Blendsrf, intersection curves and naked edges

Rhino
,
1

Hi. I’m on Rhino 7, been using Rhino happily for many years.

Just trying to blend two surfaces, one of which is trimmed. Blend is created looks fine. If I join the two surfaces and the blend, there is no naked edge which is great. All bread and butter stuff.

Chk2
Chk22137×1404 352 KB

If I don’t join but ask the perform the intersect command, I’m getting a broken line (highlighted in yellow).

Chk1
Chk12129×1365 382 KB

I pretty much always check blends with intersect and have never had issues that I can’t fix logically.

Feels funny to me that the intersections largely start and finish at isocurves.

Trying to bug check, messing with Rhino unit settings doesn’t seem to cure it.

The surface throwing up the intermittent intersection is not trimmed, doesn’t have anything clever like different point weights etc so for me should be easy to run this blend along.

I’m getting similar issues with the networksrf command too, where I never used to.

I don’t seem to have issues with loft or say sweep 2 rails, but these commands give me less control.

I haven’t knowingly done anything in settings. My Rhino toolbars looked a bit different after a windows 10 update a few weeks back.

May something have changed? Should I just not worry because if it says no naked edges when joined that is the acid test for downstream machining etc?

Exported surfaces attached in case anyone has any bright ideas. File that this sits in is 136Mb or some such.

Thanks in advance

Blendcheck.3dm (231.8 KB)

2

Using Intersect to check if edges of adjacent surfaces are close is new to me. My guess is that is not an intended use of Intersect.

To check the deviation of surface edges I use CrvDeviation and select the two edges to check. Doing so on the upper edge of the blend surface and the adjecent edge results in:
Maximum deviation = 2.26898e-12
For the lower edge:
Maximum deviation = 8.66707e-13
Those results are in the numerical roundoff range. No problems.

You can also use EdgeContinuity to check the deviation between adjacent edges.

If Join results in no naked edges then the maximum deviation of adjacent edges will be less than two times the absolute tolerance. (If JoinEdge was used it’s s a different story and it can cause problems downstream.) So that is usually the test for whether the result is a closed surface for manufacturing, assuming the absolute tolerance is appropriate. The usual suggestion is the absolute tolerance should be about an order of magnitude tighter than the acceptable manufacturing tolerances. The absolute tolerance in this file is 0.001 mm which is around 1/100 of the diameter fo the average human hair, so it should be more than tight enough for a boat.

3

Thanks for the response, all makes perfect sense. I will get over myself and get on with it. I just can’t help the feeling that something has changed on my machine or the software.