Problem in large assemblies with losing projections, components not having moved.

[PrintBoy97]

Hello,

I am having a consistent issue with fusion losing my projections of components which do not change.

It happens both in normal assemblies, but also when using assembly context. My experience in fusion I would say is quite advanced but this problem I feel is quite basic and nobody seems to have answered this on forums before. Perhaps there is a problem with my workflow and I would love to hear from someone who would have a better approach.

Say for example that I have a sketch which contains projections of many different components. I project and intersect, lets say, around 20-30+ different parts, faces, edges, anything I need into my sketch. This is a lot and can be confusing to look at times, but can be necessary to build up the shape of certain parts.

My example is of a drone fuselage, similar shape to a normal aircraft but smaller, I trace around my components so that I can create mounting points and positions for reinforcing structure and other things. These sketches therefore become the backbone of my model.

I carry on with my model and everything is fine of course until I wish to manipulate one of the sub-parts in some way, perhaps an extrude feature I need to change, or move something a few mm left or right. Maybe the entire model of a motor we want to use is changed.

Of course, it makes sense, I just changed a part, now re-select the edges. But now I see the sketch and am faced with disaster. The entire sketch is yellow, and I am asked to re-point 20-30+ different lines, and click just for the outline to be the exact same way it was! This is not only very tedious and time consuming, but if I am not concentrating properly and select the wrong geometry, I end up making errors and changing my model without knowing.

Potentially, this happens not only when I manipulate a part, but when I add/remove one. I imagine that within the program code, each projection is assigned a line (eg. project 1, project 2,3,4 etc...) and when something is added or removed in the chain, the list shifts in name assignment and "loses" its position. This is just my hypothesis but anyway...

There must be a better way. Again perhaps my workflow needs adjusting and the answer is simple. Should I separate this process into many smaller sketches? I have read on these forums that many sketches are computationally expensive and bad practice, and if you need many sketches in one plane, it's better to use just one. Is this perhaps a case to use many sketches on one plane?

I am sorry that I cannot share my model as it is confidential and belongs to my employer. I hope that I explained myself well enough. I look forward to your answers.

[g-andresen]

Hi,

Without insight into the design, it is not possible to provide reliable information.
Share a dummy file with the described problem.

günther

[PrintBoy97]

Hi,

I appreciate that you wish to help for my specific case, and I have tried to make a quick dummy file recreating this issue, but I have not been able to. This is because it would take too much work to build up a complex enough file.

I feel that I have explained my scenario well enough for at least some kinds of pointers or best practices. I found this post on reddit describing a similar issue:

https://www.reddit.com/r/Fusion360/comments/z30cro/how_often_do_you_encounter_errors_in_fusion_360/

The most popular reply suggests that some sort of circular dependency could be the issue on a recompute. For me I do not think this is the case, but  the next part of the reply seems to line up with what I am thinking:

"When you're changing history, changes that create new regions in sketches and new projected edges from geometry can't always be resolved automatically.

If you split up your sketches into smaller units you can often avoid big topology changes when parameters change, allowing smooth edits."

Just for an idea, this is the kind of sketch I am dealing with when I talk about this.

Moving forward, I plan to make many small sketches relating to fewer parts at once, in the hope that the collapses in the model are smaller and more manageable when I edit a part. I would love to be proven wrong, but I do not have faith in fusion being fully error free when making changes in this complex a model. Though I do accept some responsibility also for not being a perfect user. The above lost projections come from assembly context, external to the work model from within the larger assembly file.

Please let me know if there is any more advice on this you can think of.

Thanks

[davebYYPCU]

You are the master of your own destiny.

(Even here , no file, pic, (someone else’s - helps) or video, dictates that we restrict answers to generalities.)

A sketch like that - umm.  If those projections were valid, how much time is required to weed just what each feature is needing from that sketch? 

KISS exists for a reason.

As discussion, you have 20 > 30 parts in a model, boss says I need .....

At this late stage, what are you asked to do, that needs those 20 plus parts’ outlines included into this sketch?

when I edit a part - isn’t that job done - Why was it later captured in this sketch?

[nater3796]

I didn't see your post until I created mine but, yes, I have had what sounds like the same issues, now that I hear you describe your problems. I wish I could help you but good to know an advanced user is having the same issues.