Hobbyist Mill Conversion - Reasonable Accuracy

[Upnorth]

I am not sure what you mean by negative or positive stock? Are you saying I should adjust the dimensions of the model?

No it's where you set up your CAM. You make the model in full size. When you are generating tool paths it is a setting in the cam that I use. You tell the tool path to leave or remove stock as required. I use Fusion 360 to generate my CAM.

[suntravel]

Here you can see that a cheap hobby mill can probe with <+-0.005mm deviation.
So, it is no problem to mill with < +-0.01 mm deviation in the first run

But I must admit my profession is building machines and making parts for the automotive industry.

On my hobby mill are only cheap CL7 ball screws and CL Stepper from Stepperonline, but with 2:1 belt reduction 4000 pulse/rev.
The gibs and slideways are reworked with a little hand scraping.
I use screw compensation and now the new volumetric compensation.

To keep the mill precise, it needs regular maintenance and readjustment, Chinese cast iron is very soft and wears fast.

I use industrial grade tools and keep the cutting rates under a point where deflection becomes a problem.

In fusion you can use stock to leave or tool diameter to manipulate the size of the part, or use radius compensation in the machine and manipulate the tool diameter in the offset lib from CNC12.

Uwe

[RJS100]

Thanks for all input here. This is exactly what I was looking for. Considering all of these posts and looking back at my original figures, it looks like my measurements for the part varied by .00036 for the first part and .0005 the second time I ran the part. These are very close to Uwe's figure that I should shoot for .00039 tolerance. With a bit of gib adjustment I think I will certainly be in the .00039 ish range. Also, by adjusting the tool diameter and tweaking the stock to leave I belive I will be able to get very close to a target dimension.

I am still waiting to receive the precision ring gauge that I ordered so this should help with more precise probe readings.

Thanks again for all your help... Richard

[Muzzer]

Adjusting the cutter comp and setting backlash values is one way to frig the results but ultimately minimising mechanical backlash in the mechanism should always be your first step.

I'd recommend you use Swissi's Probe App to tell you what your mechanism is capable of. As Uwe says, careful setup and maintenance is essential. The results from the app may be rather depressing but you need to know your starting point.

[RJS100]

Thanks... Will do.

[Upnorth]

I'm impressed with your accuracy Uwe. I have not seen one yet that was consistently that good.

[RJS100]

One more thought regarding accuracy. To further test the limits of this mill I am getting ready to machine a 2.5" Turners cube as shown below. I thought it would be a good idea to use my new KP3 probe to see how well the mill was trammed. The table measures 8 1/4" x 33". I measured four points on the table approximately as show below with the back left corner set as a zero reference. The others points measured .0004, -0.0006 and -.0007.

Does this amount of variation seem acceptable for a hobbyist mill? These figures suggest to me that the table has a bit of twist.

Thanks as always... Richard

[suntravel]

This is not the way to measure if x y is straight and 90 deg

Place a precision toolmaker angle on the table, align it with a dial indicator in X on both ends to 0
Measure with a dial indicator set to 0 on one end of the other shank, if it stays to 0 while moving with y

If other end of y is not 0, you have the axis not in 90deg
If the dial indicator is moving while driving x or y the axis is not moving straight

On a mill like yours you are busted, no easy way to correct this with hardware.

But with volumetric compensation easy to compensate in CNC12.

Uwe

[RJS100]

I just purchased a precision square. Just to confirm that I understand you properly:

1. Place the square along the x axis. Run and edge finder up and down the square and confirm that the edge finder indicates a constant value. If it does not, adjust the head until the edge finder shows a constant value while running up and down the square.

2. Repeat this on the y axis and adjust the nodd if necessary, until the edge finder shows a constant value while running up and down the square.

Can you please confirm that is correct.

Best... Richard

[suntravel]

Here is it better explained than I can do with my broken English

https://www.haascnc.com/service/trouble ... ries-.html

Uwe