Does Zero-Thickness-Geometry exist?

[mike miller]

Does it?

[matt]

Does it?

Oh, good one. You know I can't resist. I'm going to have to find a troll emoji now.

Here's the thing. You can't do it in a single body because SW and/or Parasolid made the decision to not allow it for the sake of integrity of the data. For every case you can name there is a way to achieve it using multiple bodies.

And the age old argument against it - what do you get if you make zero thickness geometry on a mill or lathe? Scrap.

I say just get over it. There are other battles you might win, this isn't one of them. Plus, this isn't even a battle worth fighting. Even if you won (or whoever is making the argument), what would you gain? Somewhere between very little and Zero.

I still get caught by it now and then, and it's never geometry that I intend to make or need to make. Sometimes it takes a while to find, even when I know what I'm looking for. Better diagnostic tools to identify it when it happens or is the cause of an error would be helpful, but if they started allowing zero thickness on parts, I (and a whole lot of other people) would probably start making more junk parts.

[mike miller]

Does it?

Oh, good one. You know I can't resist. I'm going to have to find a troll emoji now.

Here's the thing. You can't do it in a single body because SW and/or Parasolid made the decision to not allow it for the sake of integrity of the data. For every case you can name there is a way to achieve it using multiple bodies.

And the age old argument against it - what do you get if you make zero thickness geometry on a mill or lathe? Scrap.

I say just get over it. There are other battles you might win, this isn't one of them. Plus, this isn't even a battle worth fighting. Even if you won (or whoever is making the argument), what would you gain? Somewhere between very little and Zero.

I still get caught by it now and then, and it's never geometry that I intend to make or need to make. Sometimes it takes a while to find, even when I know what I'm looking for. Better diagnostic tools to identify it when it happens or is the cause of an error would be helpful, but if they started allowing zero thickness on parts, I (and a whole lot of other people) would probably start making more junk parts.

Trolling? I'm the troll here! What are you doing stealing my thunder!?!? Get back under that bridge!!

It keeps coming up on "the other forum" and I thought I'd give those guys a chance to share their heart. Actually, I've got a new argument I want to test........

[Lapuo]

Good question
I will try to explain my opinion best as i can , but please forgive me for my English , because i am "self learned" so i am having a hard time to write what i mean

I will take example of hole which is tangent to a edge.
I understand that you cant have zero material between hole and edge of a part , although you can drill (or mill) that hole on that exactly position where you would put dimensions in SW if zero thickness is allowed. Deformation of material would "destroy" that edge where hole is tangent to edge.
But i will compare this with plate 1 meter long which is 100% flat (and that is what you will never accomplish in real life, you will always have some "deformation" which will affect flatness )

So i am really having trouble with understanding why zero thickness is not allowed in SW. There is a ton of stuff you can model in SW but you cant make it in real life , so why zero thickness geometry is a different?

[matt]

So i am really having trouble with understanding why zero thickness is not allowed in SW. There is a ton of stuff you can model in SW but you cant make it in real life , so why zero thickness geometry is a different?

Solid models are all about volume, and one of the ways to identify a solid body is as a "contiguous" or uninterrupted volume. This definition becomes ambiguous when you have a condition with an edge or a point that has no thickness. Thickness is kind of a basic definition of a solid.

In short, it is harder to guarantee valid solid data if you allow zero thickness points and edges.

With your example of the hole tangent to an edge, there is always a tolerance on the hole size and the placement, so use that to make something that the computer can work with.

I don't agree or disagree with the policy to disallow zero thickness conditions. I merely accept it. That's the way it is. They have a reason for it which is reasonable. I keep getting my work done. It would be easier - I'd get fewer errors - if they allowed it, but I want valid geometry more than I want to get rid of a minor inconvenience.

