Thursday, November 1, 2012

Squareness Comparator Part 1

A month or two ago I wrote an article about squaring techniques on the milling machine. Since then I have posed the question to a few people I know and asked them how they would square something to very close limits without a standard of some sort. A couple of interesting potential projects came out of the discussions. It got me thinking about the entire squaring and verification process again. Squaring is such a recurring theme in machine work it deserves some detailed discussion and special tools. If you can master this one technique it will pay you back in so many ways on all the work you do. At one point I owned one of the really nice Taft Pierce squareness comparators and a cylindrical square. For what ever reason I let it get away from me in a moment of weakness. They are a pretty expensive tool versus how much you use them.
Both the square and the comparator are pretty simple tools to make. The cylindrical square is self proving by rotation if you have a surface plate and the comparator is just a fancy version of a re-purposed surface gage.
In this picture a test indicator is attached to the stem of the surface gage. The ball bearing sits in the vee in the front of the surface gage and acts as the pivot point when using the setup. The Taft Peirce commercial model has a machined in radius on the base that bears against the workpiece, or the cylindrical square when you zero it.
This works fine and uses stuff most folks already have around. I saw a nicer setup on a website where the toolmaker ground a groove in the surface gage and inserted a blade of spring steel with a curve in it. It was much nicer than having loose balls. After all the discussions and writing about squareness and squaring techniques I decided I wanted to make one for myself. For a month or so I doodled in my notebook working out the details of the gage making it easy and inexpensive to make.
Solidworks is a great tool when you have a pretty good plan already. All I used it for was to work out the finer dimensional details and make some really nice drawings. This is the final configuration from the study sketches and modeling work. It is loosely based on a Hermann Schmidt indicator stand that has a flexure in the front to make fine movements of the indicator.

For the squareness comparator a pivot is needed to butt the gage against the part to be checked. The indicator is moved up and down as required for the surface to be evaluated. The picot in my design is made from blue tempered shim stock and is the same thickness as the flexure.
Solidworks makes such good assembly drawings. If you have ever had to make a full assembly drawing on the drafting board or with 2D CAD its almost like cheating its so easy.
I have been collecting the materials for the project as the design was frozen for each of the parts. The base is made from pre-hardened 4142 ground flat bar. Its reasonable hard as it comes but is still machinable. The guide rod is a short length of linear bearing shafting. It is hard and smooth so the indicator mount slides up and down easily. You can even get them with a hole already tapped in one end. The flexure is blue tempered shim stock and is also used for the pivot plate.
I decided to start with the rod base first. It has two large chamfers on it that I have been thinking about how I'm going to set them up so its been on my mind. Here I'm using a large high speed steel roughing end mill to knock the thickness down. I like them because their quiet and don't throw red hot chips down your shirt. I can take off as much as I like in one pass with them also. So one roughing pass and one finish and I'm done.
Marking some of the major features with sharpie helps with sequence sometimes. I just got in the habit to help me do less thinking when I was at the machine. Do your thinking at the bench or the design office so you can move fast at the machine.
In the rod base there is a one inch diameter flat bottom hole. In general I really try to avoid flat bottom holes unless I plan on using a CNC machine. In this case it was easier to do one flat bottom hole that make several other parts or have to make the shaft from scratch. I actually tried something new to make the flat bottomed hole which worked out well so I'll share it with you.
I'm a big fan of Hougen Rotabroaches. They make accurate holes quickly and with less chips and arm power. Their claim to fame is in the annular cut like a hole saw. More of the cutting force goes into the annular cut instead of making the entire hole into chips like a drill bit. I have a bunch of these in various sizes so I tried to make most of the hole with the rotabroaches. The first cut was with the one inch diameter. This is the size of the actual indicator rod which might be risky to get an accurate hole plunging with an on size cutter. I have done this enough to know the limitations of the tool. Rotabroaches can be relied on to make holes within +/-.003 of the nominal size of the cutter. This is just what I wanted dimensionally anyway. Design around your machines and tooling and you won't ever be disappointed.
The first cut with the one inch diameter is already in. I followed with a three quarter inch diameter to the same depth to try and take out the remaining part. Remember this is a blind hole which I would never use a Rotabroach on.
The results were pretty good. It left a thin cylinder of metal in the center which broke when I ran the brush over the hole to clear the chips. For many flat bottom holes this would probably suffice. For this project I wanted a nicer bottom to the hole for the rod to sit on.
This is an interesting tool. It is a boring and facing head made by the Chandler machine company. When you hold the knurled ring at the top while its rotating the lower toolholder part of the head feeds outward. You can face the bottom of a hole like this or cut an internal snap ring groove in something that won't fit in the lathe. The alternative is using a boring head and stepping your way out to the finish diameter leaving a series of lines on the bottom of the hole. If you try to make a broad cut on the bottom with a flat bottom boring tool you risk serious chatter which you would have to go even deeper to remove.
This is the angle setup for the knob area. I laid out the angle and flat on the surface plate and roughed the angle by Kellering up to the scribe marks. This was to get most of the material out of the way for the end mill with a corner radius on it which I didn't want to use for major metal removal. It was important to have a rounded transition between the flat and the angle mainly for looks in the knob area.
Ready to mill the angled surface. You can see the roughing steps in this picture easily. The small hole is for the adjustment thumbscrew.
An afternoon in the shop. I took a couple of days off work because its my birthday. I borrowed a rule from somebody I used to work with that always took his birthday off. See the stuff I do on my birthday. I wouldn't be anywhere else given a choice.

More soon. I still have a couple of days off.


  1. I am following this project with interest, as this is something I have been meaning to make for years. At the moment I use a modified scribing block (surface gauge) very similar to the ToolandDieGuy's, except I have a doweled and bolted arc attached to mine.

    I notice that in the USA you like the Interapid DTI. I have never seen one in the UK, though they do one in metric. I like the swivel post idea on them. I use a Verdict T30 as my everyday beat around DTI, especially in the VMC in case of a mishap. One interesting feature of this indicator is it has an involute stylus, which in theory eliminates cosine error.

    I use a Compac Vertical, Mercer (both 10 micron clocks) and a Mahr (2 micron) DTI's for my inspection.

    As I use all DTI's as comparators, one thing (as well as not being anti-magnetic) that has put me off the Interapid is the 12 degree measurement angle. This can lead to the clock under-reading. In use I would always prefer the clock to over-read. I know we are only talking about 2% error, but I feel this is a design flaw.

    Once again, thanks for the post (I am also a fan of Solidworks, though I have a very old version, 1997 I think).

    P.s.: I must be robot, your verification to comment is a nightmare!