Another issue with circles

I read a number of threads related to circles not coming out round but none of them seemed to explain exactly what I am experiencing so I thought why not add one more…

I am working on cutting out the bottom of an ammo box for 9mm rounds. This picture is of my first test cut. It came out pretty good and the holes were accurate to my model but they were a bit small.

After doing a couple more tests in pine I found that the idea hole size is 10.2mm. With that all figured out I moved on to cutting it out of cherry. Just to be safe and to improve surface finish I dropped the feed on all of the operations by 100 or 200 mm/min (depending on what the operation was).

I cut the part out of cherry and the surface finish was acceptable and everything seemed to go well. I did some minor cleanup sanding and then starting inspecting the part.

First issue… The thing was not square. It was about 1mm out of square over a distance of ~200mm.

Second issue… The holes were not round. Cutting a 10.2mm hole was previously giving me about .5mm on all sides of the bullet. Now in the y direction I have zero clearance (it is actually an interferance fit) and in the x direction I have about .75 - 1mm on either side.

I adjusted the machine to square by snugging of the x/y tensioner bolts on the center section and manually setting the steppers to be square before powering the steppers on. Now the outside of the box is square.

I did a bunch of tests boring just the holes in cherry… slowing the operation wayyyyyyy down (from 600mm/min 100mm/min)… Doing a repeat pass to account for tool deflection… moving the part to the far back corner of the machine (that area rarely sees any action)… rotating the piece of cherry so the grain is running along the opposite axis.

The only thing that seemed to make any difference at all was rotating the wood. Things were slightly improved but not enough to say it wasn’t just a fluke. After all of this, I noticed that the wood was burning and the bit was heat stained so I ran to HD and picked up a new freud spiral up cut 1/4" bit.

I ran a test with the brand new bit in pine running real slow after making sure the gantry was square and the result was a box with square sides but holes that are not round.

You can find pictures of my build here New MPCNC build in Detroit.

Parts were all printed by a friend of mine. I’m running a RAMPS 1.4 with the standard (non-endstop) firmware and some no-name A4988’s I got on amazon in a printed case with two 40mm fans.

The part was designed in fusion 360 and CAM was done in fusion as well using guffy’s post processor.

When I get home from work I can upload some pics of the holes I am getting now.

I am running out of ideas so any suggestions would be much appreciated.

2 Likes

Great bug report!

Not sure how to help though. It sounds like a limit in milling cherry. Is it crazy to make the circles ellipses and make the Y direction 0.5mm larger? Not happy about that answer.

That is actually my plan if I cant come up with a better option. The only issue is that now cutting those same bores in pine also produces an oval. They used to come out round in pine… maybe not 100% perfect but they weren’t visibly ovals.

Here are some images showing my results.

A. (first pic) The first test in pine with 10.2mm holes
~.050" play inside bores along x axis
~.025" play inside bores along y axis

  • The dimensions of these holes are prefectly acceptable for my purposes.
  • Looking at them its extremely hard to tell they are not perfectly round

B. (not pictured) First attempt in cherry with 10.2mm holes
~.030" play inside bores along x axis
~.007" play insidebores along y axis

C. (second and third pics) Second attempt in cherry with 10.2mm holes
~.030" play inside bores along x axis
no play at all along y axis

  • Bit was dull at this time

D. (last pic) 10.2 mm bores with brand new bit and boring feed set at 100mm/min
~.040" play inside bores along x axis
~.005-.010" play inside bores along y axis

After measuring all of this it looks like I have a fairly consistent difference of .030" of play between the x and y planes. I took those measurements by sticking a bullet in the bore and moving it back and forth with a dial indicator touching the side. Obviously, they are not perfect… more of just a reference.

check for flex in your x-y gantry. If you grab the top of the Z axis, can you flex it in one way or the other?

I had an issue with corners being rounded over, but only when moving in one direction. I finally discovered that my XY axis could be twisted in the X direction but not the y. For most tasks, it didn’t do anything, but if I was applying too much pressure in a cut, then the entire Z axis wouldn’t be vertical any more.

If you have a lot of space between your spoiler board and the bottom of your bit, you could try putting another block of wood under the board your cutting to reduce this torque affect (basically move the part closer to the xy gantry so the Z axis isn’t sticking so far down.

Maybe it’s not the cherry wood and something happened to your machine between A and B.

Install a pen and try a ruler and maybe also a crown. It will tell you if you have backlash or an overall scale issue, or if the movement is uneven like from an eccentric pulley (hypothetically).

3 Likes

Really uneven tension in your belts?

That’s what I was going to say, sounds like uneven belt tension or something loose like a stepper gear. VectorTest.zip (5.6 KB) I have added a zip file with a DXF vector drawing that will let you know what is out of alignment or at least what axis is causing the issue. I use it with our lasers to make sure everything is running true.

1 Like

Thanks for the replies everyone. I will give your suggestions a try and see what shakes loose.

I made no changes to the machine setup Other than cleaning all of the rails with some wd-40 and ran each of the tests at 20mm/sec and again at 5mm/sec. In all cases the 5mm/sec version was much smoother.

The x axis of the ruler is a dead on perfect match for an actual steel ruler I have. The y axis has slightly uneven spacing but the total error over 200mm is about double the thickness of the line. It does look like there could be some backlash issues with the y axis I guess… not quite sure.

2 Likes

Do you have arcs turned on or off? You are still making that with fusion CAM?

Arcs are turned off. The CAM for the piece I am making was generated by fusion.

Hello, Im having similar isues. Any solution or idea?

I wasn’t able to find a good solution. I ended up just boring the holes out manually since I was only making one finished piece.

1 Like