|
 Why isn't this creating bend reliefs, and how do I fix it? I think bend reliefs are what they are called. The little slots on each side of the bend, so the adjacent edge doesn't get distorted. e: this is for a little tab with a rivet hole. I am going to bend it with pliers and rivet it to another part. is there an easy way to pattern/replicate this tab in the 10 places I need it on this part? ryanrs fucked around with this message at 21:25 on May 9, 2024 |
#
?
May 9, 2024 21:12
|
|
|
|
| # ? Jun 3, 2024 09:09 |
|
|
Don't pre-cut your little reliefs, make a flange on the entire edge and then use the 'edit flange profile' button to draw the locations you want. You can use the flange position to set how the flange surface sits relative to the face you bent off of. I don't think you can do multiple flanges with one operation, but you should be able to do a feature pattern of some sort to arrange them how you like.
|
|
#
?
May 9, 2024 21:39
|
|
|
thx that fixed it
|
|
#
?
May 9, 2024 21:47
|
|
|
    Mounting box for an Aruba AP-303H wifi AP. This metal box gets attached to the surface of a wooden beam. Then the AP-303H plastic wall mount bracket gets screwed onto the sheet metal. The access point then clips onto the wall bracket. The metal box provides clearance behind the AP for cabling (not necessary when installing on drywall). This project uses no CNC bending and no tapped holes, so it's relatively cheap for a custom metal box. $27 for 1.2mm galvanized steel at Sendcutsend. Assembly will involve hand-bending with pliers and maybe a vice, pop rivets, and self-tapping screws. The skirt has thin slots on the 45 degree bends, but they're not visible on the 3D renders. I add them at the bend lines after the model is flattened. This would be a prime candidate for 3D printing, like Newfatmike's outlet cover plate, except I don't have a 3D printer. I think any commercial 3D printing service will cost more than that for 160 x 86 x 31mm (~thick paperback book). If I needed several, then 3D printing starts looks better because of much less assembly time. But since I only need one unit, I'm going with laser-cut, galvanized steel coil. It's like Simpson Strong-Tie made a mounting bracket for my router! I kept the OEM wall mount bracket because it has a pretty good quick-release system. So it's easy to pop the AP off the bracket and bring it inside. This is going to be installed in an apartment complex parking area.
|
|
#
?
May 10, 2024 06:16
|
|
|
Out of curiousity, why the two part design instead of the face piece with the 4 sides being bent down?
|
|
#
?
May 10, 2024 14:48
|
|
|
Originally, the skirt was going to wrap wrap the flat plate in a way that perfectly matched the rounded corners of the Aruba AP. But then I simplified it a lot, and now it can be build in one piece, I just didn't think of it. Hmm. I maybe I should redesign it.
|
|
#
?
May 10, 2024 16:30
|
|
|
How does an electronics and cad man just not have a 3d printer? Is that allowed?
|
|
#
?
May 10, 2024 17:08
|
|
|
Rexxed posted:How does an electronics and cad man just not have a 3d printer? Is that allowed? When I first looked at it, 3D printing seemed to involve a surprising amount of cost and technical knowledge that was tied to your specific <printer, material> pairing. Yes, there was a lot of broad "3D printing" knowledge, which is cool and good. But there also seemed to be a lot of learning about your particular machine's hot end, or weirdo vendor firmware, or weirdo open source firmware. So instead I pay $30-100/piece to get stuff printed by commercial 3D print farms. This is for functional parts weighing a couple hundred grams, iphone-size or coffee-mug-size using processes like HP MJF or Markforged Onyx. FDM ASA would be fine for a lot of my stuff, but it generally costs more than MJF. I think the powder bed machines might scale better than FDM for high-throughput commercial shops? Pros: - no up-front cost - get to try every material - perfect prints first time Cons: - $30/part realistic minimum - no instant gratification - no iterative design - better be perfect first time
|
|
#
?
May 10, 2024 18:52
|
|
|
The more turnkey printers like the bambu labs are a lot less tuning and futzing than printers used to be, I've been in the same boat as you with regards to owning a printer being a hobby in of itself but that seems to be changing. It'd be nice for the iterative design phase and then going to print suppliers for final parts. I will say MJF can produce very nice structural parts and when you're filling a full print bed its pretty economical for prints, we had one at a previous job and ran it all the time.
|
|
#
?
