CNC doesn't provide the requested measurements

[Stan900]

Hey folks, can anyone help me out?

I'm almost done assembling a CNC machine, but I've run into a single issue: the machining is offset by approximately +-0.5 mm.

When I mill a pocket, it leaves it 0.5 mm higher, and when I profile, it's 0.5 mm lower. Sometimes it gives me more or less measurement, but generally it's around +- 0.5 mm. Occasionally, it gives me the exact measuremen

I've meticulously calibrated the motors (using Mach3) and even manually input exact coordinates, as I'm using 1204 ball screws.

I'm starting to feel desperate here. I'm programming with Vectric Aspire, and it's strange because I don't have any issues with another machine using the same software. It feels like there's something I'm configuring incorrectly.

Any ideas, guys? Could it be that the board is defective? I´m using the red bitsensor board

Similar Threads:

• Newbie- provide me help linux based cnc machine
• Provide CNC machining parts
• Provide CNC control software.
• provide CNC machining service

[routalot]

If the numbers in the code are correct then you need to take a look at the physical aspects of the machine.It might be a coupling slipping slightly or a motor mount flexing.In the worst case it might even be the structure of the machine flexing if it is one of those made of extrusions held together by brackets,but we can't see enough information to give more precise pointers.

[Stan900]

Thank you for the response.

I also thought it might be a flexing issue, so I reinforced the Z-axis, which was the weakest. However, it's still the same. The entire machine is made of 3/8" thick 6061 aluminum plates. It has 16 mm round rail guides. It has Nema 23 motors with 4.5 amps (something like that).

So I don't know, as I mentioned, I have another machine with less capacity, and it works without problems. I'm starting to suspect it's the board. I tried setting it to 0 and moving it to rough coordinates, and it does give them and returns to point 0, so I don't think it's steps lost either.

[routalot]

If the numbers in the code are correct,it has to be a machine related problem.Do you have the means of measuring Z axis travel?It would seem to be a logical step to move the axis over a large part of it's range and to physically verify the distance travelled.A height gauge would be ideal and the result might highlight calibration issues which could allow you to introduce a factor that would correct things.If the travel is correct things might get a bit more "interesting".