[gta18]

Hi Guys!

My first post here.

I just finished cnc'ng my HF Micro mill and doing some tests on it. I am using mastercam x5 with hobbycnc driver and nema23 on all 3 axis(mach3 backlash is disabled for now).

I drew up 50 holes 5rows by 10columns with the same distances in between on my drawing.

I did the drill toolpath to be S-shaped left to right and going up one row every 10 (x) columns. I noticed that the distance between the holes get bigger (say 1mm apart and at the 9th hole its about 1.8mm already. this is true both ways and happens on both axis) it doesn't matter if its left to right or right to left. Its consistent that the beginning holes have closer distances in between.


Is this due to backlash? I would think if the axis is going in one direction backlash would not be a problem right? and if it was backlash I would think that the successive holes should still have the same distance in between.


what do u guys think? any suggestions?


thanks!
arnold

[RicknBeachcrest]

Have you gone through the axis calibration on the settings screen of Mach3?

[gta18]

I actually have done the calibration. I used a dial to calibrate the steps per inch on each axis.

Is there anything else that will need to be calibrated in mach3?

Thanks

[kawazuki]

Have you gone through the axis calibration on the settings screen of Mach3?

What you describe would point to a steps per error as its seems to be cumulative and the same on both directions - backlash would be exposed on direction reversals then be fairly constant whilst moving in the same direction. Hoss has a very good axis calibration video on his site when using he was using a modified X2

[pippin88]

A steps per distance error would not be cumulative like that.
If the machine were moving say 10% further than actually commanded then a 1mm move command would result in 1.1mm move. This would mean 1.1mm between each hole center to the next, though overall after 5 holes there would be .5 error (moved 5.5mm).

The opening post describes incremental error between each hole. (1mm spacing at start becomes 1.8mm by end)

Check all mechanical connections. Tighten screws. Put a line mark (sharpie/texta) on the ball screw and couplers to see if they are slipping

[Fastest1]

Also which version of Mach are you running?

[kawazuki]

A steps per distance error would not be cumulative like that.
If the machine were moving say 10% further than actually commanded then a 1mm move command would result in 1.1mm move. This would mean 1.1mm between each hole center to the next, though overall after 5 holes there would be .5 error (moved 5.5mm).

The opening post describes incremental error between each hole. (1mm spacing at start becomes 1.8mm by end)

That would be true if using incremental relative moves but if moves were relative to a reference 0,0,0 point wouldn't the spacing grow incrementally? I don't know - i'm thinking out loud?

[awerby]

Yes, if you measure to each successive point from the zero point the error would increase. But what the original poster seemed to be saying is that the distance between each successive point was progressively larger. I've never seen that happen, but a simple error in the steps per unit settings wouldn't account for that - I don't what would.

Andrew Werby
www.computersculpture.com

[gta18]

I solved the successive increase in spacing, I watched the mill closely while doing the holes(it was losing steps, I had the Gibs to tight.)

now I have another issue, I'll post the picture and see what you think.

On my drawing all 60 dots are aligned but see the picture on how my holes come out.

I am puzzled about my Y-axis backlash not being consistent.

[pippin88]

That would be true if using incremental relative moves but if moves were relative to a reference 0,0,0 point wouldn't the spacing grow incrementally? I don't know - i'm thinking out loud?

Nope.

As I explained a steps per distance error will result in each hole being a fixed distance apart.

Original poster described (I've elaborated):
Hole one to two = 1mm
Hole two to three = 1.2mm
Hole three to four = 1.4mm
Hole four to five = 1.6mm
Center distance from hole one to five = 1 + 1.2 + 1.4 + 1.6 = 5.2mm
This is only likely to occur from mechanical problems.

A steps per distance error of say 10% would give:
Hole one to two 1.1mm
How two to three 1.1mm
Hole three to four 1.1mm
Hole four to five 1.1mm
Center distance hole one to five = 1.1 + 1.1 + 1.1 + 1.1 = 4.4
The machine is commanded to move a set distance and does move a set distance. It's just that the distance the machine thinks it's moving and the distance moved in reality are not the same.

[digits]

I solved the successive increase in spacing, I watched the mill closely while doing the holes(it was losing steps, I had the Gibs to tight.)

now I have another issue, I'll post the picture and see what you think.

On my drawing all 60 dots are aligned but see the picture on how my holes come out.

I am puzzled about my Y-axis backlash not being consistent.

If you have variable backlash, IMHO you may have a loose fit somewhere between the screw & bearing block or nut and machine. Sometimes it will slip, until it binds, and then your backlash will be much larger. I've even had repeatable, but variable backlash on machines before caused by loose screws.

I would suggest you set up a plunge indicator and measure the real travel at 0-10mm in 0.5mm steps, both going forwards and backwards. You might find a sudden jump between two successive moves...