## The Ultimate Resource for TurboCAD Knowledge

Remember to do a search before you ask a question!!!

Sort Of Trunking Solid Summing and Fillet Problem.
July 24, 2019, 06:58:01 AM
It is very difficult to apply bandages to subtract the lower corners from this type of object.

Logged

July 24, 2019, 07:36:21 AM
In Another attempt to make sharp edges more able to be rounded I tried creating a 3D polyline up on of the vertical edges then sweeping a circle along the 3D polyline. I then attempted to subtract the swept circle solid from the features of the main solid shape.

I found that I could not subtract the swept cylinder corresponding to the section of the the lofted Cyan part from the lofted Cyan part!

Why can I not use a boolean operation to subtract a swept part from a lofted part?!
« Last Edit: July 24, 2019, 07:41:24 AM by MicheliDE »

Logged

July 24, 2019, 08:33:56 AM
Then I took the solid into a subractive space by designing a box around the entire model and subtracting the model out the summed solid. This created even more problems however there is a clue. When I subtracted the model out of the box Turbocad changed the mid section which is the long section that has the same profile on both sides into a hollow object!

I went back and looked at the midsection before subtraction and it is infact a solid!

With the Cavity sliced into left and right sides I tried quick pulling to complete what should have been the subtraction from the mid section and it wouldn't remove it with a quick pull! Could it be that the ends are also surfaces? Did this occur during adding the top and bottom together? How did this get automatically converted to a surface when it started as a solid?

And I have to go through this and I don't even know if I am going to be able to fillet in the negative space when I wasn't able to fillet in the positive space.

Logged

July 24, 2019, 09:29:30 AM
The Mid solid is a hollow solid that has surfaces on the ends! How does two identical closed polylines form a hollow loft with surfaces on the ends?

With a relofted midsection I tried multi-add and the solid still would not add all at once.

This time after the multi-add was used it generated an error and would not let me use boolean add afterward. This is another clue of why I am having issues.
« Last Edit: July 24, 2019, 09:42:07 AM by MicheliDE »

Logged

July 24, 2019, 09:54:32 AM
And so because of the fixed mid section I once again tried filletting from the bottom to the top. Just as previously I noticed that the fillet tool is creating different kinds of ends on two of the junctions. You can see them here as the ends of the fillet tool go past the Blue closed polygon profile. Two are pointy on one side and four are pointy in the middle and one of the four that are pointy in the middle has four dots that could be filleted on it. That is so weird and I know that the two pointy fillets will not produce a round corner. All of these fillets were made with 0.040" Start radius and the corner fillet button pressed.

Logged

July 24, 2019, 10:02:26 AM
Going back to the Negative space attempt. I can no longer subtract the summed solid from the surrounding box that I created, error. I also cannot slice the summed solid in either the part or ACIS form. I cannot subtract either part or ACIS from the surrounding box.

Logged

July 24, 2019, 10:10:03 AM
And so even though I replaced the mid section with a new loft that was a solid it has still caused a problem when trying to subtract from the surround box solid that I created.

I am able to subtract all the other lofts from the surrounding box except the mid section. I am also still unable to slice the surrounding box with the top and bottom lofts already subtracted.

The bounding box and all of the lofts do slice if I separate the slice into top, midsection, and bottom and the bounding box.

Another file is attached showing current step.
« Last Edit: July 24, 2019, 10:21:09 AM by MicheliDE »

Logged

July 24, 2019, 10:29:30 AM
Then there was an insufficient memory error when rendering all these pieces and turbocad crashed. Good thing i saved before trying to subtract.

Also In the negative space I can see that there are problems with some of the edges. The edges are visibly thicker than others and they pixelate more when you are moving where you are viewing the solid from.

How is it that I was able to subtract out the top and the bottom before when my mid section had a problem but now I can't subtract out the top. This would mean that the way that the top was added even though neither method generated an error is now causing me a subtraction problem!

So now I have to take the subtraction from the surrounding box with the bad mid section and attempt to take the new mid section and subtract it from it.

