Hello,
I new here on this forum so I'm going introduce myself quickly: I'm a 25-year old engineer from Belgium and recently I changed to Mach3 because the other CNC programs I worked with (DinCNC, PCNC) didn't support the use of more than 3 axes on the same time.
Last week, I noticed a problem: at the start of my milling program the mill is standing off 1 mm of the workpiece in the Y-direction. At the end of the milling program I return to the same start position (correctly programmed)  but in reality the mill is standing off more than 1 mm.
So I programmed a simple program in which the mill moves away 2 mm of the workpiece and than returns to its start position. Yet, the same problem occurs. I thought, maybe the speed is too high, but lowering the speed gave still the same results.
Is this a known problem? Is this caused by the change in direction (+Y directly followed by -Y) or is there some parameter I have to adjust? Entering a backlash ain't going to be the solution I guess (compensation by the move in the two directions).

Kind regards,

Kevin

What you describe is "textbook" a backlash.  What makes you think that using backlash compensation will not fix your problem?  Have you tried?

Regards,
Sid

Dear Sid,
maybe I didn't explained the problem accurated enough. I have written a simple program in which the mill moves 2 mm away from the workpiece in the Y-direction and than it returns to the start position, away and back again (so 2 times moving away and coming back). The more I run this program, the further the start point moves away from the workpiece and it's initial position.
Entering a backlash isn't a solution because in this case the start point moves further away depending on the number of cycles (away and back). If the movement of the start point, compared to it's initial position, would stay the same independent of the number of cycles, a backlash would be the right solution. Altough, that's what I understand under "backlash". To prove this, I entered a backlash and ran the program a couple of times. Result: the start point kept moving away from it's initial position.
Changing the speed, acceleration, pc... had no influence.
I also ran the same (adjusted: y becomes x) program for the X-direction. The same story here; the more cycles, the further the start point moves away.
To find out what triggers this problem, I ran the same program (both for x and y) with DinCNC. Result: the mill comes back to the same position, independent of the number of cycles. Butt in the future I need to run 4 axes at the same time and this isn't possible with DinCNC. So I downloaded Mach2 and I ran the program again. Result: the mill returns to it's initial position.
Problem not solved, but I got around it.
Although, I'm still wondering what went wrong with Mach3. I double checked all the parameters, but none of them indicates some sort of delay. If someone has any suggestion on how I can solve this Mach3 problem, just reply because I'm still very curious.

Regards,
Kevin

 

Hey Kevin,
   I don't understand why it would do it with Mach 3 and not do it with Mach 2. Have you tried to calibrate your steps per in Mach 3?

Brett

Brett,

I think I calibrated the steps, but now you make me doubt it.  I'm not 100% sure of it. I tried so much the last few days. But indeed, it is strange that everything works well in Mach2 but not in Mach3.
For the moment, I'm happy that I can run the program properly. Time is money you know.
I bumbed also into a problem using Mach2 today. Under "Motor Tuning", I give the x direction a velocity of 200 units/s. However, when I jog in this direction at 100%, I get a much much higher velocity (not sure how much). Jogging at 1% is still relatively fast (50mm/s). Any clou what could be the problem here?

Kevin

Not really Kevin other than maybe the steps per differ in two and three. I have never messed with two. I have sent you a personal message.

Brett

This is very strange, even if the steps per were wrong the tool would come back to the same point.

Are you using cutter compensation in your program?  If so are you cancelling with G40 before the end of the program?

Graham.

Is the error increment exactly the same each run thru of your testing program?
Is the end point shy of the start point or "ahead" of the start point?
Is the motor/lead screw coupler (whatever you have slipping?
    just a couple thoughts for you....gpdc