Hobbyist Mill Conversion - Reasonable Accuracy

[RJS100]

I spent the last two years converting a small 2HP benchtop mill to cnc. I did not spend lots of money on high end ball screws or bearings as this is truly a fun hobby.

I spent quite a bit of time fine tuning the turns ratio for the ball screw as well as lash compensation. If I line up two 1x2x3 gauge blocks in measure their length with the KP3 probe, I consistently get results like 6.000 or 5.9998 inches. The mechanical backlash was .0008, .001 and .0012.

I recently milled a small aluminum block which is supposed to be .75" x .75" at the top profile.
The actual dimensions of the top of the block ended up at .7539" x .75375". I used a 3D adaptive tool path leaving .005 stock radially, then I used a contour path to remove .004 with a final finish pass of .001. I maintained the same stock position throughout the machining process.

Having no past machining experience and was wondering if results like .75339 and .75375 seem reasonable for a hobby machine?

Interestingly enough, I made a second block and the dimensions were .7546 x.7541

Thanks... Richard

[suntravel]

This is not ok, should be in a +-0.00039" range If everything is setup right on a benchtop mill.
Try a second finish pass, if it takes chips, cutting rates were too high and bringing the machine into deflection.
Small mills needs sharp tools and slow feed rates.
Also check if the gibs are not lose adjusted, quill is clamped, tool diameter is exactly matching CAM / tool lib.

Uwe

[RJS100]

Thanks Uwe. I did try a second pass and it removed virtually nothing.

I have been using YG1 aluminum specific tools and very conservative feeds. I used a 3 flute 1/2 inch endmill. For the 3d adaptive clearing tool path I ran 3000 rpm and 18 ipm with an optimal load of 12.5%. I left .005 radial stock. For the contour tool path I ran 3000 rpm and 18 imp with the first pass of .003 and final pass of .001. While these were the programmed speeds and feeds, I think when I ran the job I actually ran it slightly slower.

I re-checked the gibs and found the Y-axis a bit loose. I will try to tighten it up a bit.

I do have the diameter of the tools set to exactly .5 which could be the problem. This would make the part oversized as it is. I'm thinking of reducing the endmill diameter by 1/2 the error observed above. Does this make sense? This brings me to a final question. Is there a trick to measuring the diameter of a 3 flute high helix endmil?

Thanks... Richard

[suntravel]

You can use one like these for 3 flutes:

https://www.cnc-tool.com/store/p961/CNC ... .Z3.html#/

Or use the probe and touch of the probed surface.

Uwe

[Upnorth]

I think you may be expecting a lot from a hobby mill. Are you actually measuring to 5 decimal places? Being off by a few thousandths of an inch can be easily explained by deflection on a hobby mill with a small end mill with lots of stick out. I generally consider something within .005" to be ok. Obviously you need to be more accurate for something like a bearing fit. I'm usually within .002" on the first try though. If I need to be more accurate I make adjustments in the program after a test cut.

[suntravel]

I think you may be expecting a lot from a hobby mill. Are you actually measuring to 5 decimal places? Being off by a few thousandths of an inch can be easily explained by deflection on a hobby mill with a small end mill with lots of stick out. I generally consider something within .005" to be ok. Obviously you need to be more accurate for something like a bearing fit. I'm usually within .002" on the first try though. If I need to be more accurate I make adjustments in the program after a test cut.

Yes I do, but in metric units on a mill with 0,01mm resolution by default, for precise parts 0,001 mm resolution.

Uwe

[Upnorth]

I should have clarified my response was to the original poster. He quoted dimensions like .75339 and .75375. That is way beyond what can be expected for accuracy on a hobby mill.

[RJS100]

Thanks for all your thoughts. This is very helpful. The part I was trying to make had dimensions of .75 x .75 on the top portion and then the faces below were supposed to be recessed by .028, .029 and .030 respectively.

From your comments I understand that I could get the .75 x .75 much closer by adjusting the tool diameter (something I forgot about). However with a target accuracy of .00039 that Uwe suggested, trying to machine the other faces that only differ by .001 would be unrealistic for a hobbyist mill.

I very much appreciate all your help. Best... Richard

[Upnorth]

You can also change the finished dimensions by using negative and positive stock to leave in your tool paths instead of adjusting tool diameter.

[RJS100]

I am not sure what you mean by negative or positive stock? Are you saying negative or positive stock to leave?