Airfoil profile inspection

[DaSalo]

We a a tricky inspection dilemma which I am hoping to get some advice on:

- Imagine a part that is 36" long, 4" tall, and shaped like a slightly twisted airplane wing.
- The datum structure of the part is defined with a 6 point nest:
Primary datum: 3 point on pressure side of wing, two of which are in line top to bottom near one end and the third is at the midpoint of the other two down at the other end.
Secondary datum: two points on the leading edge located on the same cross top-to-bottom cross sections as the primary points.
Tertiary datum: one point on the tip.
- We are trying to verify the amount of bow and twist that exists in the part in the free state. The intent is to allow .150 bow and twist along the length of the part (the more flexible direction) and half that amount along the width (the more rigid direction).
- Currently this is done with a profile of surface control applied to a central cross section of the part. There is a note that specifies that one tolerance applies along one vector and a different tolerance applies along a vector perpendicular to that.

The question is: can something like this be checked with a hard gage? The problem we are facing is that the part is so light weight and flexible that it moves when contacted by the CMM probe even if the probe is dialed down to minimum force. What we are hoping to do is place the part in a physical nest and have a "guillotine" type profile gage fixed at the nominal position of the central profile. If the part can pass through the gage plate then it is in tolerance. This presents a ton of issues though. How do you define the profile of a zone with a complex shape and variable tolerance? How do you account for variation in part thickness? As the part increases in thickness the suction side moves out from the primary datum so the central profile cross section must move with it. But what about a situation where the areas of the part near the datum points are at MMC but the central area is at LMC?

Is profile even the correct control for this situation? It seems like what is really being checked is the position of a feature of size. We want to know how far the vertical and horizontal centerlines of that cross section have departed from their basic position and orientation. We aren't actually trying to verify the shape of the cross section itself, just its position relative to the datum structure of the part as an indicator of bow and twist on a macro level. If we move to a position control we can now use the MMC modifier and define a hard outer boundary. We still face the problem of that boundary floating as a result of the part moving away from the primary datum and secondary as it increases in size.

I've got myself tied up in some mental knots over this. Any help much appreciated.

 

[ctopher]

There are laser scanning inspection equipment that will check such parts.
The process you mention is a tricky one.

Chris
SolidWorks 10 SP4.0
ctopher's home
SolidWorks Legion

 

[DaSalo]

Chris,

You've hit right to the heart of the issue. We do not have the appropriate technology to check this requirment properly and, instead of purchasing the correct technology, are trying to engineer a less expensive work-around using a hard gage.

I realize this is a really not a great application for a hard gage so what I am wondering is, if you were forced to do it that way, what would be the best approach to capture as much of the intent as possible?

If we were to request a drawing change from our customer what would be the best type of control to recommend?

 

[btrueblood]

"How do you define the profile of a zone with a complex shape and variable tolerance? "

Well, you would need, at least, a MMC and LMC profile gage for each section to be inspected. Advance gage to make contact, and then inspect the open gap, verify it is within allowable tolerances. Repeat for additional spanwise sections (requiring more gages from what you have described?). Gets expensive quickly.

 

[ctopher]

I agree with btrueblood.

Check into the expense of making tooling/gauges to inspect vs a purchased tool. You may be better long term purchasing the tool.
Machining/programing/material costs are not cheap, especially for a complex part.
To make your tooling, you will need to design them within tolerance of certain points of the part. Maybe go/no-go gauges.
I don't have an easier answer.

Chris
SolidWorks 10 SP4.0
ctopher's home
SolidWorks Legion

 

[DeanD3W]

DaSalo,
Since the part lacks sufficient stiffness for even minimal touch probe force to work, I don't think you will have much success with a hard gage either. I think a vision or laser measurement system will be the only good choice

Vision measurement might work well enough... Repeatability within a few (3 or so) microns is possible with typical vision systems, using focus to measure in the Z direction, and laser measurement can be more accurate and repeatable. Of course, both vision and laser measurement capabilities depend upon the material and surface finish, but either vision or laser should work quite well, after a bit of development. Since the part has only a slight twist it sounds like maybe you could leave the part fixtured flat and use vision or a laser to gather Z data at specified X, Y locations... The difficult part may then be analyzing the data to get what you want and tailoring the spec to clearly say what you want. Could you use measurement point data along lines that are perpendicular to the leading edge, then just best fit a line to each and limit the orientation difference between those two lines and the nominal/designed angle? This approach isn't using GD&T, but maybe your need is specialized and a different approach would work better..? Is there a line down the middle of the upper or lower surface that the twist rotates about, so that the line is nominally/designed straight? If so, maybe you can apply surface line straightness just to that line, to evaluate bow.

With a system that has 1000mm X & Y travel, you should be all set. Is there an inspection provider in your area that might have a system like this?

Dean

http://www.d3w-engineering.com

 

[MikeHalloran]

I would use a camera obscura.

Basically, it's a large pinhole camera placed at an angle oblique to the part, so that if the part is uniformly illuminated, an image of the part forms on a screen inside the camera opposite the pinhole. You don't want an 'end view' or a 'plan view', but a view midway between.

Now turn off the uniform illumination, and instead illuminate the part with a series of planar light sources, in planes normal to the part's long axis. You could do it with, e.g. an array of cylinder lenses, or multiple parallel laser beams oscillating so as to illuminate transverse sections of the wing.

The camera obscura's screen would then carry an obliquely distorted image of each transverse section of the 'wing', which could be recorded by using photo paper as the screen, or traced with a pencil onto ordinary paper.

Or, one could pre-plot the limits of the allowable outline on paper based on the ideal geometry and some trig to account for the oblique view and such.
Or, one could make an ideal part from rigid material, and photograph it (in the camera obscura) in order to make a master checking template.

Okay, it's not super simple or super cheap, but it's the sort of thing where one can trade time for money, construct it with simple tools, and get a usable implementation for not a lot of cash.


Mike Halloran
Pembroke Pines, FL, USA