Visually matching radii
I know this is getting a bit geeky but it might come in handy someday when you are pumping CAD.
I tend to get into small decisions like this all the time. When laying out 2d curves in CAD and you want similar “looking” radii at the corners it tends to be a judgment call if the straighter sections come off at different angles. The below sketch shows a 12mm radii on the left off of two perpendicular lines. The same size radii on the right is placed between two lines that come together at a 45. The size of the radii does not look similar to me, although they are identical.
Decreasing angle as seen below on the right to 22.5 degrees ( half of 45) makes the radii look even more different.
So here’s a try at double the radius value of twelve on the left. It’s a judgment call at this point. I’m gonna saw these radii look about the same at this point. Close enough to fool most people.
So now I’m going to try to visually match the radii again at the 22.5 degrees (half of 45 degrees) side. I raised the value of the radii to 48 (4X 12). Look about right? Again a judgment call. Looks good to me.
What is visually the same is a judgment call. But I do have something else to measure here. Here is the last scenario I showed with the for times radius value on the 22.5 degree angle. I thought it looked about right. I then measured the arc length and it appears they are exactly the same.
I knew this was going to happen, I just want to point out the fact then when created visually matching radii on a series of straighter segments, the radius value should not be matched, it should be the arc length. Typically a bit harder to do when creating sketches in CAD. But it seems to be more accurate according to the eye.

Leave a comment





hi tom.
first of all, im an ID student from Argentina and is a real privilege to me to find designers like you who share their experiences. about the article, you are not geeky at all, i think that this kind of details in design is what make the difference at the end. im not sure if u get my idea, my english is still very limited.
thanks for reading!
alan
alan
July 15, 2010
Hi Tom,
Nicely put, sir.
Sometimes I’ll also just dimension the “width” of the sketched fillets and link their values so that they’re always the same (if that’s what I’m after). So in the case of the left fillet (18.85) that “aligned” dimension would be pulled out at a 135 degree angle and the right fillet would be pulled out at, um, 67.5 degrees-ish from horizontal, I think. But doing it by eye definitely works as well.
I’ve noticed lots of fillets (as oppposed to blends, y’know C2 or better) on power tools and have always wondered why that is. I’m sure you’ve seen Gray Holland’s periodic table of form PDF. While I’ve always noticed the difference between fillets and blends on tooled parts, I never thought to term the junction of a fillet and primary surface as a ‘blemish’, but I think that’s about accurate. On the interior faces, sure, y’know, do whatever. But I’d figure on the outside when aesthetics are important, that blends would be the way to go. Is there a cost- or leadtime-related constraint that prevents you from blending (y’know, C2 or better) in your industry?
Is this Catia that you’re running? If so, how do you like it?
And sick work, btw.
Cheers,
G!
Gamal
August 6, 2010