Are there plans to make sheet metal easier to use anytime soon?

shawn_crocker

Lets face it. Onshape sheet metal is simultaneously great and bad. It's great because when you are using it to make very predictable sheet metal pieces, its very fast. But, when you are trying to do something with it slightly out of the ordinary, but for some reason the developers overlook it as something to account for in code, its super annoying. The main issue I find is that sheet metal seems to always be trying to correct what you are doing. Sometimes it like playing wack-a-mole. You trim over here and then something over there pops out. Or, for no reason whatsoever other then clear oversight in the featurescript functions, will not allow you to do something.

For example, often I need to split a sheet metal part into too pieces. OK, so I can't use split. Haven't been able to do use split since forever. Why is that?? What does it take, to trigger Onshape to realize that this is just something that needs to be done? I live in Quebec, Canada. Our roads are notorious for being poorly maintained. Sometimes I drive down the same road that I have driven down for years and it is in the same nearly undrivable state as when I started using it, and I ask myself the same question. What is the trigger for repair here? What does it take to just get an overhaul to make this thing work the way most people want it too? I get agile mindset and releasing now, fixing later. In the case of sheet metal, is later ever going to come?

Like most people, I get around sheet metal part splitting by using an extrude. But, sometimes even that will not work. Depending on if there are other sheet metal parts from the same feature close by, sometimes it won't let even that simple operation pass.

lana

Luckily, there is a mechanism for this

https://forum.onshape.com/discussion/11765/split-active-sheet-metal-parts

We do pay attention.

shawn_crocker

https://forum.onshape.com/discussion/comment/128456#Comment_128456

I think what I'm getting at is, does Onshape really need the community to tell them this should work? I see lots of improvements coming into the system that are really great. I'm sure some of them were tasked internally and are not necessarily directly driven by a specific user improvement request. Am I wrong? If nobody complained about it, is Onshape really going to just assume everyone likes not being able to use the split feature? Is the lack of support for the split feature on active sheet metal actually logged somewhere as intended and desirable behavior? Sometimes I move a sheet metal face and another one pops out which I then need to perform another move face on. If I don't take the time to write up a formal improvement request, is Onshape really just going to assume everyone likes it that way? Is complaining about these things in support tickets and through the forum in this manner not a good enough indicator?

My main point here is, I personally, do not think Onshape is investing enough effort into the sheet metal world. It feels like it is constantly being left in a mildly dysfunctional state. Same goes for the Frames feature in my opinion. There is just so much about it that is awkward I feel like I would have to write a 5 page report just to complain about it.

bryan_lagrange

Would love to have a Teams meeting with the Onshape sheet metal team along with Onshape sheet metal power users like @shawn_crocker , @john_mcclary @Derek_Van_Allen_BD @eric_pesty @sebastian_glanzner and others to share some thoughts on the sheet metal functionality in Onshape. I agree with Shawn that some times their are some gremlins lurking in sheet metal that pop up out of no where and you have to do a work around to figure out how to solve them. Would this be possible @lana @NeilCooke @mlaflecheCAD @Vajrang_Parvate

john_mcclary

I would make myself available for such an opportunity

shawn_crocker

@bryan_lagrange I'm down with it. 😊

Derek_Van_Allen_BD

Speaking as someone in the community that does extensive work with both frames and sheet metal and spends a lot of time bouncing sheet metal issues back and forth with @lana and others on the dev team, they do really good work taking in feedback and making changes and improvements when requests are made through the improvements requests page or via support tickets when something gets buggy. The reality is that developing features for either sheet metal or frames is indeed a complex undertaking from an architectural standpoint and that some things are less about adding a feature to the list and more about making sure that new integrations don't completely destroy other features with new bugs. For example conic sheet metal was technically supported by the engine for months ahead of public release and you could access it if you were savvy enough with the featurescript side of Onshape to get a preview, but it didn't make its public debut until the (literal) corner cases around flanges and other sheet metal tools went through some bug handling to make sure they didn't push a build to their users that they would get even more support threads generated.

It's pretty likely I could produce a split feature that does what you're looking for with sheet metal in a weekend without a ton of friction, but the difference between myself and the core developers is that my scripts don't go through the same unit tests and validation that the real devs put into the product. I often publish tools that I know break other processes but I can get away with doing so by telling people on my team "just don't do that thing after the other thing" instead of building in real error handling.

Additionally, some of those gremlins you guys are encountering with extra faces being generated from cut extrudes or move face operations are technically speaking correct from an engine standpoint, if unintuitive. The little stick-out bits that result from a cut across a rip are due to the edges on that corner of the rip becoming single sided edges on the surface and unassociated from their surrounding rip geometry on either side, so they extend to the virtual sharp of that corner of the surface definition. Partial flange generation operates in a very similar way and we're pre-disposed to accepting the results as they appear because we see partial flanges that extend all the way to the virtual corner regularly, but we rarely or never see rips being done the same way. Maybe there's something to be said about the difference in engine behavior from user expectations but some amount of "gremlin squashing" would require additions or changes to the engine that are behaving the way they do with good reason.

