Complex 3-Dimension Measurements

[kthompson]

ok... I'm gaining alot of really great knowledge here and feel I'm doing pretty good learning to use the CMM and PC-DMIS from you guys... thanks! NOW....

Most of the things I've been checking are in z-plus workplane, checking X & Y coordinates... flat stampings... etc. I just helped someone with a part where we had to change workplanes, offset the rotation a few degrees, and check some funky dimensions... probably no big deal for alot of you guys... my head hurts... (geometry was my best and favorite subject in high-school, but I just had to look in my profile because I couldn't remember how old I am... my head isn't what it used to be!) anyways... in wanting to learn more about checking complex parts, multiple workplanes, etc... where would you guys direct me for a good resource in learning it properly? or just keep up the trial and error stuff as I usually do and learn from my own mistakes?

suggestions?

thanks!

Kev

[Wes Cisco]

This forum and the good old school of hard knocks are the only resources I know of for what you seek to learn. I handle the kind of complex geometry you are talking about almost every day. Try it yourself first. When you get stuck or it just plain don't make no sense at all, then post it here. Remember to describe things as completely as possible. Pictures are always very helpful. I find the Graphics screen to be very valuable, especially if you have version 3.7. You can always send me a PM, I will help if I can. I hope you like trigonometry too.

[kmcmm]

oh yes trig, wait till you get to program body corrdinates without CAD or a print that is drawn to part specs and not all car position. Sine, Cosine, and Tangent are me best friends.

[#2#]

My suggestion is to find something that you know the dimensions of or you can measure using other methods. Then write up a program to measure the same features. Keep fooling around with it until it comes out right. Start with a fairly simple part and work your way through.

[Matthew D. Hoedeman]

Well, one trick to getting around the workplane issue (not that it is an issue, really, but this is another option) is to measure a plane, on that funky angle you talked about, then when you go to measure a circle or a line that is in that measured plane, you can change the feature so that it is created in the measured plane instead of in a work plane. You simply double-click where it says WORKPLANE in the feature and hit F8, it will change to FEATURE (there are other options, but feature is the first one), then you can click elsewhere, then go back and put in the FEATURE ID of the measured plane. It will then put that feature 'into' the measured plane. Pcdmis will then also use that as the default for all learned features until you change it back. I would suggest re-measuring the first feature (the one you changed) so that it is already 'looking' in that direction for the creation of the feature. This will eliminate the need to rotate alignments to get square to an angled surface and still let you learn the features without using CTRL-F features.


I have found this function VERY handy when the need arises to map out (digitize, check, align) parts or details that I can not sit square on the machine. If your level plane is not square to the machine axis and you have a rotate line in that plane, you will not get proper probe comp and direction unless you do. The biggest example of this I can give is I have a detail that I check for a local machine shop that I have to lay it on it's side so I can check everything that needs to be checked. Well, this makes the bottom at about a 40 degree angle to any machine axis. If I were to measure the plane, then the 2 sides for the rotate and origins, the lines would get comped to a machine workplane, not to the bottom plane of the detail. This would make my origin off by the sine of 40 degrees times the radius of the probe. BUT, by changing the WORKPLANE to FEATURE=PLN1 (feature ID), then it WILL comp correctly.

You could also get around this by having an alignment that just levels to the measure plane, but I want as few alignments in a program as I can, so I don't do this.