What's new
What's new

Making/checking a square

Conrad Hoffman

Diamond
Joined
May 10, 2009
Location
Canandaigua, NY, USA
Doing this well seems more difficult than it should be. The go-to method seems to be a circular square, which can be made and verified without too much trouble, though best done on an OD grinder. Haven't got that. Does the following method make sense as another way of doing it?

Machine a small steel monolith, a rectangular or square bar. Grind two sides parallel, something that can be done very accurately, and can be checked with mics. Grind an end to whatever squareness can be achieved without heroic effort.

The bar can now be checked on the two parallel faces for squareness using an old height gauge with the V-front for a ball. Similar to this tool, but way less expensive to DIY yourself- Taft-Peirce 9146 Comparator Square A lot of old scribe rod type gauges have a V in the front where a ball can be seated.

Assuming the square is out a bit, a near certainty, perfect square will be halfway between the two indicator readings. Naturally the base surface can be scraped or reground until true, but it isn't absolutely necessary.
 
The problem I see with your method is that the height gauge isn't necessarily vertical. As a height gauge, the sine error from being slightly tilted is negligible, but would have a big effect on you. My understanding of the comparator square is that it transfers square from a master square, so the straightness/perpendicularity of the column doesn't matter. Since you are trying to make a square from scratch, this won't work.

Someone makes a perpendicularity checker that has an indicator running on a precision granite square, but I don't remember who.

If you want to make three of them, you could use your surface plate and check the right angle faces to each other.
 
Yes, Conrad, that will work. Once you figure out the error you have, you can step grind the top to use as a reference for regrinding the base, eliminating the error you find.
This method can produce true square to the limits of the measuring equipment you are using and your surface grinder. With care you should be able to produce something with an error of <.0002" over several inches.
 
The problem I see with your method is that the height gauge isn't necessarily vertical. As a height gauge, the sine error from being slightly tilted is negligible, but would have a big effect on you. My understanding of the comparator square is that it transfers square from a master square, so the straightness/perpendicularity of the column doesn't matter. Since you are trying to make a square from scratch, this won't work.

Someone makes a perpendicularity checker that has an indicator running on a precision granite square, but I don't remember who.

If you want to make three of them, you could use your surface plate and check the right angle faces to each other.

I'm not understanding how sine error has much effect here. He's using the indicator only comparing the two sides. They should be pretty close to square to start with so the readings will be very close to one another, I'd expect less than .001" difference.
 
I'm not understanding how sine error has much effect here. He's using the indicator only comparing the two sides. They should be pretty close to square to start with so the readings will be very close to one another, I'd expect less than .001" difference.

You're correct, I was misunderstanding what he was measuring.
 
Doing this well seems more difficult than it should be. The go-to method seems to be a circular square, which can be made and verified without too much trouble, though best done on an OD grinder. Haven't got that. Does the following method make sense as another way of doing it?

Machine a small steel monolith, a rectangular or square bar. Grind two sides parallel, something that can be done very accurately, and can be checked with mics. Grind an end to whatever squareness can be achieved without heroic effort.

The bar can now be checked on the two parallel faces for squareness using an old height gauge with the V-front for a ball. Similar to this tool, but way less expensive to DIY yourself- Taft-Peirce 9146 Comparator Square A lot of old scribe rod type gauges have a V in the front where a ball can be seated.

Assuming the square is out a bit, a near certainty, perfect square will be halfway between the two indicator readings. Naturally the base surface can be scraped or reground until true, but it isn't absolutely necessary.

.
sure surface height gage with test indicator where part is in contact with the base of surface gage AND parallel sided part is flipped on flat surface plate and surface gage brought into contact again with part side at bottom of surface gage is traditional way to check perpendicularity.
.
the old method is so old that many machinist need a $2000. or more gage to measure and have never done it with a $50 surface height gage with $50. test indicator. sure you can buy a more expensive height gage and indicator if it makes you feel better. you need to be in flat contact with front of surface height gage base as if you rock it reading changes of course if part is rectangular not a cyclinder
 
Flat surface of some length, true and flat.
Mount an upright perpendicular to the flat surface.
Attach two bearings to the upright at a distance that will span the blade of the square.
The bearings can be positioned left or right and then held.

Square inspection:
Set the beam on the flat surface and bring the blade up to the bearings.
Adjust the bearing that is loose till it has contact and does not turn, just like the first bearing.
Flip the beam 180 and bring the blade up to the bearings.
If the square is true both bearings will be in contact.
If not, the error can be determined via fine shim or paper.
John
 
Conrad --

Don't make the base flat, but recess the center a bit. You want the square-to-be's base to contact the surface table on narrow lands under the parallel faces. Once you know the deviation from square, correcting the error is a straight-forward matter of shortening the too-tall land.

I know of more than one square that was fine-tuned by "lapping" the tall lip with wet-or-dry paper on a cutout remnant of granite countertop.

John
 








 
Back
Top