Derek_Van_Allen_BD

On the subject of frames: I agree that there are many limitations in the current implementation of frames that need navigated both with the default tools available and with featurescript development. I have an improvement request thread I made about this where I talk about some of the pain points but I can re-summarize here. In short, many of the challenges of drawing frames are very analogous to things which are inherently easy with sheet metal due to the effort the team has expended in making the simultaneous sheet metal engine. Things like aligning all frame members to lie on the inside edges of a volume in one feature, or being able to draw parametric secondary or tertiary cross members of frames, or advanced corner treatment, or unrolling groups of frames for laser cutting processes could be a lot simpler if there were hidden wire bodies that could be used to construct and hold centerline information or the other alignment pips that frame attributes currently don't carry. The cut list math alone spends a lot of effort reconstructing a centerline path from geometric heuristics which can fail if non-attribute sensitive operations interact with the frame bodies and one of the scripts on my list is a frame re-attribution feature that fixes cap faces when they get destroyed. A lot of that logic can be skipped with a similar approach taken by the sheet metal engine, but also a lot of that logic happens on the proprietary side of Onshape and not the featurescript side.

shawn_crocker

I have also assumed the funny little issues with new geometry popping up when removing other geometry is happening because of some other underlying functions that are attempting to do there job. I don't think all sheet metal systems suffer from this type of thing so glaringly(Its been a long time but my mind goes to solidworks here), and in my case, so often. I almost never make a new sheet metal design with out having to dance around undesirable behavior many times during my work. At this point, I don't think we can consider the under lying engine as actually doing its job here. If it was, people wouldn't be complaining about it. If the engine is consistently doing something that almost every user has to immediately undo, in my opinion, that is a underlying function that is not doing its job. The underlying function that is reforming sheet metal as you edit to maintain a proper sheet metal part needs to be improved so that is does not just move things around when it is 90 percent certain the user wouldn't want that. I get that would be hard for them do fix this far into things.

I have suggested it before and I suggest it again, we need a new feature that would be sort of like a partial finish feature. The sheet metal would remain active but the helper functions underneath would be turned off. I would love this because I would be able to generate the bulk of my operations with full sheet metal geometry automation, then turn off the sheet metal helper functions, step outside and breath a big breath of fresh air anticipating the piece and tranquility I am about to experience, and then continue the final edits to my parts knowing the system is not going to do anything I didn't tell it to.

bryan_lagrange

Solid Edge, which also uses the parisold kernel, is what I compare sheet metal functionality in Onshape with. When a sheet metal gremlin pops up I look to see if the same pops up in Edge, if it doesn't I wonder what is going on under the hood to make it act differently.

S1mon

I don't use sheet metal as much as some of you, but when I do, I'm often frustrated with the bend reliefs. Inevitably I either have to live with weird little problems, or do a lot of manual move face, replace face, cuts, etc. to make something that doesn't look amateurish to a sheet metal vendor. I fear the need to keep things backwards compatible will prevent any wholesale clean up and refactoring of those algorithms.

john_mcclary

don't get me wrong, I love Onshape's sheet-metal. For the most part Onshape's sheet metal has bugs but is still miles ahead of what I've used in other cad.

Sheet-Metal is mostly of what we use. for example here's MDR conveyor with 56 unique manufactured parts. only 2 aren't sheetmetal. The tower tube, and the lexan window.

So, it's fair for me to say. I live and breathe in Onshape's sheet-metal environment. There is not much more important in a CAD system than that for us. Yes, there are gremlins, most of them are more 'annoyances' than deal breakers. Remembering back to SolidWorks. I still feel Onshape's automatic relief system, (even with its random quirks) Is still far better than having to manually do relief everywhere in SolidWorks.

There are long conversations I could have with some struggles, most of them I have tickets for. Most of those they have already rolled out improvements. I would love to see more sheet-metal focus again from the Onshape team.

Here is a quirk I just encountered for the first time today. I'm deriving a sheet-metal part from another part studio (it's built off adjacent geometry) I checked the box to import as active sheet-metal (which is a nice recent feature) But for some reason this one behaves.. different…

I ended up having to add 7 steps into the tree to fix it so I could push the flat pattern stocksize into the part properties. Usually my featurescript for this does it automatically for all parts in the studio at once. but it was measuring the bounding box of the formed body instead of the flat pattern. even though my debug report was showing it went through the flat pattern sub routines.

