Cylinder diameter problem

[rerun_lives]

I am checking a cylindrical bore using a 4MM probe (.1574") using three rows of hits 12 hits total. The bore pin checks .256 but the CMM tells me the Diameter is .2623 - WHY IS THIS? The probe calibration checks out fine!

[#2#]

I'm not sure as to why in your case. Some possibilities...

Your cylinder is not straight. The diameter at one "level" in the cylinder would be larger than the gage pin that would slide through the cylinder.

Your cylinder is dimensioned as the average diameter of the cylinder or the min. circumscribed circle. If you dim. it as the max. inscribed circle your results would correlate better with a gage pin.

Your cmm just isn't good at measuring diameters - none of them are. I would think you should be closer than 6 thou though.

Did you calibrate the probe? Is the location of the cylinder good? If it were bad your hit vectors would be less than ideal and that could introduce error.

[Jan d.]

CMM's are not the best to measure cylinder size, but you are waaaaay off. CMM's ought to do better than that. I typically get within a few 10ths of the pin.

Do you measure 3 individual circles or autocylinder at 3 heights?

I would use 3 autocircles. Output X,Y,D,RN. Look at the XY for all 3 and you can get a good feel whether they are in line or not.

From earlier post I got the hunch that before V4.0, the construction of a cylinder from 3 circles is sometimes unreliable ( I may be totally wrong here because I have never used PC-DMIS before V4.0). I know that in V4.1 you can construct the cylinder properly from the 3 circles.

Just my thoughts, Jan.

[craiger_ny]

What Jan said. All of my cylinders are constructed from circles simply for ease of trouble shooting. Each circle can be evaluated individually, it makes things quick when it comes time to do nitty gritty feature evaluation. It is an old habit that was started back when cylinders were very unreliable, I used to do the 2 circles/line trick (once in a while I'd construct from points but the enemy was to actually measure a cylinder) for orientation call-outs as you couldn't construct them from circles once upon a time. I do measure cylinders once in a while for constructing because for whatever reason you can not construct from a constructed cylinder (can't intersect a cylider and plane to construct a circle) and I do see the orientation is better than it used to be but are cylinders reliable? I don't know as I don't use them enough to be able to say. I know once upon a time cylinder was the least reliable feature in PCDMIS. What do you get when you do circles?

Craig

[JoeRZ]

What everyone says + check out probing could be shanking if cylinder is not square to machine.

Is that bore machined, cast with draft? may be good to look inside the bore for irregularities, I know when checking castings sometimes you can hit a void and will throw your readings off.

JRZ

[John Riggins]

I am checking a cylindrical bore using a 4MM probe (.1574") using three rows of hits 12 hits total. The bore pin checks .256 but the CMM tells me the Diameter is .2623 - WHY IS THIS? The probe calibration checks out fine!

Please note that your probe diameter is very close to your cylinder diameter. Well, what I consider close anyway.
I have found that when the probe dia and hole dia are very close you can get very different results depending on the speed that you are set up for. Touch speed and actual speed controlled by the potentiometer (or the knob, twisty-thingy).

With a 4MM probe, results will be unreliable until your hole size approaches .350". Location will be good, but you may need to play around with the measuring speed to get close to reality.

When I need very accurate results from the CMM for hole size I do some special setup.
I get a standard very close to the hole size I will be checking and then measure and adjust the speed until I get repeatable numbers close to the standard.
This can be a pain in the Bush (W variety), but it really does work.

[kpayne]

I will normally use surface points created by using the paste with pattern function at everyone 20 degrees with 3 or more levels then construct cylinder. This will allow you to analyze measured feature by points individually, as a circle or as a cylinder.

I measure a lot of plastic molded parts which have draft on most surfaces and I found this method works well for me.

I am using version 4.1 and have not tried constructing cylinder from measured circles since before version 3.5 because all I got was a 3D line and lost all my diameter and form information. I guess I should try constructing cylinders in version 4.1.