[AlexLachance]

Good question
I will try to explain my opinion best as i can , but please forgive me for my English , because i am "self learned" so i am having a hard time to write what i mean

I will take example of hole which is tangent to a edge.
I understand that you cant have zero material between hole and edge of a part , although you can drill (or mill) that hole on that exactly position where you would put dimensions in SW if zero thickness is allowed. Deformation of material would "destroy" that edge where hole is tangent to edge.
But i will compare this with plate 1 meter long which is 100% flat (and that is what you will never accomplish in real life, you will always have some "deformation" which will affect flatness )

So i am really having trouble with understanding why zero thickness is not allowed in SW. There is a ton of stuff you can model in SW but you cant make it in real life , so why zero thickness geometry is a different?

Sorry for the off-topicness, but I wanted to congratulate you on teaching yourself english. You're doing prety good!

[Lapuo]

Good question
I will try to explain my opinion best as i can , but please forgive me for my English , because i am "self learned" so i am having a hard time to write what i mean

I will take example of hole which is tangent to a edge.
I understand that you cant have zero material between hole and edge of a part , although you can drill (or mill) that hole on that exactly position where you would put dimensions in SW if zero thickness is allowed. Deformation of material would "destroy" that edge where hole is tangent to edge.
But i will compare this with plate 1 meter long which is 100% flat (and that is what you will never accomplish in real life, you will always have some "deformation" which will affect flatness )

So i am really having trouble with understanding why zero thickness is not allowed in SW. There is a ton of stuff you can model in SW but you cant make it in real life , so why zero thickness geometry is a different?

Sorry for the off-topicness, but I wanted to congratulate you on teaching yourself english. You're doing prety good!

Thank you
I am better at understanding it than writing or speaking but i am trying my best

[matt]

Here, tell me what this is:

image.png

This is a case where SolidWorks has a switch (verification on rebuild) that allows you to create bad geometry. The inside of the box pokes through the outside of the box. And some people still get surprised when SolidWorks gives them impossible geometry.

Do you want to get this kind of error? No, but some people still throw the switch. They prefer ease of use over quality data.

SolidWorks usually checks every face against every adjacent face to see if you have a self intersecting model. The Verification On Rebuild switch forces SolidWorks to check every face in a model against every other face. This takes way more time, but it helps guarantee a good model. Some people turn it off.

[AlexLachance]

Here, tell me what this is:
image.png

This is a case where SolidWorks has a switch (verification on rebuild) that allows you to create bad geometry. The inside of the box pokes through the outside of the box. And some people still get surprised when SolidWorks gives them impossible geometry.

Do you want to get this kind of error? No, but some people still throw the switch. They prefer ease of use over quality data.

SolidWorks usually checks every face against every adjacent face to see if you have a self intersecting model. The Verification On Rebuild switch forces SolidWorks to check every face in a model against every other face. This takes way more time, but it helps guarantee a good model. Some people turn it off.

Geez Matt, are your books in Audio versions? I'm not much of a reader but I sure wish I had that sort of knowledge.

[matt]

Geez Matt, are your books in Audio versions? I'm not much of a reader but I sure wish I had that sort of knowledge.

Thanks, but sorry, no audio. It's very long (~1200 pages), you've got to be really dedicated to read it. Lots of pictures, though. Black and white in print, but color in the ebooks. They never translated them to French either. One year they made a short version in Chinese.

[mike miller]

Geez Matt, are your books in Audio versions? I'm not much of a reader but I sure wish I had that sort of knowledge.

Thanks, but sorry, no audio. It's very long (~1200 pages), you've got to be really dedicated to read it. Lots of pictures, though. Black and white in print, but color in the ebooks. They never translated them to French either. One year they made a short version in Chinese.

I'm reading sections in it every day. It's well worth it. I learn something new in every chapter.

[mike miller]

Good question
I will try to explain my opinion best as i can , but please forgive me for my English , because i am "self learned" so i am having a hard time to write what i mean

I will take example of hole which is tangent to a edge.
I understand that you cant have zero material between hole and edge of a part , although you can drill (or mill) that hole on that exactly position where you would put dimensions in SW if zero thickness is allowed. Deformation of material would "destroy" that edge where hole is tangent to edge.
But i will compare this with plate 1 meter long which is 100% flat (and that is what you will never accomplish in real life, you will always have some "deformation" which will affect flatness )

So i am really having trouble with understanding why zero thickness is not allowed in SW. There is a ton of stuff you can model in SW but you cant make it in real life , so why zero thickness geometry is a different?