May 10, 2024 19:05
|
|
|
I noticed that with MJF there doesn't seem to be a quantity discount, even when going from 1 -> 2 parts. They must be turning over their print volumes like clockwork. Why couldn't they have over-invested in capacity? 
|
|
#
?
May 10, 2024 19:34
|
|
|
I had very similar concerns with 3D printing but I've found my little Prusa Mini and prusa slicer makes it pretty much plug and play. There's a little bit to think about with layer height/shell thickness/infill but mostly I've found it dead simple for the little odds and ends I've done with PLA
|
|
#
?
May 10, 2024 20:30
|
|
|
Prusa Mini has been exceptional for me, but recent releases like the 4 and XL have really been a step backwards in overall quality by a lot. A lot of tuning is required and there was even a bad firmware for the 3.5 and 4 that caused layers to shift for no reason. That version was current for over a year. It’s really a shame. My experiences with Bambu machines have all been excellent so far, it’s hard to recommend anything else these days especially since the X1/C/E can do some pretty solid engineering materials.
|
|
#
?
May 10, 2024 20:43
|
|
|
My company had Prusas for a long time and I thought those were fairly low effort. Until someone bought a Bambu. Within a couple months we had 4 and during busy projects they're running constantly. They print faster and more reliably with less hassle than anything I've used.
|
|
#
?
May 10, 2024 21:36
|
|
|
NewFatMike posted:Prusa Mini has been exceptional for me, but recent releases like the 4 and XL have really been a step backwards in overall quality by a lot. A lot of tuning is required and there was even a bad firmware for the 3.5 and 4 that caused layers to shift for no reason. That version was current for over a year. It’s really a shame. That reminds me I should get a refund for my XL preorder
|
|
#
?
May 10, 2024 22:14
|
|
|
 Sendcutsend: DIY Sheet Metal Bending Solidworks question: What should my workflow look like for using these bend reliefs? And what's the theory with these weird reliefs? It looks like it achieves 0mm bend radius on the inside corner? What's the K-factor for that? I think I need to figure out the K-factor, which will be based on the slot width and material thickness. Then design the part as sheet metal. Then flatten it and put the squiggle reliefs on the bend lines. Is that right? ryanrs fucked around with this message at 23:04 on May 10, 2024 |
|
#
?
May 10, 2024 23:00
|
|
|
Kfactor is a pain in the rear end given that you still need to know how the actual manufacturer wants bends noted on the print. Different tooling and machinery will produce equivalent bends slightly different ways/with different notation. I know at a previous employer the guys in the sheet metal area of the aerospace shop hated Solidwork's default settings and gave me a custom bend table to use for the grades of metal/tools/etc. they worked with.
|
|
#
?
May 11, 2024 02:08
|
|
|
Just a guess but the opposing zig zag might help keep the bends aligned Also looks like a really excellent outside tig welding seam
|
|
#
?
May 11, 2024 02:09
|
|
|
Hadlock posted:Also looks like a really excellent outside tig welding seam lol I'm gonna make it out of galvanized
|
|
#
?
May 11, 2024 03:06
|
|
|
I watched the video on the page and it looks like it helps with fixturing. Imagine fixturing that upright with individual flat parts. This trades easy bends for fewer parts. I would not ask it to hold a load through a bend like that. The structure is compromised to allow easy bending. Stitching the corner would bring strength back. I think it will help with non 90d structures and smaller things. I'd say it can be good enough for an electrical enclosure, heat shield, faraday cage, sun shade, etc.
|
|
#
?
May 11, 2024 03:20
|
|
|
ryanrs posted:lol I'm gonna make it out of galvanized I feel obligated to point out what you probably already know, don't weld galvanized steel unless you absolutely have to, and even then use proper ventilation and respirator
|
|
#
?
May 11, 2024 04:57
|
|
|
Oh, I didn't even notice that video. But car suspension guy is not using the wiggly reliefs shown on the web page. Maybe I'll go with my original 2-piece design, and also order some test pieces with zigzag reliefs of different size. Then I can bend them and measure the results. e: It doesn't need welding to attain the required load rating of 0.5 kg. And galvanized steel laser cuts SO CLEANLY. You get a nicer edge than e.g. aluminum. ryanrs fucked around with this message at 05:36 on May 11, 2024 |
|
#
?
