Torus request

[[Ti...]]

Is a torus measurement being considered?

Half of our parts could benefit from this feature. Is there a manual way to create this type of feature, with capturing some type of average diameter and/or location. I've never played with a rotational section cut, but it seems this would create many section cuts that I would have to include in the user defined section to average out. And I'm not sure how I could get a single position from a torus like feature....yes, I know this is a bad GDT callout, but companies do call this out.

Thanks.

Edited June 12, 2023

[[Mi...]]

Hi Tim,

although GD&T inspections for tori may be introduced to Zeiss Inspect at some point in the future, there are no immediate plans to do this in one of the upcoming software versions. The required center point for a torus will be probably have to be the result of a Gaussing fitting algorithm, but this is not part of the software yet.

[[Be...]]

Hi Tim,

one further hint: If you are interested in such features based on specific geometries it is always helpful to have concrete real world data. This would help us in the future to develop an application which fits really your needs. Therefore I would kindly ask you to supply some real world data, may be together with some technical drawings. The data should be supplied via our profesional support or via your contact person at ZEISS. Of course, data is treated confidentially by us.

Regards,

Bernd

[[Ti...]]

Please sign in to view this username.

  That would be nice, but without the GDT it would just be nice to be able to capture a sizable average radius for a turned part with a rounded edge for instance without having to section it and create circles. There are times the circles are so small or limited in arc degrees that those don't suffice. I'm thinking a feature might be able to capture the rounded radius of a lathe part might be more robust.   We do have some location of a torus so that would be nice one day.

 

Please sign in to view this username.

 I have supplied some stuff to our contacts inside Zeiss, so we can do that. I may send something off here soon.

[[Ti...]]

I know you all said this isn't up for consideration soon. But what is this:

 

[[Ch...]]

This are the remains from our first try to implement a torus element that was taken a decade ago. Of course it would be better if these token would be invisible but it is not a critical bug at all, so do not expect that we fix it very soon. 

During this first try we have found out that we doesn't understood the problem and the needed use cases deeply enough. That is one of the reason that we always asking for real projects and inspection task to find out what is really needed.

Sorry, but this token doesn't mean that we have started to work on the torus topic since your last post.

 

Best regards

Christoph Schult

[[Ti...]]

Damn...I thought yall had something set to come out with this feature 🙂

 

I'll send in a torus example this week to my contact at Zeiss and with this request and an explanation. I ended up having to make my own torus by taking 12 slices of a donut shaped feature and averaging out the 12 circles to hone in on the centroid of the feature.

Edited November 6, 2023

[[Ja...]]

I think its hard to realise as i cant imagine there are many ' perfect' toroid geometries that are needing insoection.   Ive digitised / measured a lot of things across a lot of industries . I can only think of one or two very special cases where there was a torus .

The other option could be to treat the surfaces as a positional call out ?   Maybe datum axis of centre of torus? 

 

[[Ti...]]

We have plenty of torus features that could use the help of a torus feature.

Imagine an internal thread relief cutout on a lathe using a boring bar...you can easily measure that b/c it is cut as a cylinder. What if this was cut with an undercutting ball nose endmill like this:

https://www.harveytool.com/products/specialty-profiles/undercutting-end-mills

Now ask yourself on an internal feature with a CT scan, are you happy with a simple 100 deg arc section cut circle for a radius check and then a position check based on this simple circle, or if you section cut the other direction you get one slice at the pie for a location.

I am having to slice a 120 arc torus into 12-20 section cuts and create a circle from each section and then go into the user defined check to make an average location of all these circles to better represent the location of this feature....yes...I am well away this should be called out as a surface profile, i get it. But we deal with the customer print the best we can and they state a location for a technical non-feature of size without opposing sides to confine the feature which most standards say is required.

 

Now what about gathering a radius for a rounded edge on a lathe part?  Are you all simply doing a section cut and measuring the radius of a circle feature you create? What happens if this rounded edge is very small and a scan cannot get a good circle from a circle? Should you then pattern a section cut and end up with 10 section cuts and 10 circles to average, or would it be more simple to have a torus that can average the radius and if you need it do a min max?  Fillets and edge blends can be measured with a torus vs making a bunch of section cuts and making circles.  Am I crazy?

 

[[Ti...]]

Here is a torus a created manually....I am trying to determine if I need to section 20,15, or 10 circles to get a reliable location right now. Yes this should be a profile callout, but as I said, you can't control everything a customer throws at you.

 

 

 

[[Ja...]]

This is a tricky one indeed. The bottom of these cylinders with a rad and the outer rads pictured would be a quarter toroid shape, not a full torus.  So even if there was the gdant maths for toroids because you dont have over 50 percent of the feature the maths would be invalid anyway id suspect.