My opinion on ZTG..........it doesn't exist except in the mind of AD and PTC fanboys.

Think about it. Can you shine a light beam through a ZTG corner? Yes or no? Either answer proves it is either has a gap or is combined.

[HerrTick]

The only semi-legit cases for t=0 geometry I can think of...

• sheet metal flanges bent to where corners meet

• spring coils

Others???

[Frederick_Law]

A very simple ZTG:
Two pipe/round bars touching.

Two volume share a line or point create ZTG in SW.
In real life, there is no zero thickness zero point.
Sharpest point is one molecule wide.
One day we might get it to one electron, proton, bison but still not zero.

[Glenn Schroeder]

A very simple ZTG:
Two pipe/round bars touching.

. . . and that can be modeled in SW just fine.

[mattpeneguy]

A very simple ZTG:
Two pipe/round bars touching.
ZTG-02.jpg

Two volume share a line or point create ZTG in SW.
In real life, there is no zero thickness zero point.
Sharpest point is one molecule wide.
One day we might get it to one electron, proton, bison but still not zero.

I don't know...if your margins are measured in bisons, you may need to look at adjusting your tolerances...:

image.png (145.21 KiB) Viewed 10280 times

[HerrTick]

I don't consider the two-cylinder shape a genuine zero-thickness problem. It's just two pieces.

[mike miller]

This is a classic example of ZTG.



You can't make it in real life. Change my mind.

[jcapriotti]

This is a classic example of ZTG.

2021-03-29 20_18_52.jpg

You can't make it in real life. Change my mind.

Does any CAD software handle this?

[Lapuo]

This is a classic example of ZTG.

2021-03-29 20_18_52.jpg

You can't make it in real life. Change my mind.

You can't make geometry like so(material will deform so you won't have zero material at tangent point) , BUT you can drill hole at exact location which SW does not allow.
You can make ton of things in SW , which you can't in real life (but SW allows it anyway )

Does any CAD software handle this?

Inventor does, i believe.

[SPerman]

I get why cad systems can't handle ZTG. But why can't I create a section view (in the model) that results in theoretical ZTG? I'm pretty sure other software can handle this, but I think other software handles section views differently.

[bradb]

My biggest complaint regarding ZTG and SW is if I do something that creates the condition and the software is smart enough to recognize it, then show me where it is so I can fix it. It's not always as obvious as the previous example.

[Frederick_Law]

I don't consider the two-cylinder shape a genuine zero-thickness problem. It's just two pieces.

Most ZTG can be 2 pcs.

Self touching is forbidden in SW

[Matthew]

This is a classic example of ZTG.

2021-03-29 20_18_52.jpg

You can't make it in real life. Change my mind.

If the toolroom at work wasn't so swamped, I would submit a job for the wire EDM to see just how close they could actually make that.

[Jim Elias]

ZTG is a curse. Sure you can create it in some systems -- but they then choke when attempting to use ZTG-intersected surfaces downstream. Thickens, offsets, derived geometry etc. will at some point fail and you can then slog back to the scene of the crime.

[Rob]

This conversation reminds me of post match pints in the pub.

If anyone was discussing if it was offside, or a red card, etc? my mate JohnJohn would shut them down.

It obviously [WAS|WASN'T] offside because that's the decision of the referee.

The question is moot. No matter your opinion - the parasolid kernel says it doesn't exist.

[Jim Elias]

My opinion on ZTG..........it doesn't exist except in the mind of AD and PTC fanboys.

Which brings us to another ZTG Fun Fact... IV and Creo apparently have different internal mechanisms for dealing with ZTG. If you interoperate a Creo ZTG model into IV, you can do stuff with it which IV won't allow on its native ZTG models. And so on. Ambiguous, flukey, mercurial... ZTG "fake math" sums up what CAD isn't supposed to be.

