Unbalance measurement

Hi guys, I have a wheel bearing drawn in Pro/E.

Does anybody knows how to evaluate the unbalance in [g*mm] ?

Thanks!

 

1) In WF 3 (similar in previous versions): Measure>Model>Mass
Properties
2) Pick coordinate system at center of axis of rotation.
3) If the density is defined correctly, this will give the mass and
center of gravity (CG). Multiply the mass (gram) by the distance (mm)
from the center of gravity to the axis of rotation. This is the
imbalance in g*mm.

The distance from the CG to the axis of rotatation can be calculated
or you can create a point at the CG location and measure the distance
to the axis of rotation. Measure to an axis not a coordinate system
because the axial direction does not affect imbalance.

Dave Parker

 

Excellent. Thanks a lot!

 

Sorry, I really don't get this. How can the CG differ from the center of
rotation? Accuracy factor? Or at least if you have a constant section
rotated 360 degrees around an axis. Especially when you're talking about
digital models. Don't think you'd pick up string theory's vibrations at this
scale. So, center of mass and center of rotation ought to be coincident. At
least in the plane normal to the axis.

David Janes

 

I assume the original poster has a valid reason to expect the part to
be imbalanced. If the wheel bearing is a simple revolved feature I
agree that the CG and axis of revolution will be coincindent and the
imbalance will be zero. However, the CG can be offset from the axis of
rotation if the feature is not a simple revolved feature and has some
asymmetric feature. Or perhaps the nominal part is perfectly balanced,
but the original poster is interested in modeling the imbalance at
worst case tolerances.

Dave Parker

 

Maybe we're expecting too much of digital models. Granted, he may know, FOR
A FACT, that the part is unbalanced (the wheel balancing machine told him
so!) Then, is he somehow modeling an unbalanced wheel trying to simulate the
balance results in Pro/e? Good trick, since you can't spin your model and
get the unbalance that way. The unbalance you get with a model is, at best,
static. With an actual spinning wheel, it is dynamic: some unbalance factor
turns into some changing amount of vibration (what the wheel balancer
detects)

Thanks for acknowledging this. It seems to be the dominant one.

Yes, an interesting possibility. But, he did say "a bearing" so probably not
a keyway. I'm stumped, maybe marcellen could weigh in with some particulars
about how this could be offset from the center of rotation.

Can Pro/e do this? Take the nominally perfect revolved section (whose CG and
axis of rotation are both 0,0) and produce, by virtue of a profile tolerance
of .001, an egg shaped bearing, 90% of whose profile is at min and 10% is at
max allowed diameter and then produce some offset of CG from center of
rotation? Would be nice but I don't know of any such function within Pro/e,
especially since there are an infinite number of profiles and offsets that
could accomodate some such scenario. I could see making this a function of
something like Pro/CAST and placing the imported data from Pro/SCAN TOOLS
into this type of analysis to calculate if a maching allowance can be
accomdated by an inspected casting. So, as a way to compare scanned parts
(incoming inspection) with engineering models, this could be of great use.
But to compare theoretically perfect models with theoretically imperfect
ones? Don't think Pro/e has the chops for that. Couple of orders of
magnitude beyond Pro/e's capabilities.

David Janes

 

David I appreciate your knowledge and efforts in contributing to this
newsgroup. The original post was a valid question and I assume the
poster had a good reason to ask it. My answer was a valid answer for
how to calculate the static imbalance as a mass*moment arm. This is
what the original poster implied when he asked for a single value in
g*mm.

Dynamic imbalance moments can also be calculated using ProE (without
spinning the CAD model) using these equations:
Mx = Iyz*w^2
My = - Ixz*w2

Where z as the axis of rotation, w is the constant rotational speed,
Iyz and Ixz are the moments of inertia about CSYS aligned with axis of
rotation, and Mx & My are the resulting "imbalance" moments about the
x and y axes.

I was only speculating that the original poster might be considering
tolerances (e.g. concentricity of the inner and outer race). I did not
mean to imply that ProE could automatically apply blueprint tolerances
to the mass properties. However, you can create a CAD model with
geometry at worst case tolerances for analysis purposes. In my company
we typically analyze nominal parts. Then if any region of concern is
noted, we create a separate CAD model with worst case geometry for
further analysis. Note that "worst-case" may be min material, max
material, or some combination of min/max material conditions in
different regions. Furthermore, geometry that is "worst-case" for a
particular region may not be "worst-case" for another.

Dave Parker
West Palm Beach, FL

 

David I appreciate your knowledge and efforts in contributing to this
newsgroup. The original post was a valid question and I assume the
poster had a good reason to ask it. My answer was a valid answer for
how to calculate the static imbalance as a mass*moment arm. This is
what the original poster implied when he asked for a single value in
g*mm.

Dynamic imbalance moments can also be calculated using ProE (without
spinning the CAD model) using these equations:
Mx = Iyz*w^2
My = - Ixz*w2

Where z as the axis of rotation, w is the constant rotational speed,
Iyz and Ixz are the moments of inertia about CSYS aligned with axis of
rotation, and Mx & My are the resulting "imbalance" moments about the
x and y axes.

I was only speculating that the original poster might be considering
tolerances (e.g. concentricity of the inner and outer race). I did not
mean to imply that ProE could automatically apply blueprint tolerances
to the mass properties. However, you can create a CAD model with
geometry at worst case tolerances for analysis purposes. In my company
we typically analyze nominal parts. Then if any region of concern is
noted, we create a separate CAD model with worst case geometry for
further analysis. Note that "worst-case" may be min material, max
material, or some combination of min/max material conditions in
different regions. Furthermore, geometry that is "worst-case" for a
particular region may not be "worst-case" for another.


Dave, what is this "worst-case" scenario geometry? If it's simply a smaller diameter, it will not run out of round and vibrate (the OP's concern?). Will it be some kind of eccentric of a profile tolerance? where and how much variation? you must specify and, as I said before, there's an infinite number of possible variations within any profile.

You provided the correct theoretical answer. At the same time, production has managed to solve, within well known limits, most of these problems. What I'm not seeing from the educational community, the theoretical community, the mathematical community is a) a recognition that these well known engineering problems have well known practical solutions, b) the design tools can reflect and incorporate those design features. So, each time the "theoretical" problem is posed, real world engineering doesn't get set back 300 years, as if the problems had not, in fact, been solved and c) 95% of engineering grads know nothing about those practical solutions to the well known problems and act, for a decade at least, as if they're required to reinvent the wheel. In contrast, a very experienced engineer I met recently gave me the opposite dictum: don't make a custom part when an off-the-shelf part will do. Generally, in spite of the prejudices of the junior engineers, these are the better made parts (with realistic and consistently realizable tolerances). They're also cheaper and over-night deliverable.

David Janes