Visualize this, a piece of aluminum about the size of a credit card and about 3mm thick.

Here is the problem.

The part is machined under convention methods and received with pretty decent parrallelism.

One side (large surface area) must be as flat. We mate a silicon wafer on this side(similar dimensions). The flatness must be closely matched to the silicon wafer. (thermal issues)

I don't have access to expensive flatness measurement systems.

I tried to use an optical flat with a monochromatic light but can't seem to see any fringes.

Is it reasonable to use a somewhat calibrated straight edge against a light source to see whether it is concave or convex?

I realize that I have not mentioned any numbers on spec's , but lets assume the silicon wafer is nearly optically flat (but not hubble telescope flat)

Regards.

Quote:

---

Originally Posted by marcoawg

Visualize this, a piece of aluminum about the size of a credit card and about 3mm thick.

Here is the problem.

the part is machined under convention methods and recieved with pretty decent parrallelism.

one side (large surface area) must be as flat. We mate a silicon wafer on this side(similar dimensions). the flatness must be closely matched to the silicon wafer. (thermal issues)

I dont have access to expensive flatness measurement systems.

I tried to use an optical flat with a monochromatic light but can't seem to see any fringes.

is it reasonable to use a somewhat calibrated straight edge against a light source to see whether it is concave or convex?

I realize that i have not mentioned any numbers on spec's , but lets assume the silicon wafer is nearly optically flat (but not hubble telescope flat)

regards.

---

Your biggest issue is the thickness of your aluminum piece. 3mm is about an 1/8" thick. Trying to check for flatness mechanically is a problem because your part is going to immediately conform to the shape set against it. The usual method would be to apply bluing to the surface. Bluing is like dykem except that it does not dry. A very thin coat can be applied. A straight edge would then be drawn across the surface and the removal of the bluing viewed. If bluing was removed consistantly across the whole surface then your part is as flat as the straight edge. If not, the areas on which the bluing remains is not flat.

What I've decribed is a proceedure we use in our shop to check the flatness of a "thrust face" on our machined rotors. We have allowable concavity and flatness requirements of .001" or better.

If you could somehow apply bluing and then lay your sheet out in a "free state" meaning no clamping or other influences. Then bring an edge across the face that you could lower in minute increments so that you could monitor the removal of the bluing. You would be doing essentially the same thing.

Quote:

---

Originally Posted by marcoawg

Visualize this, a piece of aluminum about the size of a credit card and about 3mm thick.

Here is the problem.

the part is machined under convention methods and recieved with pretty decent parrallelism.

one side (large surface area) must be as flat. We mate a silicon wafer on this side(similar dimensions). the flatness must be closely matched to the silicon wafer. (thermal issues)

I dont have access to expensive flatness measurement systems.

I tried to use an optical flat with a monochromatic light but can't seem to see any fringes.

is it reasonable to use a somewhat calibrated straight edge against a light source to see whether it is concave or convex?

I realize that i have not mentioned any numbers on spec's , but lets assume the silicon wafer is nearly optically flat (but not hubble telescope flat)

regards.

---

It's me again.

I was giving it further thought. What if you made or had made a picture frame out of magnetic steel that was slightly larger (.01") wider and longer than your credit card shape and of various thicknesses starting with .005" thicker than your aluminum card.

If you then placed the frame down upon a magnetic table (such as those on a machine shop's surface grinder) and put the aluminum card with bluing face up inside your frame, you could turn the magnet on and run a straight edge across the face of the frame.

The frame need not be super flat but must be parallel. The magnetism will draw the steel flat. By experimenting with the thickness of the frame (thinner until bluing is removed) you can monitor the bluing removal and therefore the flatness.

I believe Starrett sells the kind of flat straight edge that you will need.

Good Luck!