Simple CNC MILL test part

Anyone have a favorite simple test part for cutting on a "for sale" used machine? I have done several; simple round post, and a square to put a micrometer or caliper on to check for backlash, parallelism, squareness etc... On-site inspection. Is a stainless part going to show any more machines flaws than an Aluminum part? Like to do something more than the onsite crowbar, indicator test. Seller is willing to let us run a part.

Though about adding a simple hemisphere to see the ballscrew and way wear and z-axis. I expect the spindle bearings to be worn. More concerned about wear on the box ways. 1/2 inch flat and ball end mills.

I can't be the only one to do this. Like to see others input.

Biggest id circle you can mill. You can really see the quadrants and the backlash associated.

Circle-diamond-square test. Mic and sweep the circle, mic the diamond and square, check the diamond corners with a precision square. Use a big, stiff endmill and take a spring pass.

Heard of, but never tried:

have the mill drill (or for bonus points, interpolate the holes with an end mill) a square pattern of 4 holes equal distance apart in x and y. Run 2 parts. Then see if they fit together with slip fit dowel pins, and again after turning one 90 and/or 180 degree's from how they sat in the machine.

Can tell if machine is repeatable, and if x and y movements are matching what the display is saying, without using measuring tools. If it fails that test, no it doesn't tell you where the problem is, but it just depends if you're shopping for a running machine or a project.

It's not a part... but I have a habit of always slapping an indicator on the spindle and sweeping the table. Then run the table to its positive and negative limits in X and sweep the table again at both ends.

PM

To check backlash:
Go into the control and find the settings for backlash compensation, note down the numbers. Then mount an indicator to the spindle nose with a 123 block bolted to the table, jog each axis up to the block, then away and note any lost motion. Add this to the numbers from the control for the *real* amount of backlash in each axis.

The above procedure does not isolate the source of the backlash, it could be thrust bearings or ballscrews. To isolate the source (if you really want to) pull back the way covers and use grease to stick a ball bearing in the center hole that is usually exposed on the end of the ballscrew, get an indicator on the bearing, movement here when changing directions indicates the amount of backlash caused by the thrust bearings, remainder will be in the screw.

Then inspect the spindle taper:
Blue up a new toolholder from a reputable source (Maritool!), clamp it in the spindle, then remove and look at the contact pattern. An old machine will likely be bellmouthed, but at least this will tell you how much by and whether there are any dings and so on in the taper. A drawbar force gage is also a good investment for machine inspection, however low drawbar force is not normally a big problem as you can simply rebuild the drawbar.

The reasoning:
Then take your test cuts and look for problems. The reason I like to do these separately is that the test cuts only tell you if there's a problem, but not what it is... If you take a test cut and you get a lot of chatter and so on, but spindle taper print is good while the drawbar force is low then you can negotiate on the basis of the machine 'not being able to take a cut' while knowing that rebuilding the drawbar is a relatively minor task and is likely to fix the problem.

Same goes with seeing quadrant marks on your test piece. You see heavy marks and you have indicated a lot of backlash when testing with an indicator, but testing the backlash on the screws tells you that it's all in the thrust bearings, you can negotiate down given the backlash but you know it will be easy to fix... The biggest issue with this process is just the time needed to inspect a machine to this level, however I think it's worth it as you will know much more about what you're getting into.

Just my $0.02.. I don't think test cuts by themselves are terribly useful except as a final peace of mind measure.