Neither mid section will subtract out of the surrounding box. Now I have to slice the box and pull the box from one side to the other!
« Last Edit: July 24, 2019, 11:08:08 AM by MicheliDE »

Logged

July 24, 2019, 11:19:29 AM
Dean,
I haven't look at the drawing yet that you posted in Reply#49 but my first guess is that most of these problems that you are having might be that TC is not internally calculating the object as an ACIS solid object even though it's being reported as one, I know, a bit quirky
I haven't read all of your posts but did you try starting over and make sure to use compound profiles on any open lofted profiles.
I would recommend staying away from the Quick Pull tool when working with a complex object, It's also a bit quirky....

Logged
"What do I know, I'm still learning"..
Windows 7 64 bit HP TurboCad Platinum >2020- 2019-2018-2017-21-20-19 and 18.Using only LightWorks (NO Redsdk Plugin installed), Anilab lab 5.  HP all in one  23" Touch smart screen with Nvidia Geforce 210, open gl 3.3. Intel Core i5 ,2.27GHz,

July 24, 2019, 11:26:10 AM
What I do is I save files at critical points so I can go to and from any point to any other point and introduce objects from another drawing. This usually helps to get further.

At this point was unable to subtract any more from the two halves. the mid section will no go away and I cannot even slice the two halves either.

I will go back to the lofts with the new mid section profiles to re build all the profiles again.

Logged

July 25, 2019, 05:34:06 AM
Thanks for your help everyone I was able to move forward on this.

Logged

July 26, 2019, 02:57:37 PM
Dean, if you pass a section through one of the circle-to-polygon transition lofts of the file you originally posted, there's a convex region of the surface that restricts the cross-section area.  I experimented with the shape, rebuilding the main body and reconstructing the transition with a faceted loft that's had tapered fillets applied, blended into the edges of the main body that have .04" fillets.  The fillets have been applied sequentially and the order is slightly different to each. One has had the tapered fillets applied in the same operation as the .04" edges that they flow into, the other has had those fillets applied in separate operations.  The fillets can't be applied in one global operation, successive fillets are dependent on those preceding.  I built the models in V2015, V21 got crashier and more frustrating the further in I got.
Their part trees give some insight into the filleting sequences and the profiles I built the lofts with are in a hidden category.
« Last Edit: July 26, 2019, 03:05:36 PM by murray dickinson »

Logged

July 28, 2019, 05:53:34 AM
Hi Murry,

I attempted uploading the file that you attached in TC2019 and it sent the program into a not responding state for quite some time but it finally exited. It certainly seems like there is some trouble with this solid as you have found. I see that you are experimenting with a shape that does not have such a pronounced cylinder section. I think I am having edge problems that are being caused by problems resulting from snapping profiles to other profiles.

In the side by side profiles from the original file I have gone back and inspected each junction between the horizontal profiles and when they are 2D profiles they line up perfectly however as soon as I loft them there are slight gaps that are forming between the solids. I am not sure why this is but it requires zooming all the way into them to see them. This appears to be what is causing my filleting issues.

I think am experiencing the double edged sword of using guide lines in a loft. The result is that the guide lines help the solid become more mirror able however they seem to make the edges not square. Either I do not use guide lines and end up with a slight twist or I use guide lines and end up with unsquare edges. It seems that I can't add the guide lines in the corners of the lofted polygon either because that interferes with the profile.

What difficulty.

Logged

July 28, 2019, 08:45:16 PM
When objects are built and saved using the part tree, the file is a "recipe" for the part, and TC rebuilds it following the order and parameters of the part tree when it's opened.  That's what takes a while here, and by "a while", I mean that my uninspiring six-year 1.4GHz Celeron with the sole benefit of 8Gb RAM takes five minutes to rebuild it, but I'm old and patient, and that leads me to reply that I don't think it's got "troubles".  I rebuilt the basis of your part over your file, but saved the new construction.  I've just saved it back in 2015 to V21, and it opens with parts sound and part trees intact.  I think that the fairly acute relationship of circle to polygon in your original file puts a lot of expectation on the loft tool, even with guide curves.  My preference, which has seen TC give me pretty good returns over the years, is to loft through an intermediate profile or profiles, giving a more obvious indication of the way the loft tool should surface the shape.   Transitions between polygons and curved profiles are especially subjective, and I've a fair bit of sympathy for the programmers who create a kernel to try to satisfy as many people as possible, with a high possibility leaving many less than satisfied.