If you give ZTG-containing models to clients or other contractors, you *will* end up looking like a noob before the day is out.

If there was a ZTG switch in IV and Creo, I would leave it shut off. 'Nuff said.

[Tera]

Everyone is talking about creating a body that causes a ZTG. I can understand that.
I simply want to section view an assembly and Solidworks refuses because of a ZTG. It's just a section view and nothing more. I need to view the result from that position. That's all.

Is it that hard for SW?

[Jim Elias]

Everyone is talking about creating a body that causes a ZTG. I can understand that.
I simply want to section view an assembly and Solidworks refuses because of a ZTG. It's just a section view and nothing more. I need to view the result from that position. That's all.

Is it that hard for SW?

The SW section view is basically a temporary assembly cut. It has to calculate and generate geometry in order to allow you to do measurements off the theoretical section edges, so the view creation is very likely constrained by the modeling rules for that reason. (Same thing in drafting.)

If you do a "graphics-only section," ZTG is ignored and you can see the section, but there are no section edges generated for measuring.

I've been doing the 0.001mm offset workaround for years, it doesn't bother me.

That said, NX will allow ZTG in true measurable section views, and generates the sections much faster. But there are limitations compared to SW (you can't do a "pie-wedge" combo section view, for example)

[Rob]

This is the only part I struggle to model in SW that has ZTG IRL.

[Merovingien]

This is the only part I struggle to model in SW that has ZTG IRL.

where ?

[jcapriotti]

This is the only part I struggle to model in SW that has ZTG IRL.
image.png

I see a gap......

[jcapriotti]

Sheet metal does allow this in some cases.

[Rob]

There is a slightly rounded edge to the bar, but the curl is forged pretty close

image.png (124.16 KiB) Viewed 9982 times

[Merovingien]

i'm sure, by forcing a little, we can pass through it with a 0.01 mm titane sheet...

Rob :
...that has ZTG IRL

Lies and Slander (Quote from Futurama)

Disbeliever.png (96.08 KiB) Viewed 9975 times

[jcapriotti]

There is a slightly rounded edge to the bar, but the curl is forged pretty close

image.png

I see a bigger gap than I first thought

Seriously though, in this example there is no such thing as ZTG.....there is a gap there. You can keep making the gap smaller and smaller. You could, with enough pressure and heat fuse it together, in which there would be no gap but instead merged material so still no ZTG. I guess true ZTG would be a black hole?

I looked at the hole on edge example in Inventor and it does allow the creation while SolidWorks does not. IV is splitting the face which makes me think it's detecting the ZTG and slightly cutting thru.......maybe at like .00000001 or something.

[Rob]

i'm sure, by forcing a little, we can pass through it with a 0.01 mm titane sheet...
Lies and Slander (Quote from Futurama)

Disbeliever.png

Maybe it has shrunk back a bit, but it was definitely touching whilst it being was made

[Roasted By John]

There was a big discussion in the Old Forum about ZTG - is some instances it might make a difference, depending on what you're designing, such as laser etc..

If it's a manufactured component, then my question was always, "Can you manufacture it as modeled", and if the answer is no, then why need it. For models you can always offset a line by .00001" - there is no way you can measure that with general shop inspection tools, plus change of thermals will easily move molecules more then that amount, again why do you need to model it?

[MJuric]

This is a classic example of ZTG.

2021-03-29 20_18_52.jpg

You can't make it in real life. Change my mind.

Theoretically you can't make anything you model in any CAD system in real life....that's why everything has a tolerance on it. Go out and make a perfectly sized, perfectly square, perfectly flat on all sides 1" cube. Not going to happen ever. Yet in any CAD system it is the most basic of functions.

We all know that there's no such thing as "Zero Thickness Geometry" what I never understood about this was why can't you just make the software realize that when you have a theoretical ZTG you really have an infinitesimally small break and just put the damn hole where I told you to?

This isn't really a question of whether ZTG exists or doesn't it's a flaw in how the software identifies a ZTG.

