mesuring too much or not enough?

[dragonblade]

Hi
I have attached a picture of a part that I hope can be seen.

the small OD on the left and right sides are dat -A-'s. I measure 3 circles with 6 points, one circle at the begining and end of the cylinder and one in the middle of the cylinder and I do this on both dat -A- cylinders on the left and right sides and then construct center points on those circles and construct a 3D line through those points and use that line as my dat -A-. My supplier takes one circle with 5 points on each cylinder then constructs a 3D line through those two circles and uses that line as dat -A-.


The ID cylinder facing the operator in this view is one of two cylinders, there is one on the opposite side and these cylidners are dat -B-. I do the same with those cylidners, I measure 3 circles on each cylinder and create a 3D line and use that line as my dat -B-.
My supplier is only measuring one of those dat -B- cylinder and using that as dat -B-.

After I contruct my 3D lines, I create a intersction point, axis origin to that point and rotate my axis to lines. In my opinion, I have just set a fairly accurate axis on my part to the datums and I can begin measuring all my features and dimensioning them.

My questions:
My philosophy and training is that as a minimum, to take circles on a cylinder in order to best represent the cylinders direction and use the circles to create a centerline, a minimum is to take two circles, one at the begining and end of each cylinder, but I prefer three and take a third in the middle of a cylidner.
But, would I be better off mesuring those circles and instead of creating a 3D line, constructing a cylinder through those circles instead?
I am i doing overkill by measuring three circles or two cirlces on each cylinder? or is my supplier who is only measuring one circle on each cylinder correct?

thanks

[Roberto]

No picture...

[dragonblade]

it's there now

[Roberto]

I agree with you Dragonblade on taking 2 or 3 circles for best representing a cylinder, especially if print calls for squarness o cylindrcity on each of them.

However, in my opinion a 3D line constructed with 2 circles (on each side) is correct for aligning the part.
Now, instead of making a single line off the 6 circles, I would costruct cylinder on each side, pierce points to the face wall and then use those 2 points for constructing the 3D line.
Let's see what the gurus have to say...

p.s. Nice setup though!!

[dragonblade]

""However, in my opinion a 3D line constructed with 2 circles (on each side) is correct for aligning the part.""
Supplier said they were doing 1 circle each side and I don't think that is sufficient because I don't think that one circle on each side represents the cylinder correctly. I agree it is ok for a quick alignment, but I wouldn't use that one 3D line for dimensioning, I would contruct a tighter line for dimensioning by measuring more circles to better represent the cylinders.

Another question, after thought. Instead using a 3D line, what about constructing a Cylinder through the 6 circles and using that as dat -A- instead of the 3D line?
If I were to use pierce points, I would rather mesaure each cylinder as cylinders and then create pierce points instead of measuring circles and construction cylinders. I guess that would be a personal preference.
Being that the center line of the parts is through the two cylinders, I would be less inclined to create pierce points becuase I would have to measure planes on the end surfaces and they may not be perfectly flat and parallel to each other as they need to be, which could offset the consrtucted 3D line, which could cause an OOT condition when it isn't OOT, right?

I know there are plenty of ways to skin this cat, but I am wondering what is the best approach.

thanks

[keego]

Generally Datum_A is the primary datum, the leveling datum which locks down your first 3 degrees of freedom. I would measure both sides as a single cylinder, (3 circles on each side all from left to right), and level to that. Always use at least 3 levels when measuring a cylinder.

[STANLEY]

You don't need to overkill it. One circle at the beginning and one at the end of the cylende is good enough

[JoBLOCK]

how do you level to 2 cylinders? or a 3d line. Oh so you could level to the line too.

[Otto Square]

What the heck is that thing?

[dragonblade]

i don't have a problem with doing only two circles on a cyliner, but I do think that only one circle isn't enough.

as far as constructing a line vs. a cylinder, I have played with this and it seems to me with the software that construting a line is better than constructing a cylinder through the cirlces that I measure.
thoughts?

[MARTY Syed]

Construction of 3D Line

Hi Guys,
1st of all, I'd like to say that please pardon me if I suggest anything out of line. Reason being I have'nt worked on CMM for almost 3 years, due to a work related injury.Good news is that I am starting a new job tomorrow, at a forging and a machining house.I want to jump in just for the heck of it. 1st think I noticed in the picture is that you have Clamped the ? out of this part.It seems like a heavy enough Aluminum hogout. If you have to program this part. How the heck are you going to avoid all those clamps ? I have even forgotton the terminologies used. I believe it's called the clearplane and a passthrough clear plane. I'll find out tommorow. You would need to either use a lot of movepoints or if you are using clearplane in 2 directions you would need pretty big numbers, to back off rotate and move over for your clearplanes. I would think simply a glue gun would have done the trick, leaving you all kinds of room to move your probe around the part without having to worry about crashing.When I used to program or even just check the part, I would always try to measure my primary in one shot. 3 step cylinder would have been the best. You can still measere it as a cylinder with 6 poits 3 @ each end. Then use it for leveling the part. The way part is sitting, I would apply the right thumb rule and go from left to right, and level Z. if you donot measure the cylinder from left to right since it is a 3D element PC-Dmis gets confused and flips your axis's.Even if you are telling it to level Z, it would still flip your X and Y axis. As far as using the pierce points, they give very good results when checkings positions. for alignment purpose, they dont work too good, because if your surface is even off by a few tenths, it throws the location of the point quite a bit. That is the reason I always avoided using pierce points for alignment purposes. You did'nt mention in which direction are you rotating your secondary. Again looking at the picture if the bore on the front face is Datum B, I guess it can be measured as a cylinder using a longer extension, from back to front and rotating the positive y to this cylinder. Your tertiary is not mentioned, but I think a point shoud be enough to use as your Datum -c-since you R already leveled and rotated. Now in my mind I'm thinking that this alignment should give me the realistic results. If all this does'nt make sense to any of you.Please let me know. That would mean I have forgotten what I knew very well once. And that means that I will have to work very hard at my new job,because I will be doing all the programming at this new Job. This also means that all those Pain killers (Methadon and Oxycontin) all the good stuff have really screwd up my memory and I dont remember squat.

Marty Party Hardy

[dwade]

Just food for thought.

When making decisions for measurement methods I try to find out how the hole is made. How the hole made makes a difference how I am going to check it. If it is bored with a boring head I don't worry too much about form, not much can go wrong other poor surface finish. If it's interpolated I will take more hit on each circle because I know some of our mills don't interpolate well. Same thing when a hole is drilled, drills wobble and walk unless they are short shanked carbide.

A circle on each end is normally eougn unless you question the form.

Just my .02.

Duane

[Snaggy]

Just food for thought.

When making decisions for measurement methods I try to find out how the hole is made. How the hole made makes a difference how I am going to check it. If it is bored with a boring head I don't worry too much about form, not much can go wrong other poor surface finish. If it's interpolated I will take more hit on each circle because I know some of our mills don't interpolate well. Same thing when a hole is drilled, drills wobble and walk unless they are short shanked carbide.

A circle on each end is normally eougn unless you question the form.

Just my .02.

Duane

Add on to that what kind of tolerances are you looking for? 0.1mm, 0.01mm, 0.001mm ? Dont go chasing microns if you aint looking for them!

[JoBLOCK]

if left OD and right OD is Datum A, wouldn't it be simpler to measure those cylinders as individual points and construct 1 cylinder with those points and level to it?

[justncredible]

if left OD and right OD is Datum A, wouldn't it be simpler to measure those cylinders as individual points and construct 1 cylinder with those points and level to it?

No........