Logged

July 29, 2019, 02:14:18 AM
I had a simple design that was extremely slow to load and display on screen, eventually found that it was caused by audit history was checked. It had parttree parts as well. Uncheck all audit options and retry.

Logged
Daz
TCW V21, 2015-2020 PP, Animation Lab V5.2 & Redsdk enabled, LightWorks rendering mostly. Ender 3 3D printer.

July 29, 2019, 06:15:11 AM
I think that this the shape that is the difficulty here. Any processing miss step causes turbo cad to go slower. I was finding that the more that I tried to line up the profiles the slower it would go in some instance. I still believe that at maximum zoom turbocad is having trouble putting the four edges together when they are added from different solids. I have rebuilt this an inspected every corner before adding more then several times and despite snapping everything, using a snap line ruler that I made to maintain squareness in the object as a check, and redrawing completely from primitives I am still getting edges at maximum zoom. I have since increased the precision of turbocad to level 6.

I haven't yet added this computers profile to my description yet however this computer is a Intel I7, X920 2.0GHz, with 8GB ram, 64 Bit and an NVIDIA Quadro FX 1800M It is about 10 years old. I do have a new computer on the way. Could this be a video card problem?

Logged

July 29, 2019, 03:23:16 PM
Ever seen an error message that tells you "the ACIS solid is degenerated" when you try a boolean action?  That's what you're building into your part when you try to match edges and vertices exactly as you're doing.   In CAD-speak, "degeneracy" is coincident features: edges, vertices, faces.  Better practice (probability of success) is to overlap parts and boolean them against each other with intersections and subtractions so the program is determining the geometric relationship of coincidental features itself.

Logged

July 30, 2019, 05:38:47 AM
I think that this the shape that is the difficulty here. Any processing miss step causes turbo cad to go slower. I was finding that the more that I tried to line up the profiles the slower it would go in some instance. I still believe that at maximum zoom turbocad is having trouble putting the four edges together when they are added from different solids. I have rebuilt this an inspected every corner before adding more then several times and despite snapping everything, using a snap line ruler that I made to maintain squareness in the object as a check, and redrawing completely from primitives I am still getting edges at maximum zoom. I have since increased the precision of turbocad to level 6.

I haven't yet added this computers profile to my description yet however this computer is a Intel I7, X920 2.0GHz, with 8GB ram, 64 Bit and an NVIDIA Quadro FX 1800M It is about 10 years old. I do have a new computer on the way. Could this be a video card problem?

Dean

Can you confirm what you are trying to achieve with the node points from your last attached drawing?
I have an inkling that all these nodes should be positioned on the z axis in the same place and looking using 3D Spheres there is some discrepancy with Precision set to 10.

The top row (the Red Line) Green and Blue Spheres are different to Yellow and Orchid Z positions
The Bottom row (Green Line) the Wheat Sphere is different to Brown, Light Blue and Dim Grey z Positions.

Logged
Daz…

V2020 & V2019 Plat 64bit + REDsdk Plugin, Lenovo P72 Laptop, Window 10 Pro for Workstations, Intel Xeon E-2186 CPU @ 2.90 Ghz (6 cores/12 threads), 32GB RAM, 512GB & 1TB SSD's, Nvidia P5200 c/w Max=Q Design GPU, Display UHD 3840 x 2160 pixels

July 30, 2019, 06:13:07 AM
Ever seen an error message that tells you "the ACIS solid is degenerated" when you try a boolean action?  That's what you're building into your part when you try to match edges and vertices exactly as you're doing.   In CAD-speak, "degeneracy" is coincident features: edges, vertices, faces.  Better practice (probability of success) is to overlap parts and boolean them against each other with intersections and subtractions so the program is determining the geometric relationship of coincidental features itself.

Hi Murray,

The last several iterations of the the boolean actions did not report any ACIS solid degeneration. There were a few instances where I was intentionally overlapping the solids and in those instances I was getting shards left over when performing a subtraction from the solid. I'm not sure how this is possible however I have attached an image of this occurring. This was when I switched to not overlapping solids. In the instances that I was getting these shards I was unable to subtract out.

