Cone Angle

[[Ko...]]

Hi everyone,

I’m struggling to measure a very small internal cone angle.
The cone has a hypotenuse of  0.5 mm and an adjacent side will be 0.35 mm

The main problems are:

• The cone feature itself gives a result out of tolerance angle with both helix vs. linear lines.

• Creating and measuring with two opposite linear lines features  on the cone and constraining the normal vector gives a stable result, but it doesn’t fully match results from the contracer or comparator. withoiut constrainingt he normal vetor the readings are closer to the actual but better than measuring the cone as a feature.

• The internal cone is also 0.1 mm ±0.1 mm away from the reference face, which makes positioning tricky for the CMM to find the location of the angle.

Does anyone have suggestions for improving accuracy on such a small, short cone?
What scan strategy or evaluation settings (constraints, filters, etc.) would help the CMM find the correct angle more reliably?

 

Thanks in advance

 

[[Ma...]]

I would try curves or if you can scan that cone, then tryout locking ( restricting ) feature's axis ( vector )

[[Da...]]

difficult task 🙂, i would: 

Use contour tester if i can , its simple easy and may be bether if have no time. If i cant use contour tester and have time: 

The part should be good quality lets say Ra 0.4 chamfers, drawing Ra 6.3 on such tolerance and such small part is WRONG you just cant have two same results on contour tester if you put on, measure, rotate part few times results will be stupud. How can you trust your results if they are in range lets say +/- 1deg.  This must be PREETY in order to measure it acuratelly. Another way i allways say "sorry this measurement is bad" you cant trust it and i cant make it bether. They will hate me later on ;]

 

1 Good alignment is very important, the minimum is Align to some Z plane (top plane where the hole is made) and some center of hole as X0 Y0, may be good to repeat this alignment 2-3 times as its really smal thing. Do some tests to make sure after re fixing results are repeatable 

2 Use probe ~0.3mm max in diameter. 

3 Use ~2000 points in minium 2 circles (do first runing tests for 6 such circles) measure it very slow 

4 Check 3D grafic, does it looks preety ? does it look reasonable ? may the sigma be too high ? are the deviations of cone leting it to measure it propelly ? lets say its max 0.015mm in form of cone. Are the results coresponding to cone shape - nothing is not measured on hole or champfer. 

5 Measure first cone in "safe mode" in areas to be 100% secure of chamfers, then make intersection of hole and cone, then first circle set on high of this intersection + lets say ~0.07, second circle under Z plane i champfer - some safe lets say ~0.07mm. would be good to add some more circles between. Connect them in cone and look. 

6 make tests to se if its reapetable at least, iif it is, make also test after tool change / machine change and so on ... its lots of work 

write back what you did and how it works, its interesting. 

There are also themes about small cones here

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[[Wo...]]

Some one, (I think Laura?) has posted these instructions some time ago to measure tiny countersinks. It was more about not hitting the csink correctly, but maybe it will help:

"

Since, by adding up all tolerances, it's going to be allowed to move around by more than its size. I have been running tiny countersinks for years without ever getting errors.

Here's the recipe:

1) measure the bore and mating plane as well as needed. The plane might have to be only the part near the hole, depending on how good you expect it to be.

2) measure countersink (use filtering) very coarsely, 5 traces evenly spaced between 1/3 and 2/3 of it's nominal height. [The reason you will use 5 traces, instead of 2 or 3, is that should one "fall off the edge", the computer can tell that's an outlier, with 2 or 3 traces, it goes crazy]

3) create 2 shell intersections, one between cylinder and countersink, the other between plane and countersink

4) now make a fine countersink, with 5 traces (not 3, sometimes they're not really cones), spaced evenly from 1/4 to 3/4 of it's nominal height. Every single nominal field (cone angle optional) in this new countersink, including trace heights in strategy, will be a formula, based on what it could tell about the course cone measurement. Example, in x (getactual(small cone intersection.x)) - same for heights, diameter, everything.

This works, 99% of the time. With my method, if you get that error message, it's because the countersink either isn't there, or it's a radius. Guaranteed. 

 "