Checking Flatness

dusty · Post by **dusty** » Tue Oct 16, 2012 9:24 am

I put this here in community because it has nothing to do with woodworking and little to do with maintenance or repair.

If you needed to ascertain the flatness (or straightness) of a surface over its entire length (and you are not in a precision machine shop) how would you proceed.

I have at my disposal a dial gauge, digital calipers, a reliable 6" steel rule and a reliable 4' straight edge.

I am going to attach a sketch of the surfaces I am working with, to aid in visualization of what I am trying to do, but the true nature of my question is meant to be generic.

BTW: I am preparing (that means playing in the shop) for the arrival of a Ryobi Sliding Miter Table which I intend to mount on the Shopsmith Mark V Model 520. I don't know why because I have a very dependable miter gauge already. I guess it is like I said - I am playing.

[ATTACH]18775[/ATTACH]

holsgo · Post by **holsgo** » Tue Oct 16, 2012 9:29 am

I just did this on my milling machine. Tramming is the term. You can use a dial indicator so that the plunger contacts the work, then swing to various locations taking readings establishing the high and low spots. This assumes, for example, that you can mount the DI to an arm which is mounted into the drill chuck if in drill press position. Then swing the arm and take DI readings. That's an example but seeing as I just did the process it's fresh in my mind.
Less scientific is a reliable steel rule.

pennview · Post by **pennview** » Tue Oct 16, 2012 9:40 am

A machinist's precision straight edge would allow you to do that. Here's a 24" one from Starrett that is flat to +/- .0002 per 12 inches -- http://www.starrett.com/metrology/produ ... ges/380-24

dusty · Post by **dusty** » Tue Oct 16, 2012 10:01 am

holsgo wrote:I just did this on my milling machine. Tramming is the term. You can use a dial indicator so that the plunger contacts the work, then swing to various locations taking readings establishing the high and low spots. This assumes, for example, that you can mount the DI to an arm which is mounted into the drill chuck if in drill press position. Then swing the arm and take DI readings. That's an example but seeing as I just did the process it's fresh in my mind.
Less scientific is a reliable steel rule.

This is SORTA what I just tried but I cannot isolate the cause of variations. I have the DI mounted on a moveable platform (on the upper surface). I am attempting to verify the lower surface. I slide the platform from end to end on the upper surface and see significant variations even though the two platforms seem to be "flat" using the straight edge. Variations approach .010".

wa2crk · Post by **wa2crk** » Tue Oct 16, 2012 10:45 am

Dusty;
If I interprate your setup correctly and according to your sketch the assumption would have to be that the upper surface is indeed flat to start with. Additionally, if the upper surface is dished or convex then the carrier for your DI would follow the deformity and this would give the illusion that the surface being tested was not flat.
To test the reference surface I would use a good quality level and use the DI and the carrier riding the level to establish the integrity of the reference surface. Any out of "flat" readings should be recorded. After this is concluded the reference surface could be used to check the other surface and at the same time compensate for any out of flat readings in the first test.
Should be good enough for woodworking.
Bill V

dusty · Post by **dusty** » Tue Oct 16, 2012 10:55 am

[quote="wa2crk"]Dusty]
I suspect that you have concluded that my reference surface (the top surface) is the Shopsmith Main Table. I started with the assumption that that surface is flat (enough) for this task. Maybe I need to confirm that. But remember, this is a learning task not a real project.

Thank you all for your inputs. Keep them coming.

nuhobby · Post by **nuhobby** » Tue Oct 16, 2012 11:02 am

While it's extraneous review material for many, this is still very interesting reading. If you've ever wondered how a really flat surface could have been first created in this imperfect world....

http://home.comcast.net/~jaswensen/mach ... _edge.html

and other web-searches for the 3 Plate method.

worknhard · Post by **worknhard** » Tue Oct 16, 2012 11:08 am

The simplest way I know to measure surface flatness is to use a surface plate (granite surface) and 3 machinist jacks that support the part being measured. Set each jacks the same height and place the surface to be measured on the jacks. Scan the surface with your indicator.

However, more specific to what I believe you are wanting to accomplish is perhaps measure flatness or maybe parallelism of the sliding miter table surface relative to the SS main table surface... to determine how high to mount your sliding miter table. You can do this using a good straightedge. Using the straightedge on you SMT and extending over the SS Main Table you can get an idea of the height variance over areas on the SS main table, then use this information to adjust your SMT to a desired height. When doing this, I have measured variances as high as .025 over my SS main table surface.

pennview · Post by **pennview** » Tue Oct 16, 2012 11:55 am

Along the lines of what Chris posted, there is a similar technique for placing screws into the edge of a three boards for making a precision reference device. Here's a reference to that procedure, but I recall having seen a much more detailed reference some years back -- http://www.sawmillcreek.org/archive/ind ... 25241.html -- it's mentioned in the eighth post.

Dusty, I'm not sure what the lower platform is made of, but the flattest material you're likely to have in the shop is a piece of MDF, provided it hasn't developed a bowed shape.

pennview · Post by **pennview** » Tue Oct 16, 2012 12:02 pm

Here's a more detailed explanation of how to make a reference bar using screws -- http://www.woodcentral.com/articles/pow ... _511.shtml