I basically had to re-create the sheet-metal from faces, to make a clone of the part. Then did things to it. Basically, measure didn't work, I can't select faces in the flat pattern to sketch there. Just weird random stuff. (side note, when making this post, Measure now works again… hmmm)

But I can add flanges, so it is truly active:

I haven't created a bug report yet, but I'm wondering if anyone else seen this, or is it just a one-off glitch?

shawn_crocker

@john_mcclary I have seen a similar weirdness where the system has trouble resolving selected edges after using a hole feature that goes though more than one face of the same sheet metal parts. My experience was, I could select an edge or something on the flat pattern to make a sketch line coincident to, but it would immediately become dangling. I haven't experience this in a while, possible from stability improvements or possible because I just haven't accidently performed the same design steps again.

I do seem to remember solidworks auto adding relief to parts in a similar way to Onshape. I would say I prefer Onshape sheet metal because of how nicely it seems to prefer making sheet metal using the convert option. But, solidworks does do that too so, it might just be I didn't really make use of it during my time using that system to realize how much more powerful it is to convert solid geometry to sheet metal.

Derek_Van_Allen_BD

https://forum.onshape.com/discussion/comment/128486#Comment_128486

Having gone back and reviewed the Solidworks version of convert to solid recently, it's way less functional than the Onshape implementation. To my knowledge it's hard limited to single sheet metal bodies being generated at a time and never allows for multibody sheet metal generation. It was good for certain things but still required a lot of manual rip geometry generation after the fact if you were doing anything complex.

Now I do wish that Onshape had more rip styles than just butt joints or corner rips, but that's a whole separate improvement request.

john_mcclary

Here is another one that started happening a few updates ago. I haven't been detailing much recently so I keep forgetting to bug report this. (I'm making a bug report for this one now)

The tangent lines is the flat pattern are showing up if you have tangent lines to solid. So, you have to manually hide the tangent lines in the flat pattern.

It used to always be hidden by default in the flat. the two outside the border are tangent phantom and tangent hidden options when inserting the view.

those two still work fine and don't show tangents in the flat pattern.

shawn_crocker

https://forum.onshape.com/discussion/comment/128478#Comment_128478

I have been gingerly snooping around solid edge. What do you make of it in general, setting assign its a yukky files based thing? I was lightly investigating its automation package for generating new design variants. Since we now have to pay extra just to use the API in Onshape so that we can use our own automation, I was scoping it out to see if the $1500-3000 per year extra just for API access was still worth it. 🤬

bryan_lagrange

@shawn_crocker I have used Solid Edge for about 26 years now (was our primary CAD system, now it is Onshape). We are not on maintenance but still run a copy for legacy data. It is more stable than SolidWorks and has better sheet metal functionality than Works. The draft environment is still one of the best out there. If you we had to go back (which I don't see happening) that would be our choice.

shawn_crocker

https://forum.onshape.com/discussion/comment/128491#Comment_128491

Nice. I see a lot of people bad mouthing it being complicated to achieve anything with lots of extra clicking. We also would most likely never go back to designing directly in a file based system but, we do want a customer facing tool on a website to automate orders/model generation. Without building one ourselves and paying twice for Onshape's API(once for our yearly agreement and once for the new, extra BS API calls charge), I'm left getting interested in files all over again. 😣

Derek_Van_Allen_BD

https://forum.onshape.com/discussion/comment/128492#Comment_128492

Have you tried the Spokbee app? I did a demo with them awhile ago that ended up not fitting the broader marketing strategy that our internal teams ended up going with but it could probably serve that niche for you guys without needing to commit to the dev time and full API budget.

shawn_crocker

Thanks @Derek_Van_Allen_BD I hadn't looked into that one. I will test out there trial. Keeping it driven by onshape would really be desirable(as long as Onshape can keep there system from crashing so often😁).

bryan_lagrange

We are working with Spokbee. Great Onshape knowledgeable team. Mac Cameron would be your contact. mac@spokbee.com

Derek_Van_Allen_BD

@shawn_crocker let's say for sake of hypotheticals, when you are doing sheet metal splits is it typically an entire sheet metal context that you're wanting to bisect with multiple parts, or is it a single part out of many that you need to cut in two while the rest remain whole with their rip and bend geometry intact? Chainsaw or scalpel?

shawn_crocker

https://forum.onshape.com/discussion/comment/128496#Comment_128496

Thanks again Bryan!

shawn_crocker

@Derek_Van_Allen_BD I'm usually splitting one part. Often it would be a single sheet metal feature that generated many parts and a view of them need to be split down to fit in the raw material. I think a single split feature should be able to split multiple parts even if they are derived from multiple features.

Derek_Van_Allen_BD

https://forum.onshape.com/discussion/comment/128513#Comment_128513

Got it, so both options then.