[neilw20]

Out of square?
Are you running 2 steppers in parallel to drive each side of a router?
You need to ensure that they both stay in sync, or you will get out of square, as it 'racks' crooked.
If running two steppers in parallel, whether by winding or driving similar drivers you MUST have a method to get them in sync.

[quinnkyle]

I appreciate your advice! Thanks for taking the time.

Here's a video of the CNC in action.

I don't have a vacuum table, so since we're cutting small parts I brad nail both the 4'x8' sheet, as well as each part after the first cut, which keeps everything from moving.

I took one video last Thursday when I noticed some slight rocking on the Z axis as the end mill drove into material. I often cut "L" shaped forms. In spite of the Z axis getting tightened, I noticed the "L" forms were square on the outer side of the "L", but not on the inside. It has led to some costly waste in just one day! Not to mention additional wear on the end mill. Last week was the first time I noticed anything being out of square.

The video showing the entire table is using a cheaper (and louder) compression bithttps://youtu.be/6KeEpmCGVEc.

So far I've been so busy trouble shooting we haven't yet attached a vacuum, so we manually vacuum everything. I also disabled the limit switches, and will be working on rewiring this week. (I was getting a false emergency stop, which I recently learned had to do with Mach3 settings.)



Also, here are some samples of the types of things we're cutting.

[neilw20]

1. With machine stationary, put lots of force on each axis, and see how much 'free play' you can see.
2. Long axis.. Can you move one side and the other does not move?
3. Now bury the stopped cutter in the bed, at various places across/along the machine and try to move things.
4. I bet there is much more flexing than you expected.
5. What method is used to keep each side of the gantry in step?

I have been there, done that on a 12M x 5M x 2M, 5 axis monster, and got it down to 10 thou.
Use normally closed wiring on your limits so they don't pickup noise, being open circuit during normal operation.
Put a 100nF capacitor across each limit switch, at the switch.
Having wiring longer than 3M really needs opto couplers, at the controller end.

[quinnkyle]

Thanks - these are great guidelines! I did have it within a few thousandths of an inch in the beginning. If we can be in 1/64th I'd somewhat happy. Will let you know after some research tomorrow.

[wizard]

The pictures of your z axis don't inspire a lot of confidence in its rigidity. While this may or may not have anything to do with your out of square issues it leave me concerned with the state of the other axis.

In any event you will need some sort of dial indicator, a precision square and possibly other tools to find out where the issues are. You need to look for looseness in the system that allows axis to deflect under load. With the V bearings running on bangle like that, I suspect that you will be getting all sorts of deflection which probably varies based in position.

In simple terms if you can get an axis to deflect undera couple of hundred pounds of manually applied force it will deflect when cutting. The question then becomes how much deflection is acceptable to you.

[wizard]

1. With machine stationary, put lots of force on each axis, and see how much 'free play' you can see.
2. Long axis.. Can you move one side and the other does not move?
3. Now bury the stopped cutter in the bed, at various places across/along the machine and try to move things.
4. I bet there is much more flexing than you expected.

Flexing and out right looseness! I really think at least part of the problem is in the axis linear motion systems. They likely have loosened up after a bit of operation. The small V rollers don't inspire me either.

5. What method is used to keep each side of the gantry in step?

I have been there, done that on a 12M x 5M x 2M, 5 axis monster, and got it down to 10 thou.
Use normally closed wiring on your limits so they don't pickup noise, being open circuit during normal operation.
Put a 100nF capacitor across each limit switch, at the switch.
Having wiring longer than 3M really needs opto couplers, at the controller end.

The interesting thing here is the indication that outside cuts are still producing square surfaces but inside don't. It may be an issue of climb va conventional milling with the spindle being deflected in the wrong direction as the load increases. This would indicate a rigidity problem or gross looseness in the linear bearings.