Thickness tool/Sheetmetal Workbench issues?

Post here for help on using FreeCAD's graphical user interface (GUI).
Forum rules
and Helpful information
IMPORTANT: Please click here and read this first, before asking for help

Also, be nice to others! Read the FreeCAD code of conduct!
Post Reply
Trak
Posts: 4
Joined: Sat Jan 21, 2023 1:49 pm

Thickness tool/Sheetmetal Workbench issues?

Post by Trak »

I'm using the realthunder branch of realcad on EndavourOS which is an arch linux based distro.

I've been working for a week trying to make a bucket seat. I got a very basic version working (Seat Original), but I wanted to get some angles in there for comfort, i.e. wider shoulders, wider knees, taper in above the shoulders.

Over the course of the week I've discovered a few bugs I'll mention here in case they help anyone else. Extrude cuts (pocket) that line up with an existing edge on the model break the thickness tool, sometimes corners do this too, offsetting the edge of the pocket slightly usually allows it to work, at the expense of incorrect geometry.

The sheetmetal make bend tool is fairly picky about the face at the end of the edge being converted into a bend (the face representing the thickness of the sheetmetal), this one I haven't really figured out but is visible in the "Seat" file, the dimension in Pad11>Sketch18 has some tipping point between 16 and 15mm which breaks the bend at the shoulders, I haven't figured out why.t

Also lofts which should in theory produce flat faces do not, because of rounding errors. This causes edges to turn into splines instead of straight lines and what should be flat faces to not be totally flat, causing thickness and make bend to sometimes fail. I "fixed" this by redoing what I did but with pipe operations. Though the thickness tool always seems to fail in new and creative ways as show in "Seat Thickness Fail" and I haven't been able to figure that out (trying to remove the two faces in the way of the seat causes thickness to fail).

These aren't the only iterations I tried, but I didn't keep all of them. "Seat" a somewhat working example of what I'm trying to achieve, but it's built more awkwardly just to try and get things working (i.e. extruding a whole parallelogram at the start rather than just the seat part, this was to get a more favourable result with the thickness tool). The shoulder bend isn't quite a projected version of what I wanted but I dare not try fix it because things keep breaking when I do, and the weird 16mm required offset doesn't sit right with me even though none of these dimensions are critical.

As a last ditch attempt while I was writing this I skipped the thickness tool and manually created a hollow shell from the blank by extruding individual faces in "Seat Final". This sort of worked. There were some oddities at the edges where two faces would meet, which I would "trim". At first the make bend tool failed here as well, but by trimming some of the parts where two faces meet to look "nicer" that made it work. So it's possible, but it's quite an awkward process.

Am I missing something? Maybe someone can point out an error in my workflow or something, or an easier way to do this. I'm aware of the ability to build a shape up by adding "walls" using the makewall tool in the sheetmetal workbench, which most tutorials bring up as the "preferred" way of doing things. The issue there is with complex shapes that have compound angles like this bucket seat. Makewall requires you to manually input the wall bend angle and any mitres on the side edges. This is fine when the walls are 90 degrees, but I'd be better off drawing a flat pattern directly than calculating all the compound angles in the bucket seat.
Attachments
Seat Thickness Fail.FCStd
(113.16 KiB) Downloaded 10 times
Seat.FCStd
(724.65 KiB) Downloaded 11 times
Seat Original.FCStd
(427.97 KiB) Downloaded 12 times
Master.FCStd
(3.02 KiB) Downloaded 7 times
User avatar
ThanklessLiving
Posts: 98
Joined: Sat May 30, 2020 1:49 pm

Re: Thickness tool/Sheetmetal Workbench issues?

Post by ThanklessLiving »

I was skeptical by how you used a base "block" followed by thickness but in the end it really seems to have worked and the part appears to unfold correctly and it can likely be produced in the real world.

Sheet metal parts can easily spiral out of control with how many operations you have to make, a "simple" door can end up with 100 features or more as you create all the bends, refine all the corners and miters and add various holes. This can be mitigated by mirroring and combining as many features at once, for example you might want to try the Relief and Junction tools in the Sheet Metal workbench. These might allow you to coming the 4ish "Pad" features for cutting out corner reliefs into one or two.
User avatar
Shalmeneser
Veteran
Posts: 9475
Joined: Wed Dec 23, 2020 12:04 am
Location: Fr

Re: Thickness tool/Sheetmetal Workbench issues?

Post by Shalmeneser »

Full recompute shows errors.
Sketch upon face = TNP
User avatar
papyblaise
Veteran
Posts: 7870
Joined: Thu Jun 13, 2019 4:28 pm
Location: France

Re: Thickness tool/Sheetmetal Workbench issues?

Post by papyblaise »

I would do the outside first, the thickness then the seat with a few nice holes to make it light
Attachments
seat1.FCStd
(391.88 KiB) Downloaded 8 times
seat1.JPG
seat1.JPG (25.78 KiB) Viewed 594 times
User avatar
papyblaise
Veteran
Posts: 7870
Joined: Thu Jun 13, 2019 4:28 pm
Location: France

Re: Thickness tool/Sheetmetal Workbench issues?

Post by papyblaise »

this time, a sheetmetal model that gave me a headache
Attachments
seatPPB.FCStd
(544.81 KiB) Downloaded 8 times
seatPPB.JPG
seatPPB.JPG (20.73 KiB) Viewed 530 times
Trak
Posts: 4
Joined: Sat Jan 21, 2023 1:49 pm

Re: Thickness tool/Sheetmetal Workbench issues?