I'm with Matt in that it's not a big deal and you deal with it. But seems to me that it wouldn't be a major mountain to over come to make it work. Maybe Fredrick can comment on this but I thought that IV did this just fine. I could be mistaken it's been 5 years since I last used IV but I could have sworn I never saw that error and created that geometry all the time.

[MJuric]

There was a big discussion in the Old Forum about ZTG - is some instances it might make a difference, depending on what you're designing, such as laser etc..

If it's a manufactured component, then my question was always, "Can you manufacture it as modeled", and if the answer is no, then why need it. For models you can always offset a line by .00001" - there is no way you can measure that with general shop inspection tools, plus change of thermals will easily move molecules more then that amount, again why do you need to model it?

Because models are a theoretical perfect. I absolutely HATE having to fudge numbers but in this case that's exactly what needs to be done. 99.99% of the time it's a complete non issue, move the hole, change the hole size etc and in most cases it's a bad design to have ZTG. But every once in a blue moon that's actually what you want and I absolute abhor having to fudge it because the software doesn't like it. I do it, I accept it....doesn't mean I have to like it :-)

[MJuric]

There is a slightly rounded edge to the bar, but the curl is forged pretty close

image.png

I see a bigger gap than I first thought

Seriously though, in this example there is no such thing as ZTG.....there is a gap there. You can keep making the gap smaller and smaller. You could, with enough pressure and heat fuse it together, in which there would be no gap but instead merged material so still no ZTG. I guess true ZTG would be a black hole?

I looked at the hole on edge example in Inventor and it does allow the creation while SolidWorks does not. IV is splitting the face which makes me think it's detecting the ZTG and slightly cutting thru.......maybe at like .00000001 or something.

Should have read this first before posting. I though IV did ZTG without error and if I recall it does exactly that, makes a tiny little break. It's the software doing a work around rather than the designer having to do it...I prefer that.

[jcapriotti]

We all know that there's no such thing as "Zero Thickness Geometry" what I never understood about this was why can't you just make the software realize that when you have a theoretical ZTG you really have an infinitesimally small break and just put the damn hole where I told you to?

That opens a big question.....which side of the face should the software "adjust" to? Outside and break the face or inside and leave it whole? Inventor breaks the face.

This isn't really a question of whether ZTG exists or doesn't it's a flaw in how the software identifies a ZTG.

I wouldn't call it a flaw. The software recognizes the condition and tells you. It just doesn't allow you to create that condition. Inventor recognizes it as well and makes a predetermined decision to offset it a small amount.

[MJuric]

Does it?

Isn't every corner, edge etc essentially ZTG? At some point doesn't the material in the model trail off into infinity? seems to me that all that the software does is say "Material stops at this point and ends at this point and does so with in the precision of it's calculation ability.

At some point you hit zero or less and you stop. I don't see why ZTG of two intersecting features should be any different. You have material all the way up to this point and you have a break after that, with in the precision of the software. If you land exactly on zero you make the decision whether zero is a break or zero is geometry.

[MJuric]

We all know that there's no such thing as "Zero Thickness Geometry" what I never understood about this was why can't you just make the software realize that when you have a theoretical ZTG you really have an infinitesimally small break and just put the damn hole where I told you to?

That opens a big question.....which side of the face should the software "adjust" to? Outside and break the face or inside and leave it whole? Inventor breaks the face.

Isn't this already done via some mechanism limited by the precision of the software? As I mentioned in the other post this same thing has to be done with any surface, corner edge. At some point you exceed the precision of the software and you have to round the number...up or down, break or no break, is material here or not.

This isn't really a question of whether ZTG exists or doesn't it's a flaw in how the software identifies a ZTG.

I wouldn't call it a flaw. The software recognizes the condition and tells you. It just doesn't allow you to create that condition. Inventor recognizes it as well and makes a predetermined decision to offset it a small amount.

I would agree it's not a flaw. Maybe a preference. I would much rather have a program that is consistent that says "If you're directly on the line we will always break it" than have to force the designer to decide and or worse yet actually force the designer to design to the wrong number.

