Hi Michael,
This will help me solve a machining question that I would like to answer in another forum. I hope you dont mind.

This is a screenshot of a cylinder with a .5 hole cutout and a fillet of .125 put on it. The fillet becomes variable and the direction that is used to do this will answer my machining question.

This is a top view and the topmost edge of the cylinder is what reflects the fillet radius of .125. The bottom part of the fillet is larger. Is this due to the fillet doing a "Straight down" look through that distorts the radius around edge of the cutout?

In other words, from a machining point of view, if I use a fillet bit, which starts its cut at .125 on that top edge, and follow it around keeping the bit "Vertical", when it got to the bottom it would naturally be cutting more material and create this variable? The other option would be for me to look at following the edge with a 4th axis rotation.

Maybe Danny could jump in here and give me a small lesson. I would mucho appreciate it.

Thanks,
Burr

Hi Burr - this actually doesn't make a variable radius fillet - what you get with fillets on curved things like this is the same fillet radius everywhere, but some of the fillet cross-section arcs are smaller and some are larger, depending on the angle that the 2 surfaces meet at in a particular location.

Here's an illustration - here I have 2 arcs both of which have radius 5, that is they are both portions of a circle that has radius 5. But you can see that even though they have equal radius, they can have different lengths:



That's what is variable in the case you've got here - the length/size of each section (I guess you would say "subtended angle"), not the radius.

I've attached a file burr_fillet_arc.3dm which has several cross-sections of the fillet extracted as individual arc curve objects, this may help you to see more clearly what is happening exactly:



If you pick each of those arcs, you'll be able to see that each of them is of radius .125 .

And if you look from the top you'll also be able to see that they are not necessarily all oriented in a static direction like you may also have been thinking:



Hope that may help to show what is going on!

- Michael

Hi Michael,
That clarifies well thank you.

Hi Burr,

>Maybe Danny could jump in here and give me a small lesson. I would mucho appreciate it.

Yeah! what Michael said ;)

I know what you mean there, I've had this situation before and I think there are a few discussions on the forum about this situation.
I've attached a work around which gives you more like the picture attached, if you want that, there are a few more steps involved and
you would need to 3d machine it with a ball nose cutter.

The attached file has the steps involved.



Cheers

Hi Danny, Thanks for the time there.

I'm not the best communicator.

I have a model with that original fillet on it (The cad/cam package fillets it that way also) and I was trying to understand How it came about that way (From Michael) and looking at how to cut it (You).

I think I solved it but wont know for a bit as our machine is tied up doing some jobs at the moment.

The issue is the original .125 circle used to cut out the hole gets stretched due to the "projection" of the curve on the surface of the cylinder. Most have already suggested using a 3d toolpath to cut it and that works well and is a done deal, though me, being the nut that I am, am trying to figure out how to do it with a profile job (Much faster than a 3d toolpath).

My solution is to take the original 1.25 circle and convert it with "axis-dim*(360/(Pi*2))/radius", which is essentially doing the wrap, then I can run a profile operation on it with a 4th axis rotation. I think this will produce the "Subtended Angle" that we want but I cant proceed until I can run some toolpath and see if the theory jives up.

Thanks again,
Burr

Hi Burr,

>I am trying to figure out how to do it with a profile job (Much faster than a 3d toolpath).

I see now.
Yes, I would of called you a nut also in the past until I joined my current company.
We do multi cavitation injection moulds and we use form cutters where ever possible, like you said, it cuts machining time enormously.
It does help when you have your own tooling department with cnc tool & cutter grinders.

Anyhow, good luck!

Cheers

Thanks Danny,
Thats what my brother said was we would have to cut our own tool for this! But one small operation for a larger gain.

I'll let you know.

Hi Danny, is this the one where you want to get a "constant distance" style fillet rather than the rolling ball type with constant radius?

Michael,
My original question was more for my "how to machine it" understanding and not that "how it should be filleted" should be different.

With that constant fillet, being the fillet that is desired, I have 2 options. I could do it with a 3d toolpath from a 3d model that has the fillet that looks like this:



Or I can cut the curve itself from the cutout:



This is esentially the 1.25 inch circle "surface projected" onto that sphere. We could get the 2d shape of that by wrapping the 2d sphere to the cylinder with a math function of:

wrap dimension*(360/(Pi*2))/radius

Then we could run a 2d toolpath that looks like this: (the light blue represents the cylinder unwrapped)





While turning round stock around its axis. Then a 1/8 inch fillet tool could just make 1 sweep around the edge and be done.

Would it be possible to get both types of fillet for the same situation though?

Thanks,
Burr

Hi Burr, I can't really help you myself with the "how to machine it" part. But when you were asking about which part was variable since when you viewed it from the top it had some skinnier and wider areas to the fillet, that made me think a bit about this alternate way to do fillets.

