Re: [CAD_CAM_EDM_DRO] Measuring accuracy

I just found out that the accuracy of my setup won't be defined my the
mill but by the software I'm using. I've looked at all of the sub-1k$
software and they all seem to work the same. There is always some
rounding in the numbers. Much more tolerance than accuracy in the mill
would cause. If I write the G-code, it's not a concern but if I use a
software to generate G-code I'll be lucky to get +/-.05mm before I even
start the spindle.

The error I'm talking about is when using a .25mm step over and my part
edge falls somewhere in between a step. Sure, I can cut parts at .01mm
stepover but who is going to wait weeks for a part to run on the mill?
If I want something pretty to look at, it's not a concern, but if I want
to use the part in an assembly, it's unreasonable.

Am I missing something?

Greg




Jon Elson wrote:

>>
>>Gregory Kamysz wrote:
>>
>>
>>
>>
>>>What kind of test does one run on a mill to define it's accuracy? Is
>>>there a standard code or method used to do this? Or do I just run some
>>>parts and see if I've made scrap or something useful?
>>>
>>>
>>>
>>>
>
> One of the other tests I do, with no special equipment, is checking
> orthogonality
> of the spindle to the actual X-Y plane of motion. (My table is banged up
> enough that there is no way to easily place a flat on the table repeatably.)
>
> I wrote a CNC program to move in a complete circle of some diameter.
> When done, it returns to the center of the circle.
> I use a 1/2" end mill to cut the circular track in a piece of scrap. I then
> mount a dial test indicator on a copy of an "Indicol" holder, and sweep
> the circle. If the height is the same all around, the head is "trammed"
> correctly. If not, the head needs to be realigned.
>
> Because of concave wear on the bottom of my table, I get a saddle shaped
> curve to the circular track. It is a small, but detectable defect in this
> 65 year old machine.
>
> Jon