Thread: Ideal method for inspecting flatness of a part feature.

I have a part that has a .005" flatness callout on one surface. the surface is roughly 7x7". There are flats on the opposite side to sit on the granite plate but not sure if there is a better way to test flatness. Is there is typical method for flatness inspection?

Map the surface with a drop indicator on a stand, while the workis being held .....ahh... err.....correctly.

If the back side has some odd geometry, the "classical" way to do this is to use (3) support points, typically a toolmaker-grade pylon type of support with the heights all ground to control the height dimension to .0001" or so, and flip the part so that the surface to be measured is contacting these 3 support points above the surface plate. Then set up your height gage or indicator stand with a suitable test indicator to sweep the bottom surface. The economy method is to just flop it onto the plate and run the DTI over the surface, as long as the back side doesn't have strange features that make this a problem. Using some parallels or the above-mentioned support points can be helpful for top surface measuring as well.

Originally Posted by viper

I have a part that has a .005" flatness callout on one surface. the surface is roughly 7x7". There are flats on the opposite side to sit on the granite plate but not sure if there is a better way to test flatness. Is there is typical method for flatness inspection?

I don't know how "typical" this is...

The trick is to measure the actual parameter that's toleranced, not some derivative based on other aspects of the work geometry. For example, you can't just set the work on a surface plate and measure the face, because the results are dependent on the flatness of the back and its parallelism with the face.

To that end you need to support the work at three points, and adjust those supports so the three points read the same height with your gauging system. Then you can measure other points to determine deviation from that reference plane.

Of course the gauging system would ride on a surface plate or on its own reference plane.

- Leigh

I agree with Specfab and Leigh: Support your workpiece on 3 points (such as a MATCHED set of 1-2-3 blocks), and sweep a test indicator over the surface that is in direct contact with those three points. The key is to indicate the feature resting on your 3 points (datum); and not indicate the side opposite the 3 points. If you do this, you're measuring the parallelism, and not the flatness.

"Flatness"of a surface can readily be determined by graphing mapped points of that surface regardless of tilt, tip or wedge relative to any other feature. No reference plane other than the one generated by the measured points themselves are needed. As mentioned, other qualities require other measurements and determinations.

I might be inclined to setup such workpieces with a spirit level.

Originally Posted by The real Leigh

To that end you need to support the work at three points, and adjust those supports so the three points read the same height with your gauging system. Then you can measure other points to determine deviation from that reference plane.

This is how I've done it in the past. I take three small screw jacks and place the part with the surface to be measured facing up and set it on top of the jacks. I then use a Sharpie to mark on the surface over the jacks (three little dots) and use a height gage while adjusting the jacks to get all three points to read the same. I then have a reference plane (through the three points) that I can measure the flatness callout from.

MO

Just remember Sharpie Juice has a thickness...

Originally Posted by beege

Just remember Sharpie Juice has a thickness...

True so let me refine that a little. I use the marks as a reference by placing the indicator tip next to the mark. Maybe using a mark like < would be better to make sure you are on the same spot each time.

MO

This is pretty much what _MO_ is describing. Very simple way to do it.

Originally Posted by beege

Just remember Sharpie Juice has a thickness...

Use Extra-Fine Sharpies.

The juice is thinner.

Alternatively, make a small circle with the Sharpie and put the indicator tip in the middle.

Of course, then we must determine the exact center of the circle.

- Leigh

Originally Posted by _MO_

This is how I've done it in the past. I take three small screw jacks and place the part with the surface to be measured facing up and set it on top of the jacks. I then use a Sharpie to mark on the surface over the jacks (three little dots) and use a height gage while adjusting the jacks to get all three points to read the same. I then have a reference plane (through the three points) that I can measure the flatness callout from.

MO

This is the correct way to check for flatness (without a CMM).

Originally Posted by beege

Just remember Sharpie Juice has a thickness...

Juice you will never see with a test indicator.


Tom

I'll use a set of parallels with an indicator sitting below the plane and hunt for the high and low spots. It's not pure, but I never had a problem getting parts past the CMM [inappropriate reference deleted - trl].

You can also give the part very light rub on black paper taped to a ground plate and check the differences between the high and low spots. Again, not pure, but if the part cleans up easily you can sleep at night knowing she's within .005.