It seems that even though I am snapping all edges and building the profiles from primitives and strait polylines without moving any of them and isolating each profile in a tower after it is lined up on a plane that is perpendicular to all of the profiles there is still some error being introduced. I suppose I could go to precision 10 to see if it fixes the problem. When I inspected the way that all of the solids resulting from the profiles lined up before adding them together at maximum zoom they all lined up. After summing them together I zoomed in and looked at the edges and they looked like the previous image. I have not yet tried to fillet this solid yet but I suspect that the fillet tool will react differently to different edges and make some not filletable again, or not filletable again in the same way as the previous iterations.

Logged

July 30, 2019, 06:48:11 AM
I think that this the shape that is the difficulty here. Any processing miss step causes turbo cad to go slower. I was finding that the more that I tried to line up the profiles the slower it would go in some instance. I still believe that at maximum zoom turbocad is having trouble putting the four edges together when they are added from different solids. I have rebuilt this an inspected every corner before adding more then several times and despite snapping everything, using a snap line ruler that I made to maintain squareness in the object as a check, and redrawing completely from primitives I am still getting edges at maximum zoom. I have since increased the precision of turbocad to level 6.

I haven't yet added this computers profile to my description yet however this computer is a Intel I7, X920 2.0GHz, with 8GB ram, 64 Bit and an NVIDIA Quadro FX 1800M It is about 10 years old. I do have a new computer on the way. Could this be a video card problem?

Dean

Can you confirm what you are trying to achieve with the node points from your last attached drawing?
I have an inkling that all these nodes should be positioned on the z axis in the same place and looking using 3D Spheres there is some discrepancy with Precision set to 10.

The top row (the Red Line) Green and Blue Spheres are different to Yellow and Orchid Z positions
The Bottom row (Green Line) the Wheat Sphere is different to Brown, Light Blue and Dim Grey z Positions.

Hi Darrel,

Thanks for the analysis. Hmm yes I see that Z difference. It seems that I need more points to snap to on each z plane align up these solids. At this level of alignment I noticed that standard tools such as slice do not always slice directly to a plane. In some instances the slice was slicing half way between the facet surface and the work plane (as visible at maximum zoom). I was also noticing that I couldn't just pull the surface by a small amount to fix these kind of problems.

It seems that I must setup something more confining to snap to then a simple 2D alignment plane for this. I will try this  and get all these Z points onto the two planes and rebuild the profiles once again.

Logged

July 30, 2019, 07:12:19 AM
Thanks for the input! The next version of the solid parts will be built with three levels of z guide planes as in the "New Alignment Guides" image. The two points that go into air and do not intersect with a green profile cannot intersect with the green profile on the same XY plane, the solid must loft out at this point. I have previously aligned four of these wider points on this lone profile with a rectangle. Lets seem how this goes.

Logged

July 30, 2019, 08:51:11 AM
I would try to reduce the splits in profiles as it seems to improve filleting as attached shows all fillets applied @ 0.04...

Try turning the profiles so it's upright into the Z axis as you can then use the align tools.

Logged
Daz…

V2020 & V2019 Plat 64bit + REDsdk Plugin, Lenovo P72 Laptop, Window 10 Pro for Workstations, Intel Xeon E-2186 CPU @ 2.90 Ghz (6 cores/12 threads), 32GB RAM, 512GB & 1TB SSD's, Nvidia P5200 c/w Max=Q Design GPU, Display UHD 3840 x 2160 pixels

July 31, 2019, 02:44:03 PM
Hi Darrel,

When you say splits in the profiles do you mean the number of sections that the dots bisect in the image?

What do you mean by turning the profiles upright into the Z axis?

I have been able to sum the model to be contiguous without breaks now. I will be attempting to fillet tomorrow.

Logged

August 01, 2019, 03:17:01 AM
Hi Darrel,

When you say splits in the profiles do you mean the number of sections that the dots bisect in the image?

What do you mean by turning the profiles upright into the Z axis?

I have been able to sum the model to be contiguous without breaks now. I will be attempting to fillet tomorrow.

Hi Dean
Just quickly tried your "Pre-fillet" drawing and there are still problems with trying to fillet certain edges as per the area attached with a very small step where the curve meets the flat.