Post by Trak »

@thanklessliving I did try using the make relief function in the sheetmetal workbench, manually making those corner reliefs was not my first course of action. Corner relief normally didn't work, and the identical symmetrical corners often failed differently which I found odd. I just left that out in the OP because I stopped using it quite quickly.

The block approach then convert to sheetmetal is how I was taught to do sheetmetal, and in my opinion works best with complex geometry as you can create the overall shape quickly and convert the necessary faces into sheetmetal once you know the overall shape is right. If you start bend by bend you can often back yourself into a corner where the 4th bend down the line interferes with something earlier in the process because that part wasn't fleshed out in the begining. Also like I mentioned in the OP, it's the only easy way to get bends and edges with compound and complex 3d angles. Bend by bend is fine and all when all the bends are 90 degrees like in most of the tutorials, but the splay and mitre angles in that seat would be a nightmare.

@shalmeneser I'm using the realthunder/linkstage3 branch which is a bit more robust with regards to the TNP. I'm still getting some errors if I change a dimension early in the feature tree but nothing huge. Not sketching on faces or referencing shape geometry goes back to the issue of manually working out all the compound bend and mitre angles.

@papyblaise I did try the thickness tool, but like I said it kept failing. It seems to work in your particular example to some degree, like the in the "Seat" file I posted. You seem to have run into the sheetmetal and unfolding issues with your first example. Your second example with a sheetmetal model that unfolds isn't ideal for fabrication. None of the joins line up quite right when you look at them closely which creates problems when it comes down to fabricating the seat. For something like this seat that isn't dimensionally demanding I can force them into alignment and grind out any corners that stick out, but for the rest of the project that requires precise fitment and dimensions it won't work. The long relief along the seat base and side is also less than ideal, but that's the sort of limitations of the sheetmetal workbench that I'm talking about. And I totally get what you mean about it giving you a headache. I've thrown everything I can think of at it to get the results that I got an they're still less than ideal from a process point of view.
User avatar
papyblaise
Veteran
Posts: 7870
Joined: Thu Jun 13, 2019 4:28 pm
Location: France

Re: Thickness tool/Sheetmetal Workbench issues?

Post by papyblaise »

controversy over sheetMetal , thickness , subtractivePipe ... is not appropriate
Unfold is not simple and requires planar faces, subtractive pipe does not do planar faces
I cut half of your "seat fail" and break it down element by element, I make fillets at the angles which will be unfolded and keep the flat side as an unfolding support
I have to edit my result by erasing the artificially created sides and aligning the seat with the back and making a symmetrical
it only took me 1 hour
50 years of sheet metal have taught me that precision cannot be rigorous by bending, a little better by stamping
Attachments
Seat Thickness no fail.FCStd
(713.65 KiB) Downloaded 7 times
Seat Thickness no fail.JPG
Seat Thickness no fail.JPG (27.21 KiB) Viewed 404 times
Trak
Posts: 4
Joined: Sat Jan 21, 2023 1:49 pm

Re: Thickness tool/Sheetmetal Workbench issues?

Post by Trak »

I tried one more approach, that is by building the object up face by face using sketches and pads to sheet metal thickness.

This is the best result I've gotten so far, it's robust and no dependencies break (for me at least) when changes are made at the start of the feature tree. The only referenced geometries are to sketch lines, which are fairly robust with regards to changes.

The approach isn't too clunky and the feature tree is relatively tidy. It's somewhat similar to the blocking approach I used before but instead of an actual 3d block object it has a bunch of plan/side/front view sketches which is a bit less intuitive but it's what works. Converting to sheetmetal was flawless, and the bends even all look right with no weird glitch artifacts that would sometimes show up at the ends of a bend (didn't affect anything anyway, but hey). The oddities created by the overlap of joining edges (since I extruded by sheetmetal inwards) don't seem to affect anything so I didn't bother trimming them.

@papyblaise You're file throws a few errors for me. The main issue I see is that it's a bit hacky like my earlier "Seat" file which messed around with un necessary geometry to get a more favourable thickness result, I'm trying to avoid "hacky" methods because they feel unrealiable. It also requires manual post processing to modify the unfold pattern. That is an issue because unfolds are non parametric and do not respond to changes in the feature tree, so any changes require significant manual work to generate a new flat pattern.

I'm also fully aware of the limitations of sheetmetal fabrication I'm not expecting micron level precision, but as a fabricator by trade I'm also fully aware that a crooked pattern will only ever make a crooked part. I've welded up my fair share of crooked patterns, and I much prefer to avoid them.
Attachments
Seat.FCStd
(350.94 KiB) Downloaded 6 times
User avatar
papyblaise
Veteran
Posts: 7870
Joined: Thu Jun 13, 2019 4:28 pm
Location: France

Re: Thickness tool/Sheetmetal Workbench issues?

Post by papyblaise »

When you talk about Sheetmetal's lack of pressure, I poof :lol: :lol:
Attachments
pouf pouf.JPG
pouf pouf.JPG (13.88 KiB) Viewed 349 times
Trak
Posts: 4
Joined: Sat Jan 21, 2023 1:49 pm

Re: Thickness tool/Sheetmetal Workbench issues?

Post by Trak »

I'm not sure what you mean, I haven't said anything about "lack of pressure" and don't really understand what that's supposed to mean. Not sure what's supposed to be in that jpg, it's too zoomed in for me to get any context.
Post Reply