[Roasted By John]

There was a big discussion in the Old Forum about ZTG - is some instances it might make a difference, depending on what you're designing, such as laser etc..

If it's a manufactured component, then my question was always, "Can you manufacture it as modeled", and if the answer is no, then why need it. For models you can always offset a line by .00001" - there is no way you can measure that with general shop inspection tools, plus change of thermals will easily move molecules more then that amount, again why do you need to model it?

Because models are a theoretical perfect. I absolutely HATE having to fudge numbers but in this case that's exactly what needs to be done. 99.99% of the time it's a complete non issue, move the hole, change the hole size etc and in most cases it's a bad design to have ZTG. But every once in a blue moon that's actually what you want and I absolute abhor having to fudge it because the software doesn't like it. I do it, I accept it....doesn't mean I have to like it :-)

If that is the case then, if the part would be manufacturable - ZTG would be water tight, is that a correct way of looking at it??

[mike miller]

Question for those who think ZTG exists: Can a light beam pass through a ZTG area? Yes or no?

Also, ZTG is different than a theoretically perfect cube. ZTG cannot even be accurately displayed on a screen (is it transparent or opaque?), but a perfect cube can. CAD does not have tolerances. (Please put down that pitchfork, it makes nervous.) CAD does truncate decimals, but it does NOT just throw geometry out there willy-nilly and call it perfect.

[jcapriotti]

Question for those who think ZTG exists:

Can a light beam pass through a ZTG area? Yes or no?

I'll answer your question with a question. Is Schrödinger's cat dead or alive?

[MJuric]

There was a big discussion in the Old Forum about ZTG - is some instances it might make a difference, depending on what you're designing, such as laser etc..

If it's a manufactured component, then my question was always, "Can you manufacture it as modeled", and if the answer is no, then why need it. For models you can always offset a line by .00001" - there is no way you can measure that with general shop inspection tools, plus change of thermals will easily move molecules more then that amount, again why do you need to model it?

Because models are a theoretical perfect. I absolutely HATE having to fudge numbers but in this case that's exactly what needs to be done. 99.99% of the time it's a complete non issue, move the hole, change the hole size etc and in most cases it's a bad design to have ZTG. But every once in a blue moon that's actually what you want and I absolute abhor having to fudge it because the software doesn't like it. I do it, I accept it....doesn't mean I have to like it :-)

If that is the case then, if the part would be manufacturable - ZTG would be water tight, is that a correct way of looking at it??

I really don't think one way or the other is "correct", whether there is an infitesimally small thickness of material there or an infintesimally small gap. The software would have to pick one or the other...I don't really care which. Having the "error" of having a gap when you want a piece of material is more acceptable to me than having a hole at .0000001 when it should be at zero. Again probably personal preference.

[MJuric]

Question for those who think ZTG exists: Can a light beam pass through a ZTG area? Yes or no?

Also, ZTG is different than a theoretically perfect cube. ZTG cannot even be accurately displayed on a screen (is it transparent or opaque?), but a perfect cube can. CAD does not have tolerances. (Please put down that pitchfork, it makes nervous.) CAD does truncate decimals, but it does NOT just throw geometry out there willy-nilly and call it perfect.

I see this as a simple question of "Does material exist or does it not exist". You can't have material and not have material at the same time. Thus the discussion of ZTG is not about does material exist and not exist at the same time, but rather what choice/Truncation/Rounding does the software do when faced with "0".

The same question exists with every piece of geometry. At some point the software has to say "Nope, no material here". So is there material at 1.0, yes, 1.00, yes, 1.0000000000000000000000000000000....no idea because we rounded at the 9th decimal place and whatever the rounded number is that determines material or no material. The same thing applies to 0. Is there material at .01, yes, .0001, yes, .0000000000000000000000000001, no idea because we rounded at the 9th decimal place. In this case it would be rounded to zero and if the software says zero has material then it does, if it says it doesn't then it doesn't.

One would think zero would mean it doesn't, which is what IV does, creates a break.

[mike miller]

I'll answer your question with a question. Is Schrödinger's cat dead or alive?