You are in the same issue with circles as you know ..if there is a quarter circle the size and position is going to be completely unstable.

There are tools in the software for cutting multiple circles round a curve and calculating the radius and also the radius delta to get entry and exit points , not gdant of course but something else to bear in mind.

It is obviously difficult if what youve been asked to do on a drawing is not possible to the standards but that does appear to be where you are at , so you can only do your best to approximate.

Frankly id look to calculate it in the round about ways you are trying and ive mentioned but also throw in the profile to try to demonstrate that it is ultimately the correct way to control it and try to draw a discussion out about it..but im probably preaching to the converted !

 

[[Ti...]]

Please sign in to view this username.

Please sign in to view this username.

Please sign in to view this username.

Please sign in to view this username.

 I know James is no longer there, but I attached him regardless....

I'd like to send some projects we have with a torus feature to Zeiss. Where should I send them? They are large files. We could really use a torus feaure if it is done correctly. Lathe parts have rounded edges and are all small 90 deg torus. What is more stable a section cut circle or a full 360 deg torus of a 90 deg rounded edge?

I'm constructing 10 section cuts, 10 circles, then averaging them in the user defined settings due to our tolerances being so tight. It would be nice to have a feature like a cylinder for a torus...both internal and external.

 

[[Ja...]]

Hi Tim, im afraid i dont know the mechanism for this anymore .  We used to deal with it as a distributor by supplying a secure file transfer location (FTA).

I keep meaning to ask who your distributor is....

I must say i still have reservations about the toroid being the approach but i do understand wanting to use the full surface data.

It also concerns me when you say this feature has tight tolerances .  As it is inherantly unstable these type of radii can be very difficult to pass grandr checks regardless of measurement system.

Ive just thought of a further possibly wild idea in the direction you are thinking...

Create a multi radial section all the way round the quarter toroid but then 'unroll' this section onto a plane .  In this way you will now have a 'quarterpipe' made up of however many sections. You could then fit a cylinder of whatever style you wish through these sections.   Furthermore after you did this you could do a deviation colour plot of the unrolled section onto the cylinder so you can understand worst cases and maybe delve deeper on their shape.

[[Be...]]

Hi,

as always: If you want to send us confidential data please contact your local ZEISS partner. They will send you a link for uploading your data confidentially.

Nevertheless, I share James concerns about the stability of these "torus" like shapes if you don't have a torus which restricts the radii (same as for single circles).

Furthermore, I can't promise if or when we would realize such a feature (other people would decide this 😉 )

Regards,

Bernd

[[Ti...]]

Please sign in to view this username.

 

Please sign in to view this username.

 

I hear many people on here comment on issues with stability....specially when someone says the tolerance is tight.

 

My question to you both is this....do you think that creating one section cut of any of the above pictured shoulder radii in one singular location is more or less robust that if there was a torus feature inside GOM to generate the 3d geometry for the torus similar to a cylinder?  I'm a big believer in 3d geometry is better than 2d geometry. If the 3d geometry of a torus is 90 degress, yes, it is unstable and not very reliable, but then how reliable and stable is the single circle created from a section cut?  How reliable is 9 section cuts averaged?  I'd prefer a 3d geometric feature that can gaussian out the shoulder radius in the form of a torus.

 

Please sign in to view this username.

  While I don't intend on unrolling a bunch of section cuts as you suggested I am crazy interested in this feature...Where is this? I haven't heard of this before...maybe I'll take a look at it.

 

 

Edited April 17

[[Be...]]

Hi,

of course, creating one section is more "unstable" or better more "unreliable" than to fit a torus where a kind of  averaging would take place. Nevertheless, if you fit for instance a cylinder to a pointset which does not cover the surface for more than 180% degree the radius and the axis position is more "unreliable".

Regards,

Bernd

[[Ja...]]

Hi tim ,

Firstly the unrolled section is in the sections - other ( i cant bring to mind exact pathway ). Help function will give you the path if you type unrolled section.

Secondly , these types of quarter circles are notoriously difficult regardless.  

Im a big advocate of using as much 3d geometry as possible , you have a system that generates the surfaces use them ! .  However as bernd pointed out whether it is a quarter circle or quarter toroid youll not be able to gdant check it as it doesnt have enough bounding surfaces/ inherantly unstable .

I think we already mentioned the 'technically correct' answer of using a surface profile in this feature but i do understand the conflict with design/usability of this result if if indeed it is the radius that needs controlling.

I do absolutely get where you are coming from .  My suggestion of the unrolled section is basically a workaround to convert your toroid into a cylinder .

Multi section radial ..100s of sections easy to create quickly .  It would be handy being able to do screen capture and discuss with you live...but i cant do that anymore easily .