Hello Guest it is April 19, 2024, 06:00:57 PM

Author Topic: Improving resolution on C axis  (Read 4877 times)

0 Members and 1 Guest are viewing this topic.

Offline derek

*
  •  200 200
    • View Profile
Improving resolution on C axis
« on: November 26, 2012, 02:23:49 PM »
Hi
I'm using my C axis as the drive for my tool change carousel. I'm using an extra servo motor I had with a 250 count encoder. I have a 4:1 belt reduction. Consequently I have to set the steps per inch at 27 (or so) to calibrate it at 360 degrees rotation. It's not too smooth and cogs like crazy. I put this together with parts I had on the shelf so before I order new pulleys is there a way to improve this in Mach?

Thanks
Derek

Offline BR549

*
  •  6,965 6,965
    • View Profile
Re: Improving resolution on C axis
« Reply #1 on: November 26, 2012, 02:36:19 PM »
250 x 4 =1000 , x 4= 4000 , /360 = 11.1111 steps per deg.  That gives you .1 deg per step.

What drive ?  Sounds more like a tuning issue.

(;-) TP

Offline BR549

*
  •  6,965 6,965
    • View Profile
Re: Improving resolution on C axis
« Reply #2 on: November 26, 2012, 02:42:14 PM »
 Try running the code in CV mode not exact stop ??

Just a thought, (;-) TP

Offline derek

*
  •  200 200
    • View Profile
Re: Improving resolution on C axis
« Reply #3 on: November 26, 2012, 02:42:38 PM »
Hi
My Mistake.
It's a 500 count encoder. Too much going on right now!
It's a rutex 990 drive. If I set the steps per inch to 2000 which is one revolution it's smooth as can be. If I had it belted at 1:1 it would be fine.

Derek

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: Improving resolution on C axis
« Reply #4 on: November 26, 2012, 04:37:28 PM »
You need steps per at double what Terry said seeing as your encoder is double what you first said, that is assuming you are wishing to use degrees.
There should be no difference between calling it a linear or rotational axis as far as motion smoothness is concerned so I would tend to think it is tuning as Terry suggested or possibly your feed rates are too low if you have it set as a rotational axis.
Hood

Offline derek

*
  •  200 200
    • View Profile
Re: Improving resolution on C axis
« Reply #5 on: November 26, 2012, 04:54:06 PM »
It seems to me 27 steps per base unit is going to be a problem regardless. If I could scale the A axis in rotational then I could theoretically use a .25 factor then set my steps at 2000 per unit. I didn't see any way to scale the A axis but then it may be hidden.
Or I can order new pulleys at 1:1 or even 2:1 I was just trying to save some time and expense.

Derek

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: Improving resolution on C axis
« Reply #6 on: November 26, 2012, 04:57:01 PM »
Where are you getting the 27 from?
Hood

Offline BR549

*
  •  6,965 6,965
    • View Profile
Re: Improving resolution on C axis
« Reply #7 on: November 26, 2012, 05:04:26 PM »
OK 500 x 4 =2000 , 2000 X4=8000 , /360 = 22.22 steps per deg. That will run smooth or should with a servo.  It will just be slower at a set kernal frequency.

I am not sure exactly what you are trying to gain , SPEED or resolution ? those are the 2 areas you are playing with.

(;-) TP

Offline derek

*
  •  200 200
    • View Profile
Re: Improving resolution on C axis
« Reply #8 on: November 26, 2012, 05:54:15 PM »
Where are you getting the 27 from?
Hood

In the setup motor screen that's the steps per unit setting I had to set to get it close to 360 when I command 360 from mach. Granted the math should be 22.22 per degree but I stopped messing with it because it was cogging so much at low speeds and because that number was so much lower than I'm used to for linear settings. Normally I'm used to settings of 8000 or more.
If numbers that low are normal then I'll look elsewhere for my problems. As you and Terry have pointed out it's probably tuning so I'll look at that.

Thanks
Derek

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: Improving resolution on C axis
« Reply #9 on: November 26, 2012, 06:20:50 PM »
The thing is it is per degree that these units correlate to, not per mm or inch.
If you set it up as a linear axis then it would be 8000steps per unit and a unit would be 1 rev.
Presume however you would be wanting to use the rotational rollover and shortest route on G0? If so then you need it set as a rotary, you just have to set the rapid speed accordingly so that you get the RPM you want.
Hood