go to 11ga If the base is truly cast iron (very likely) and you plan on wheeling this thing around more than 100ft a day then I guess the polyurethane bushing might be a good idea over a 20 year service life. While you're at it get bearings with a grease zerk so you can service the bearings on the casters too. And make it out of 304SS in case some of the paint flakes off
|
|
#
?
Aug 11, 2023 07:04
|
|
|
I got fusion 360 to boot and login on my linux laptop using this very dodgy script that runs from the internet, downloads a bunch of binary files and executes them (do not do this) https://github.com/cryinkfly/Autodesk-Fusion-360-for-Linux/wiki/Documentation#prerequisites Not going to link to the exact commands (you can click the link footgun this yourself) but it does appear to work. I did a thing and now I'm reinstalling, and it's slow, but I have a bunch of stuff locked up in fusion360 I need to port over to onshape and my windows laptop went missing during the move
|
|
#
?
Aug 11, 2023 07:08
|
|
|
I don't weld stainless, but I do know a local metal plating shop. It'll cost $140 (minimum batch fee) to zinc plate it, but it'll look amazing. They plated my recovery points in black zinc. I am liking the light gauge steel + polyurethane bushings design. And if 11 ga steel bends, maybe that's a good thing if a 1/4" bracket would have broken the casting. And I can test the light gauge bracket a bit before getting it plated. BTW, SimulationXpress agrees with Hadlock that the big blank expanse of steel between the two casters does basically nothing (cutting it out causes a small increase in max deflection). Adding a horizontal anti-taco rib likewise does nothing. Simulation was done with all load on one caster only. ryanrs fucked around with this message at 07:20 on Aug 11, 2023 |
|
#
?
Aug 11, 2023 07:16
|
|
|
I have a 20" x ~8' thing I'd be interested in getting plated or other industrial coating if you want to split the $140 fee for a single batch of black stuff although my surface area might be enough to push them over the minimum Your shop deal looks pretty sweet I need to find a place like that
|
|
#
?
Aug 11, 2023 07:31
|
|
|
ryanrs posted:The bracket doesn't sit flat against the base. There are two 32 mm dia raised pads with bolt holes. It's not great. No I don't think this is correct I have no experience building stuff like this besides a half constructed overbuilt go kart and I would just throw two grade 5 5/16" and never think about it again. Something went very wrong here. Paperclips would be too weak but you could probably epoxy some AA batteries in there, and they would hold Or maybe I'm way way off but I think you said the whole device weighs ~150 lbs. The 24 kilonewton rock climbing carabiner by my foot is aluminum and about 4-5mm in diameter and rated for human suspension in harsh outdoor conditions
|
|
#
?
Aug 11, 2023 07:40
|
|
|
20 x 8 ft sounds more like a job for hot dip galvanizing than electroplating. What is this thing? Static forces and normal rolling around the floor usage won't break 5/16" bolts. But what about the stair test? That shock load, even if you ease it down the step carefully, could be really high. These bushings should help, though. e: side view  BTW Thanks everyone for all the ideas. We are engineering the hell out of these casters, and I love it! We may not be able to accurately predict how strong this thing will be, but I think we've identified the weak spots with pretty high confidence. This led to the bushing idea, which mitigates worst-case shock loads for very little extra work. Basically all I have to do is make sure my tube ID matches a honda civic control arm, ha ha. ryanrs fucked around with this message at 09:10 on Aug 11, 2023 |
|
#
?
Aug 11, 2023 08:22
|
|
|
If you're going up and down stairs then just switch to 8" pneumatic tires. They will soak up the shock neccessary. Else go find some mountain bike coil-over shocks and give it a fully articulated suspension. Right now that load is going to go into the cast iron and... well depends on the quality of the iron but nothing good will come of it You want the fastener to fail before the cast iron cracks. At this point the caster bracket exceeds the shock loading of brittle cast iron's ability to absorb it safely Ideally you're using trailer axle bearings, so you can just add a trailer hitch to it, and tow it behind your car down the highway at 75mph, rather than having to load it into a big heavy truck edit: like this 
Hadlock fucked around with this message at 10:14 on Aug 11, 2023 |
|
#
?
Aug 11, 2023 10:10
|
|
|
Hadlock posted:If you're going up and down stairs then just switch to 8" pneumatic tires. They will soak up the shock neccessary. Else go find some mountain bike coil-over shocks and give it a fully articulated suspension. Right now that load is going to go into the cast iron and... well depends on the quality of the iron but nothing good will come of it Whenever I see that on the highway I switch lanes.
|
|
#
?
Aug 11, 2023 15:41
|
|
|
I'm just here to complain. We have our CATIA poo poo on the cloud and I'm in a place with spotty WiFi. Mixed with that is the unreliability of our home network where our CATIA stuff is hosted locally. Anyway, this means that every time there's a single dropped packet, poo poo goes loving bananas and renders CATIA useless. Like, bricked, stuck on a loading icon like it's an '08 MacBook beachball. I hate working in the cloud and I hate this software even more.
|
|
#
?
Aug 11, 2023 16:28
|
|
|
Spaghett posted:I'm just here to complain. Lmao is there no offline mode? SOLIDWORKS people revolted over it and they added one last year for the 3DX stuff. Dassault also have, hands down, the absolutely most garbage cloud connected CAD system. It’s not only THAT sensitive to dropped packets but it won’t let you loving do anything until it handshakes and it takes FOREVER. Fusion at least will give you a pop up and reconnect you. Dassault’s own internal network is hot loving garbage. They cap downloads at like 5MB/s and downloading whole manufacturing cells takes ages, let alone downloading whole factories. Just consistently punching themselves in the dick tripping over problems other people solved like twenty years ago. You are not alone in your misery ❤️ Assault France and dethrone Dassault.
|
|
#
?
Aug 11, 2023 16:58
|
|
|
|
| # ? Jun 3, 2024 11:58 |
|
|
DDS showed us some cool stuff they have in the pipeline for ABAQUS but its all being run through 3DX or some other web interface which was a bummer to hear. And yeah I think 11ga or w/e would be fine but the concern about lever forces on the bolts is a big one imo, and I think your approximation of the lever forces is generally correct. Some of the force would be in shear once the thing is tilted back but there'll be some levering on the bolt in tension.
|
|
#
?
Aug 11, 2023 17:37
|
|