measuring squareness

[Tony]

came across this video on youtube on checking for squareness:
http://www.youtube.com/watch?v=0tu9uyRG8ZM

sounds like he's referring to it as the "bump" method?
Never saw this before and thought I'd ask the group re: opinions
on the method.

For that matter, how do you check squareness in your shop?

Personally when I've had to do this in the past I've just sighted
along a known good square -- never really tried it with a DTI.

I might just make that mod to my surface gage (the bumper).
Or maybe just magnetize a large ball bearing.

Tony

ps check this gadget out: http://www.youtube.com/watch?v=MlpuLE9Fgt8

[H8Allegheny]

Why not use a cylindrical square to begin with? Simpler with less set up.

Brian
Taxachusetts

[Void]

Tony,

I use a method very similar to the one in the 'tool and die guy' video. There are many variations on the technique he was showing in the video. Some people simply nestle an appropriate size ball bearing in the "crotch" of their surface gage. Not really necessary to magnetize the bearing and it may, arguably, be undesirable to magnetize it. One reason being it will pick up particles and make it difficult to make accurate measurements as time goes by.

Another one is to use the ball end of the rod as found on some surface gages. Gets a little trickier to get the DTI probe directly over the ball. In general it makes it more difficult to set up.

There are manufacturers of such gages. Hermann Schmidt made one, very expensive, and I don't know if they still do. Flexbar also made one (or sold someone elses) and I don't know if they still do.

I made a dedicated gage from the plans in the back of "Machine Shop Trade Secrets." It works very well and there are many variations and improvements that can be made on the design. It is a good simple project.

When done correctly the advantage of the method when using a DTI over the more traditional 'use a try square and look for light' is that you get an actual numerical reading. The numerical reading can tell you exactly how much needs to be adjusted in order to get it perfect.

As some people noted in the comments to the video it also assumes the sides being indicated are flat and parallel... or at least parallel. A perfectly made pyramid would indicate as perfectly square. A /_\ kinda shape may indicate to the same zero. To avoid this error one needs either a known-good square that one can zero out on first or measure the parallelness of the sides with a micrometer.

The other somewhat obvious requirements are that one needs a surface plate that is big enough for the work and the gage and the work needs at least two surfaces that are flat, smooth, and burr free.

For most work a precision machinists square is adequate for checking and setup. If you ever doubt the accuracy of your precision squares the method outlined in the video is best way to check it. You can also verify the squareness of the "inside" angle by using some thick parallels to elevate the base of the surface gage above the base of the square.

-DU-

[Carld]

Yes, a dial indicator and surface plate is much better than eyeballing it if your trying to measure the squareness.

[Mcgyver]

Another one is to use the ball end of the rod as found on some surface gages. Gets a little trickier to get the DTI probe directly over the ball. In general it makes it more difficult to set up.
-

some guys just grind a flat on the front of the gauge, i've just used the ball on the mast...afaik thats what its purpose is

[Void]

some guys just grind a flat on the front of the gauge

I like that idea.

I think a slight curve would work better than a flat. About the radius of the length of the surface gage base. My Starrett surface gage bases have much too lovely color case hardening to commit to a grinding... gives me an excuse to get another one.

-DU-

[Forrest Addy]

Usually I have a lot to say about topics like this but so far the material has been competently covered.

Naturally I have some remarks. One thing not mentioned is vocabulary. This is a "comparative" technique. Adding the radiused piece to the base of a surface gage turned it into a "comparator square." To be useful as a gage it has to be set before quantitive readings can be taken. Thus it doesn't measure directly. It compares the work to the reference whether to a separate calibrated reference or one developed from the work itself.

Another point the ToolAndDie Guy didn't cover is that a perfectly symmetrical polygon will register zero using the apparatus he demonstrates. The preceding tests would be for parallelism and flatness of the sides and flatness of the bottom. If the sides are flat and opposite sides are parallel then the comparator square can reliably quantify squareness of the sides to the base plane.

A height gage usually has a radiused nose on the base. Add an indicator and you have a comparitor square.

And of course there is a cylinder square. If accuratly made as a true cylinder and its end ground as a radial plane, when set on end the cylinder square is about as accurate as ever needed in the tool shop's measuring room. My 4" dia x 10" cyliinder square is about 0.0001 small at he ends and has about 0.0001" taper overall so it isn't perfect. It's a cheap import but it works for me - I know its errors.

BTW, just for brags. The ToolAndDie Guy spoke of a "couple tenths in 2" being as good as you need." When I was teachiing hand scraping ALL of my students hand scraped their cast iron project pieces to less than 0.0001" of flat, parallel, and squareness the first day. But we cheated. Hand scraping is far easier to control thant a climsey ass surface grinder. Swagger. Strut.

[WaveDude]

Can one of you post up some pics of how you're using the ball/round end of the mast. Are you mounting the indicator to the same mast? Wouldn't you worry that you would be moving it around slightly when pressing it against the block?

[Mcgyver]

Wayne, you might consider buying the current issue of our benefactors mag, Home Shop Machinist, lots of pics and description of this exact topic

[Forrest Addy]

You put your finger right on the problem, Waveguide. Squareness checks with a surface gage is a very delicate and easily petrubed evolution. The radiused plate modded surface gage ToolAndDie Guy demonstrates in the video makes it gravy.