Z Axis as shown in the attached is better when using the align tools as using them in any other workplane causes problems.

Logged
Daz…

V2020 & V2019 Plat 64bit + REDsdk Plugin, Lenovo P72 Laptop, Window 10 Pro for Workstations, Intel Xeon E-2186 CPU @ 2.90 Ghz (6 cores/12 threads), 32GB RAM, 512GB & 1TB SSD's, Nvidia P5200 c/w Max=Q Design GPU, Display UHD 3840 x 2160 pixels

August 01, 2019, 05:09:19 AM
Dean,
Something is screwy with your profiles.
Can we try something.
In "world plan view" activate "plane by world" then draw all your profiles each of them on their own layer.

Then move them to their "Z" axis positions in a front view and post that file.

Logged
"What do I know, I'm still learning"..
Windows 7 64 bit HP TurboCad Platinum >2020- 2019-2018-2017-21-20-19 and 18.Using only LightWorks (NO Redsdk Plugin installed), Anilab lab 5.  HP all in one  23" Touch smart screen with Nvidia Geforce 210, open gl 3.3. Intel Core i5 ,2.27GHz,

August 01, 2019, 07:41:12 AM
Hi Darrel,

When you say splits in the profiles do you mean the number of sections that the dots bisect in the image?

What do you mean by turning the profiles upright into the Z axis?

I have been able to sum the model to be contiguous without breaks now. I will be attempting to fillet tomorrow.

Hi Dean
Just quickly tried your "Pre-fillet" drawing and there are still problems with trying to fillet certain edges as per the area attached with a very small step where the curve meets the flat.

Z Axis as shown in the attached is better when using the align tools as using them in any other workplane causes problems.

Ahh yes. I see that small edge left... I was having trouble adding together the large extruded cylinder and the loft from the polygon to the large circle to get a clean edge between the two and so I pulled the circle from the loft and that is where the alignment issue occured.  I'll need to reloft those pieces if a slight alignment can't fix the issue. I will try alignment tools first while the solids are not summed.

Logged

August 07, 2019, 08:29:53 AM
I'm still working on this project. I had to break away into something else for a bit. I seem to be doing better on this model.

Take a look at how TC applied these fillets. There were no edge problems ahead of applying the vertically as first the top down and then from the bottom up. The result is that I can't do the horizontal fillet because both of the sides end and do not go over the midsection that is round. I tried to bandage it however I cannot add a swept profile to fill in a rounded area horizontally. Any other ideas on what I might be able to do?

Logged

August 08, 2019, 04:24:17 AM
I'm still working on this project. I had to break away into something else for a bit. I seem to be doing better on this model.

Take a look at how TC applied these fillets. There were no edge problems ahead of applying the vertically as first the top down and then from the bottom up. The result is that I can't do the horizontal fillet because both of the sides end and do not go over the midsection that is round. I tried to bandage it however I cannot add a swept profile to fill in a rounded area horizontally. Any other ideas on what I might be able to do?

Can you post a file Dean so we can look at it..... before filleting if possible?

Logged
Daz…

V2020 & V2019 Plat 64bit + REDsdk Plugin, Lenovo P72 Laptop, Window 10 Pro for Workstations, Intel Xeon E-2186 CPU @ 2.90 Ghz (6 cores/12 threads), 32GB RAM, 512GB & 1TB SSD's, Nvidia P5200 c/w Max=Q Design GPU, Display UHD 3840 x 2160 pixels

August 13, 2019, 06:14:24 AM
I'm still working on this project. I had to break away into something else for a bit. I seem to be doing better on this model.

Take a look at how TC applied these fillets. There were no edge problems ahead of applying the vertically as first the top down and then from the bottom up. The result is that I can't do the horizontal fillet because both of the sides end and do not go over the midsection that is round. I tried to bandage it however I cannot add a swept profile to fill in a rounded area horizontally. Any other ideas on what I might be able to do?

Can you post a file Dean so we can look at it..... before filleting if possible?

Hi Darrel,

I had to power through this because I was on a deadline and I saved the file past this point and so I do not have it. I did resolve my issues but ended up having to use two fillet sizes and go in a specific order.

Thanks everyone for the help.

Best Regards,
Dean

Logged