May 11, 2024 05:34
|
|
|
Ya, go for it. I'd like to see how it turns out. Here's a nice video about building custom enclosures. He designs them as two U shapes that come together with flanges. https://m.youtube.com/watch?v=RopgrECLSJc
|
|
#
?
May 11, 2024 06:26
|
|
|
 See what I'm trying to do here with the overlapping tabs? Driving in 1 wood screw secures both tabs and forces the holes into alignment. I want to design this in Solidworks with the two tabs in the same plane, i.e. intersecting. But Solidworks really doesn't like it when they intersect. I can turn off intersection testing with some option under Performance, but then I get other errors (zero thickness geometry) when trying to extrude a cut on the tabs to make the hole. Changing the tab heights so they don't intersect changes the size of the reliefs and the tab length in a way I don't want. Likewise, changing one tab to 100 degrees will cause problems with hole alignment, because tracing one tab outline onto the other projects it at a slight angle. I don't think I want to accurately model the wavy S-shapes that the real tabs will take when the screw is driven home. Is the right move to model the holes after the part is flattened? ryanrs fucked around with this message at 02:23 on May 12, 2024 |
|
#
?
May 12, 2024 02:04
|
|
|
If the picture is satisfactory, can you not convert entities on a sketch plane in the tab that isn’t cut by the first extrude cut operation? That’ll give you two holes driven by the same dimension at least.
|
|
#
?
May 12, 2024 02:30
|
|
|
NewFatMike posted:If the picture is satisfactory, can you not convert entities on a sketch plane in the tab that isn’t cut by the first extrude cut operation? The picture isn't quite perfect, because the tabs are on parallel planes, but with an offset. The offset needs to be material thickness + epsilon or else I get zero thickness geometry errors. So that's 1.3mm, which is actually kinda big. Having the tabs overlap seems to work great if I turn off intersection testing and don't try to do further operations on the tab faces.   Here's what it looks like with the holes cut after flattening. In the first image, the tab outlines are fighting each other in the depth buffer or something. The edges flicker in and out when e.g. rotating the part. But the flat view looks exactly right. The curved ends of the tabs carry the center point through the flattening process, making it easy to locate the holes later, since they are concentric with the tab edge.
|
|
#
?
May 12, 2024 02:54
|
|
|
Yeah solid bodies don't like to be intersecting with themselves, it's why you're getting errors with your current approach and you're having to flex the rules in the performance stuff. I also think you'll end up with a mismatch in length irl as well if you don't model the overlap in a way that works. I'd do one at 85 and one at 95 and dimension the holes off the end of the tab vs making them concentric, or do a jog bend to lay one of them over the other.
|
|
#
?
May 12, 2024 03:04
|
|
|
There's enough clearance in the screw hole to allow for some misalignment. But it's a wood screw, not a dowel pin, so it was never going to be precision located. Although, what if I used flat head screws, and different size holes in the two tabs, so each engages the conical screw head at the corresponding tab height? Amusing, but probably a bad idea, because how precise is that wood screw's location, anyway? Oversize holes and truss head screws are probably better.
|
|
#
?
May 12, 2024 03:27
|
|
|
It seems to me like the wrong/lazy CAD answer but the right real world answer for this one-off Wi-Fi mounting bracket is probably to drill that hole through both tabs after the bracket is assembled and they overlap.
|
|
#
?
May 12, 2024 04:30
|
|
|
Perpendicular slots?
|
|
#
?
May 12, 2024 04:36
|
|
|
Drill holes? Like with my hands? Am I even in the right thread? p.s. this is 0.048" mild steel. it'll go where I tell it, within reason.
|
|
#
?
May 12, 2024 05:18
|
|
|
ryanrs posted:
 Change the indicated bend so that the higher tab ends up .001 inches above the lower one. You won't notice it in the finished part and Solidworks will settle down. Sagebrush fucked around with this message at 07:50 on May 12, 2024 |
|
#
?
May 12, 2024 07:47
|
|
|
0.001 difference, or material thickness + 0.001? If it's just 0.001, then the surfaces get disambiguated, but 99% of the volumes still intersect. I tried that and it still gives an error.  Rebuild Errors: Operation failed due to geometric condition. If I set offset = material_thickness - 0.1mm, it gives the same error. The tabs need to be fully clear of each other for the cut to work. I think I am fine with modeling the cuts after flattening. The workflow is convenient (at least for this particular model), and it looks like it'll give me what I want.
|
|
#
?
