Techniques for Measuring Flatness

[Cullet]

My company has come across a customer issue dealing with the flatness of our product. We produce reticulated ceramic filters for use in steel foundries. We currently use what we call a "toaster" gauge which acts as a dual purpose thickness measurement gauge and as a flatness gauge. There are 2 slots of varying thickness (go/no go) in which the product must pass or not pass through. If the product is too thin and goes through the "no go" side, it is rejected. If the product is too thick and won't go through the "go" side, it is rejected. Our assumption was that the maximum degree of unflatness that a part could demonstrate would be the difference between the measurements of each slot. However, the customer is applying pressure to one side of the filter and measuring the length the other side rises off of a flat surface. I tried to argue that this is magnifying the degree to which the product is not flat but they insist their method is correct. I was hoping to find some methods other people use to determine flatness. Our current specification is +1.5mm.

Steve

 

[David DeLong]

Flatness is a surface condition using ASME Y14.5M-94 but there is a new 2009 standard where it can be a centre plane and even at MMC where your method would be appropriate. That will not be your case though. It definitely is a surface condition.

There are a couple of ways to measure flatness of the surface.

First, one would have to have a dial indicator on a stand, adjustable jacks and a granite table. Make sure that the surface that requires a flatness tolerance is facing UP with the adjustable jacks on 3 places about 120 apart near the perimeter on the bottom. Zero off the surface directly above the jacks to create a plane on the surface requiring flatness. Once that has occurred, then sweep the surface with the dial indicator and take the FIM or TIR shown on the indicator. That is your actual flatness of the surface. There is no plus flatness or minus flatness only a flatness value or range.

Should the burr be included in flatness? Most people agree that a burr is a separate condition and should be reported separately. Do not include the burr in your flatness report.

If you have a CMM, then take many, many points on the surface and the CMM will give one the best fit and then the flatness value. This method is not quite as appropriate as the dial indicator since the indicator will sweep or contact more of the surface than the CMM.

Hope this helpsl.

 

[Jim Wynne]

First, one would have to have a dial indicator on a stand, adjustable jacks and a granite table. Make sure that the surface that requires a flatness tolerance is facing UP with the adjustable jacks on 3 places about 120 apart near the perimeter on the bottom. Zero off the surface directly above the jacks to create a plane on the surface requiring flatness. Once that has occurred, then sweep the surface with the dial indicator and take the FIM or TIR shown on the indicator. That is your actual flatness of the surface. There is no plus flatness or minus flatness only a flatness value or range.

It seems to me that this method, if I'm understanding it correctly, measures parallelism of the top surface of the part to the surfaces of the "jacks" the part is resting on. It doesn't isolate the plane, in other words.
In order to measure the flatness of the surface, the part would have to rest on blocks (or some reasonably flat plane) with the target surface of the part facing down towards the surface of the plate and the indicator would have to be run across the surface from underneath. Unless I'm misunderstanding.

 

[bobdoering]

I agree with Jim. See attached (page 7 in link) for some clearer ideas.

 

[David DeLong]

It seems to me that this method, if I'm understanding it correctly, measures parallelism of the top surface of the part to the surfaces of the "jacks" the part is resting on. It doesn't isolate the plane, in other words.
In order to measure the flatness of the surface, the part would have to rest on blocks (or some reasonably flat plane) with the target surface of the part facing down towards the surface of the plate and the indicator would have to be run across the surface from underneath. Unless I'm misunderstanding.

Jim:

The method that I described absolutely does not measuring parallelism.

Please not that I stated the one should use adjustable jacks on the bottom surface and zero off on top of each jack on the top surface. We are creating a plane using a 3 point set up on the top surface and not the bottom surface.

If one wanted to measure parallelism, one would use solid jacks on the bottom surface and then sweep the top surface.

Maybe it is the way that I described the method that is confusing.

 

[Jim Wynne]

I agree with Jim. See attached (page 7 in link) for some clearer ideas.

The only other way I know of to isolate a plane on a surface plate is rather crude and won't work in a lot of situations. You can lay the part on the surface plate and measure from the plate to the highest spot on the top surface of the part (determined by using an indicator to find the high spot) then subtract the thickness of the part from the measurement. What's left is flatness error. This is rather imprecise and should be used only when you can't invert the part (or don't have a CMM) and the flatness allowance is relatively generous.

 

[David DeLong]

Zero off the surface directly above the jacks to create a plane on the surface requiring flatness.

In my original explanation I used the term "zeroing off the surface" which means that one would have to adjust the 3 jacks on the bottom surface so that the points above the jacks on the "flat" top surface must be the same height or thus, zero.

Sometimes terms are not always understood and this term is really and old term.

 

[Jim Wynne]

In my original explanation I used the term "zeroing off the surface" which means that one would have to adjust the 3 jacks on the bottom surface so that the points above the jacks on the "flat" top surface must be the same height or thus, zero.

Sometimes terms are not always understood and this term is really and old term.

Just to be clear, you're doing this zeroing on the surface of the part (as it rests on the jacks) and not the surfaces of the jacks themselves? If this is the case, why do you need the jacks? Why not just rest the part on the surface plate and do as I described in my second post in this thread?

 

[ISOsandy]

QUESTION
Do you have an actual flatness requirement?
Usually flatness is not a +/- tolerance, you are allowed so much for flatness and thats it.

+/- 1.5mm is a reasonable amount of tolerance, however not knowing the product dimensions, material or processing information it is difficult to give you an opinion. What feature does the +/- 1.5mm apply to?

My experience has been the use of 3 points or jacks, locate on the surface to be measured, then use a test indicator across that surface.

CMM is another method

Depending on the part and requirements I have also placed the part directly on the surface plate and used shims or pins... again depending on all the information.

 

[JAltmann]

David,

Great post! First time many years i have come across someone who knew this method of using the 3 adjustable jacks. This includes a list of some rather knowledgeable people too.

Ahh the lost art of surface plate layouts.