 # Tolerance question

I have a question about the application of tolerances.

I am concerned that I might have misunderstood the Rhino document, “Understanding Tolerances”.

Let us say I was modelling a 5.2m long kayak, and that I want the Rhino model to be accurate to within 1mm.

My options are:

a. If I were to work in units of metres, then the absolute tolerance would be set to .001. The boat is 5.2 units.

b. If I were to work in units of centimetres, then the absolute tolerance would be set to .1. The boat is 520 units.

c. If I were to work in units of millimetres, then the absolute tolerance would be set to 1. The boat is 5200 units.

My question is, which unit setting would give be most likely to provide the most consistent modelling performance in my Rhino model, or does it make no difference to whether I work in metres, centimetres or millimetres, because the boat is the same length in the real world?

If I have understood the document correctly, the answer would be c.

Hi -

If that means that a hole of 1 mm should be accurate, then your tolerance needs to be at least 1/10th of that. You always need your tolerance to be at least one order of magnitude smaller than the smallest feature in your model.

Even if your smallest feature were to be 10 mm, I’m afraid that the correct answer is a.
But that is only because Rhino is generally happiest when the tolerance is set between .01 and .000001.
-wim

Rule of thumb:
Use a tolerance setting one order of magnitude tighter than (1/10 of) your smallest modeled detail.
In general keep the absolute tolerance setting in the range of `0.01` to `0.0001` . Never set it below `1.0e-5` – chrashes might occur and you pay with slow computing speed. If you need a tighter tolerance, use smaller units instead.

Rhino does not know the size of the physical object. It also does not use the type of units (meters, centimeters, millimeters, etc) in calculations. If you need an accuracy of 1 mm then you could work in meters with a tolerance between 0.0001 and 0.00001, centimeters with a tolerance between 0.01 and 0.00001, millimeters with a tolerance between 0.1 or 0.00001, or inches with a tolerance between 0.004 and 0.00001. Note that in each case the tolerance can be as tight as 0.00001.

The only drawback to using a tighter tolerance than needed is a few operations may be slower and depending on how the modeling is done there may be more control points and a “heavier” model. The drawback to using too loose a tolerance is the model is not acceptable and needs considerable rework or redoing.

1 Like