May 12, 2024 16:06
|
|
|
 Solidworks posted:Rebuild Errors: Operation failed due to geometric condition.
|
|
#
?
May 12, 2024 16:10
|
|
|
ryanrs posted:0.001 difference, or material thickness + 0.001? Offset is material thickness plus 0.1 mm to give clearance. Bending both tabs at the same plane doesn't work because the part is geometrically impossible  Offsetting by exactly the material thickness, while theoretically possible to manufacture, gives a warning about intersecting faces, and then the extrude-cut for the hole can fail with a zero-thickness geometry error, as you've noted. (Rhino would be able to handle this situation  ) Offsetting the cut by the material thickness plus some miniscule amount of extra tolerance produces a gap between the surfaces. In this case the stock is 1 mm thick and the extra tolerance is 0.2 mm to make it visible (so 1.2 mm total offset), but an undetectable value like 1.01 mm also works just fine.   The extrude-cut through both parts works fine with no errors with the tolerance added. 
Sagebrush fucked around with this message at 17:57 on May 12, 2024 |
|
#
?
May 12, 2024 17:51
|
|
|
I won't notice the 0.01, but I will notice the material thickness. It'll move the location of the bend upwards, creating a noticable gap between the tab face and mounting surface. The real shape, which I don't want to model, is more of a twist of the curved tab. The tab is quite compliant in twist. Modelling something like a jog bend is both less accurate in appearance, and in the final hole location (since I don't know the K-factor or other parameters). Feels like more effort for a worse result (tho it would keep the software happy).
|
|
#
?
May 12, 2024 18:18
|
|
|
Ah, I see. Well, you're kind of out of luck, then. If you want to use the software to do the calculations for you, then you have to model your part accurately. If you want to take a shortcut in the modeling (not making the proper bend), then you might have to work around a second time (manually finding the hole locations) to compensate. GIGO Rhino could handle this geometric condition perfectly, but it doesn't do sheet metal parametric stuff like Solidworks. C'est la vie
|
|
#
?
May 12, 2024 20:13
|
|
|
The swept-flange feature may be able to achieve what you're looking for, accommodating properly for the slight shift in the hole position when the material bends: For some reason though when I flatten out the pattern that swept flange disappears. Like it's not supported by the algorithm or something? And this is after drawing it with only lines and arcs because the feature doesn't support spline paths. Reminding me why I fuckin hate Solidworks lol
|
|
#
?
May 12, 2024 20:29
|
|
|
lol okay here it is  Model the first tab. Cut-extrude the mounting hole into the first tab. Model the second tab on the same bend plane, intersecting the first tab as shown, which will create the geometry warning. Offset the cut-extrude sketch inwards by 0.1mm and use that to make a second cut-extrude hole into the second tab. (You can't use the exact same sketch for both as that produces zero-thickness geometry). The geometry is technically impossible and the holes can't be exactly the same size. However, I suspect it will be fine for your purposes. Good enough! Sagebrush fucked around with this message at 20:44 on May 12, 2024 |
|
#
?
May 12, 2024 20:42
|
|
|
For the bend relief test coupons, I hit some nasty N^2 algorithm when creating the slots. The slot ends overlap and Solidworks couldn't figure out which contours should be removed (all of them!). Manually selecting 70 regions in the Extruded Cut feature window was real bad, ha ha. 5+ seconds to click a region towards the end. Maybe I need to draw these as line segments, then use offset entity to turn it into a slot? At least Solidworks should be able to automatically select all the outlines then.  e: making these wave reliefs is a massive pain in the rear end compared to splatting down a bunch of equal-size slots in hole wizard ryanrs fucked around with this message at 21:15 on May 12, 2024 |
|
#
?
May 12, 2024 21:07
|
|
|
|
| # ? Jun 3, 2024 09:09 |
|
|
I tried a sketch with a line it it with plans to use it for a sketch bend later. I drew a new sketch with the first slots as lines, offset the lines with arc caps. Extrude cut that. Then linear patterned the feature along the line. I tied the pattern to the sketch with an equation. Then I did the sketch bend on the first sketch. edit, this https://www.reddit.com/r/SolidWorks/comments/xjlofc/flat_pattern_to_apply_to_sheet_metal_bends/ says to use fold/unfold. Like design your bent part, then unfold it, add cuts, fold back. bred fucked around with this message at 23:11 on May 12, 2024 |
|
#
?
May 12, 2024 23:07
|
|