I'll attach a model here that has this alternate way of doing the fillet, if you want to check it out.

This method of creating the fillet uses a different measurement to control the fillet.

Imagine that this is a cross-section of the fillet piece between 2 surfaces:



With constant distance, instead of controlling the fillet radius you instead control this distance value for the distance between the rails:




However, this does mean that the radius is variable and is constantly changing, the radius grows or shrinks as needed to maintain that distance. So I don't know if it will help out in this particular case or not, maybe not I suppose. But if would, I'd be interested to know!

The geometry library that I'm using has an option for doing constant distance filleting but it seems to be somewhat less reliable than the regular fillet mechanism though.

- Michael

Thanks for this additional file.

With the original file you posted showing me the actual radius value of the fillet I was looking at was 1/8 inch, just along a longer arc, this made sense from a machining standpoint and I followed a path to figure out how to do it from this type of fillet. A tool couldnt change radius during its path.

However, I like the looks of this last model from an aesthetic point, though a cad program would not want this. It mentally looks right.

I, being the nut that I am, would ask if I could have an option to do "Both!"

I could see cad guy's needing the real world path but in some circumstances this last one "looks" nice.

Can we have both?

Sorry,
Burr

Hi Burr, there isn't any way to have both of these all at the same time since holding one of these things constant makes the other part to be variable.

But do you mean an option so you could switch back and forth between different methods?

A couple of problems with that are first I generally hate adding in extra UI... but maybe that will not be so bad if I put it as another thing in the Shape: dropdown list like this:



Or does that seem weird to have it put inside there?


Then the other problem is that it just does not appear to work quite as well as the standard filleter, but maybe it is not too bad.


- Michael

I personally like the option in the dropdown, But I'm not the visionary! Ha. It is a variation.

Sleep on it and if you add it, I will like it.

thanks for the time.

Just a note. In our cad/cam package, the current result produced by MoI is equal to the one in our package called "Constant Fillet".

Thought you may want to know this.

Hi Michael,

> is this the one where you want to get a "constant distance" style fillet rather than the rolling ball type with constant radius?

Yeah, that's right, as Burr mentioned, for aesthetic reasons, in design it would be nice to have as many fillet options as possible, including variable fillets one day :)

> ... but maybe that will not be so bad if I put it as another thing in the Shape: dropdown list.

That for me seems the more appropriate spot as it does relate to Shape, but I don't know why I feel that Const Dist should come after Circular then the three G's.

---------

Hi Danny,

> including variable fillets one day :)

Yup, but I'm not sure yet if that will be able to get integrated directly into the existing Fillet command or whether it will need a specialized separate command for it, since it kind of involves picking additional information like points on edges, etc...


> That for me seems the more appropriate spot as it does
> relate to Shape, but I don't know why I feel that Const Dist
> should come after Circular then the three G's.

Yeah that does seem to make sense to group it with circular since it will also generate circular cross-sections just with a different overall shape for the fillet surface.

The only thing that is bad about putting it in the Shape: dropdown is that then you won't have the possible combination of having "constant distance" style for the rails but then also having blend cross-sections rather than circular cross sections.

It is technically possible to have the "rail generation" and the "cross-section shape" as independent things. But I don't really like the proliferation of UI that would cause, putting it under Shape and just having constant distance be circular only makes for a more tidy UI. Also constant distance with a blend cross-section rather than circular cross-section seems like a kind of unusual combination.

- Michael

Hi Michael,

> Yup, but I'm not sure yet if that will be able to get
> integrated directly into the existing Fillet command or
> whether it will need a specialized separate command for it,

Definitely a separate command, especially if your thinking of extending the history function in V4 where you will have the list of varying radii to edit :)

> But I don't really like the proliferation of UI that would cause,
> putting it under Shape and just having constant distance
> be circular only makes for a more tidy UI.

Maybe something like the sweep options dialogue box ?

---------

Hi Danny,

> Maybe something like the sweep options dialogue box ?

Well, I'm not that happy with having so many options inside of sweep neither, it's not exactly a model that I want to follow unless it is really necessary.

- Michael

Oh! Ok, it's the simplest sweep options I've seen.
If you want simpler than that, then you have to design a way to have no UI and we just picture what we want in our minds and MoI will pick up our brainwaves and start modelling.
Could you imagine what would come up on the screen from the minds of most males :)

Cheers
~Danny~

>>>>The only thing that is bad about putting it in the Shape: dropdown is that then you won't have the possible combination of having "constant distance" style for the rails but then also having blend cross-sections rather than circular cross sections.>>>>


Could be poosible to have a "checked" type dropdown you would find in a Menu option where multiple items could be checked.