Solids that fail to regenerate

[line-weight]

What causes this?

 

I have one file where a particular solid fails to regenerate each time I need to recover it from a backup.

 

I can copy and paste this solid from an earlier version of the file - and pastes in fine - it's there and correct, but if I try to do anything to it, even just double-clicking into the subtraction without changing anything, I can't then exit it without it failing.

 

If I copy and paste this same object into a blank file, there doesn't seem to be a problem. There only seems to be a problem when I put this particular object in a particular file.

 

Is there any way I can fix this? Does it indicate a corruption in the file that I should be worried about?

 

It's a subtraction where some simple solids are subtracted from a 3d symbol.

[Pat Stanford]

Do you have Fillets or Chamfers on any of the objects inside the solid? It is very likely that they would not automatically regenerate.

 

Your signature says VW2018 is that correct? There has been a lot of work done on solids for VW2021 so they now understand Chamfer and Fillet history.

 

Also, you say solids subtracted from a 3D Symbol.  I don't think it is possible to subtract solids from a symbol. You would have to convert it to something other than a symbol first. Or I am misunderstanding what you are saying.

[line-weight]

1 hour ago, Pat Stanford said:

Do you have Fillets or Chamfers on any of the objects inside the solid? It is very likely that they would not automatically regenerate.

 

Your signature says VW2018 is that correct? There has been a lot of work done on solids for VW2021 so they now understand Chamfer and Fillet history.

 

Also, you say solids subtracted from a 3D Symbol.  I don't think it is possible to subtract solids from a symbol. You would have to convert it to something other than a symbol first. Or I am misunderstanding what you are saying.

 

yes - this is in 2018.

 

There aren't any fillets or chamfers, no.

 

The symbol consists of two solids: an extrude, and a tapered extrude.

 

Subtracted from this symbol is an addition of two extrudes.

 

It is in fact possible to subtract solids from a 3d Symbol - this was somewhat to my surprise.

[line-weight]

I'm having this problem with another solid now.

 

This time, it's still unhappy even if I copy it into a fresh file.

 

Any hints to troubleshoot what's causing the problem?

[Pat Stanford]

Can you post the object copied into a clean file for us to take a look at?

[line-weight]

Here you go -

 

You'll notice that it contains some symbols, and might suspect these to be the problem, but I've tried converting them to regular solids and it doesn't solve the problem.

I'm wondering whether this is in any way related to the problem I posted in another thread, about things in subtractions/additions moving location, because there seems to be a bit of that going on with this object.

failed subtraction.vwx

[line-weight]

This object is causing me so much trouble that I've virtually rebuilt it from scratch. But now, it won't let me perform an addition as part of this rebuild. With this post I've attached a file with all the components that I want to add together - but it won't allow this, and I can't work out why.

 

Any help much appreciated.

 

 

wont_add.vwx

Edited November 26, 2020 by line-weight

[line-weight]

Is anyone able to offer any insight here?

[Tom W.]

This is probably no help whatsoever but I could add most of it together with the exception of the two objects I've dragged away as shown in screenshot

Once they were dragged away from the rest of it they could then be added to form a single Solid Addition. 

I have no idea why they wouldn't add when in their original (correct) positions.

Is the aim to convert the overall object into a single generic solid?

[line-weight]

Thanks for the reply @Tom W.

 

Hrm, I can't replicate that, but also can't quite see exactly which objects you've moved. You say two objects, but it looks like more than two objects to me? In other words, the two things you've dragged away, are in fact made of two or more objects each - at least, they are when I open the file.

 

The ultimate aim is to make the whole lot into an addition, which I can then subtract things from. I'd rather not make it a generic solid if possible, because it removes the possibility of later edits to the geometry.

[Tom W.]

Yes you're right sorry: I meant to say that I just went around the model adding the various solids together until I identified the ones that wouldn't add together, and ended up with three separate solid additions, which could only be added together when they were dragged away from each other.

 

In the process of pulling the model apart I found two Line objects which I deleted so not sure if they were part of the problem? Because you can't perform Add Solids on 2D objects? When I repositioned the solid additions I'd pulled away + ran Add Solids again it formed a single solid addition:

 

Not sure I put the objects back in quite the right place. And also because I performed multiple Add Solids commands it's a nightmare to go back into the history + edit. So sorry not much help but if you can dig out the line objects + delete them that might allow you to run the command on the objects all at once? Select Similar then Invert Selection?

[line-weight]

Yeah, I don't think the lines are the problem, I tried removing them and I think it automatically deselects them when you make the first attempt anyway.

 

However - you seem to be right that the positions of the objects relative to one another is significant. I tried separating them all using the distribute command ... and then it's happy to add everything together in one go.

 

Is there any valid reason why one object's position in relation to another should make an addition impossible to compute? If not, then is this a bug?

[Tom W.]

2 hours ago, line-weight said:

Is there any valid reason why one object's position in relation to another should make an addition impossible to compute?

You'd rather hope not given the purpose of the command but hopefully someone more knowledgable can say. Be interested to know if bug or not

 

By the way what is the model of, if you can say?

[line-weight]

22 minutes ago, Tom W. said:

You'd rather hope not given the purpose of the command but hopefully someone more knowledgable can say. Be interested to know if bug or not

 

By the way what is the model of, if you can say?

 

It's a bit of victorian railway viaduct in south London. Very tricky to model even when VW isn't throwing its own spanners in the works!

