Standard Test Pattern?

[jharvey407]

Is there a standard test pattern, shape or object that is used to measure the accuracy of a CNC machine?

I know if someone asks about accuracy I could cut some squares or shapes to be measured, but it seems to me that there is probably a standard test shape. Does anyone know?

Thanks,
James

Similar Threads:

• pattern
• Servo Test pattern
• A bit of topic but interesting IR remote test, using a standard digital camera.

[adprinter]

Is there a standard test pattern, shape or object that is used to measure the accuracy of a CNC machine?

I know if someone asks about accuracy I could cut some squares or shapes to be measured, but it seems to me that there is probably a standard test shape. Does anyone know?

Thanks,
James

My first test was a hand written G code file. But the same "Test Pattern" used for testing a computer printer for aspect ratio accuracy, would be a good test for your machine as well: Draw a perfect square in your vector graphics based software, or CAD program. Next, draw diagonal lines from corner, to corner of the square. (The point where these lines intersect should be the exact center of the square). Using the exact center of the square as the reference point, draw a perfect circle within the square, with it's outer edges touching the lines of the square (this should occur at the exact midway point of the walls of the square between it's corners). Save the graphic, and output to your CAM program to generate the g code file. Do a test cut on your machine, and measure the results. They should match exactly, the measurements used in drawing the graphic. If not, the results should give you a "starting point" from which you could determine any fine tune adjustments needed to correct any inaccuracies. Hope this helps.

[underthetire]

Look up "circle diamond square". That was the old standby for most all machine builders before lasers and ballbars. It was just as it sounds, a circle at one level, a diamond (just a square turned 90) at another level, and another square parallel to the axis. Most had holes in the corners to check positioning as well.

[louieatienza]

Is there a standard test pattern, shape or object that is used to measure the accuracy of a CNC machine?

I know if someone asks about accuracy I could cut some squares or shapes to be measured, but it seems to me that there is probably a standard test shape. Does anyone know?

Thanks,
James

Guys who've seen the IncraJig tablesaw fence demo video probably remember the 3-4-5 challenge... Rip three pieces of wood, of widths 3, 4, and 5 inches, and one of 12 inches, and stack the three smaller on top of the 12 inch piece. Not easy to do with a traditional rip fence. A lot easier with the meshing racks of the Incra....

One tough test (which my machine fails) is to cut a circle of known radius, and a pocket of same radius, and take the circle and try to put it in the pocket and spin it around. There's an Onsrud demo video where they do this... with 1" acrylic! They actually also have a video of circle-diamond-square.

I also saw a Mazak video where four shapes are cut in relief on a huge block of steel, and the negatives cut on another hiuge block, and they use a forklift to flip the other piece and it fit like a piston.

[Al_The_Man]

Industry tuning pattern.
Al.

[jharvey407]

Thanks for all of these ideas. I'll probably use them all.

Right now my machine is accurate and repeatable to less than .010", but I know I can acheive better.

These test will help.

J

[victorofga]

try to cut a large square like 24x24 and measure the sides, also diagonals...
with simple tapemeasure you can measure about 1/32..
if diagonals showing same on a 24x24 inch square, that precision already ok...

cutting a circle, as large as you can measure yet with a caliper, will shows the occuring backlash...

this is the simplest way..