 

In a bit more context -

 

[zoomer]

2 hours ago, line-weight said:

Is there any valid reason why one object's position in relation to another should make an addition impossible to compute? If not, then is this a bug?

 

 

It is much easier and less critical if objects are overlapping quite a bit.

And if they are nearly matching but not exactly, Boolean Operations can fail.

Like when Faces are nearly coplanar overlapping.

 

It was a pain in Microstation but I always thought VW is overall much more

forgiving.

 

My proposal would be to not add the parapet walls but use the new VW 2021

Sketchup Style direct modeling tools to extrude Faces from the base Solid.

Or from Boundaries of Faces that you define by drawing lines on the Faces.

That makes sure you have the same vertices.

That is basically how a Mesh Modeler would work)

(It would work the same in previous VW versions, just not that comfortable)

If you are fine with Generic Solids geometry at the end, I would also deleting

modeling history after every few steps, to avoid complex hierarchies.

[Tom W.]

@line-weight It looks fabulous. You gotta love a victorian railway viaduct. What kind of survey are you working from? Thanks for sharing

[line-weight]

6 minutes ago, zoomer said:

 

 

It is much easier and less critical if objects are overlapping quite a bit.

And if they are nearly matching but not exactly, Boolean Operations can fail.

Like when Faces are nearly coplanar overlapping.

 

It was a pain in Microstation but I always thought VW is overall much more

forgiving.

 

My proposal would be to not add the parapet walls but use the new VW 2021

Sketchup Style direct modeling tools to extrude Faces from the base Solid.

Or from Boundaries of Faces that you define by drawing lines on the Faces.

That makes sure you have the same vertices.

That is basically how a Mesh Modeler would work)

(It would work the same in previous VW versions, just not that comfortable)

If you are fine with Generic Solids geometry at the end, I would also deleting

modeling history after every few steps, to avoid complex hierarchies.

 

Thanks @zoomer

 

That makes sense - but there are a few reasons why it's difficult just to extrude faces in this particular model, which are a bit hard to explain but come down to the fact the viaduct structures each follow a very gradual curve both in plan and in elevation, while the parapets and other things need to follow the same curve but with their walls and other details "vertical" if you see what I mean. As described in other threads a while back, I already had a lot of problems constructing them as truly curved objects, so they have ended up being made by lofting a profile along a segmented path (with the segments hopefully not discernable in the final renderings).

 

This means that the base viaduct objects are faceted ... although I remain a little confused about what the facets actually are (for example, when I extract some of those faces, I end up with something that is not just a simple 3d polygon).

 

Anyway, I have had various problems trying to extract edges and faces, in order to use them to build up parapets (or provide the paths for extrudes-along-paths which are also driving me mad) and I've in some cases done what you suggest, drawing a polygon directly onto a face, to avoid these.

 

It is unfortunate that VW can only tell me that it can't compute something rather than highlighting where the problem is, because if I knew exactly where the problem was I could focus on that and try and redraw that particular part in a different way. But as it is, it seems I have to go through an extremely time-consuming process of testing every single element to find out if it's the one I need to redraw.

[line-weight]

14 minutes ago, Tom W. said:

@line-weight It looks fabulous. You gotta love a victorian railway viaduct. What kind of survey are you working from? Thanks for sharing

 

Thanks. In fact this particular project is one I'm doing for my own interest; it's a kind of built history project concentrating on my local area ... so most of the surveying is done manually by me with a tape laser measure, or interpolated from aerial photos etc.

 

If it were a fee-earning project 'd have resigned from it by now because it's all incredibly time-consuming! Doing the viaducts is quite challenging and hopefully good for my 3d modelling skills in the long term.

[Tom W.]

That's amazing you've surveyed that yourself. Very impressive. I thought maybe you were working off a shapefile which look amazing on the VW videos but not sure how/where you get hold of them (ProMap weren't v helpful)

[line-weight]

21 hours ago, Tom W. said:

shapefile which look amazing on the VW videos

which are these, by the way?

 

[Tom W.]

I can't find the videos now unfortunately. The way I understand it they are like OS maps but the polygons have data attached to them which gives you lots of options for modification, one of which is to make massing models from the polygons based on building elevation data contained in the file. This thread talks a little bit about it:

 

 

[line-weight]

Ah yes. I think they are quite crude in reality, and would be of little use for building anything in detail.

 

The amount of information in google's 3d view (at least of london) is fairly impressive. It's also quite crude in a way but is surprisingly helpful in making sense of things like roof forms which might be ambiguous on a pure aerial view (even a high res one).

 

 

[Tom W.]

No probably not v precise but looked a pretty cool way to populate a 2D map with massing models with a couple of clicks. If your city has only flat-roofed buildings that is... And you could find someone to supply the shape file in the first place...

But how were you acquiring height info for your buildings/structures? 

 

[line-weight]

5 minutes ago, Tom W. said:

No probably not v precise but looked a pretty cool way to populate a 2D map with massing models with a couple of clicks. If your city has only flat-roofed buildings that is... And you could find someone to supply the shape file in the first place...

But how were you acquiring height info for your buildings/structures? 

 

Some of it can be found looking through planning application drawings (although a judgement has to be made about how reliable they are).

Counting brick courses (google streetview works for this, as well as real life!) can be surprisingly reliable/accurate.

Other than that, standing on the street with a laser measure.

Then a fair bit of interpolation.

Edited December 1, 2020 by line-weight