Hello Guest it is June 18, 2019, 01:51:14 AM

Author Topic: Machine Missing Steps? Direction on Corners  (Read 1535 times)

0 Members and 1 Guest are viewing this topic.

Machine Missing Steps? Direction on Corners
« on: April 18, 2015, 12:55:34 AM »
Hi guys
This ones been bugging me for ages. My Machine, a mill conversion, is going backwards occasionally at corners. Only for a step or two and on the odd occasion like a wild untamed thing.
I have been experimenting with cv, and shuttle numbers with good results, and last night after thinking i had it, it burst onto the scene again.
Details
A HobbyPro CNC kit.
3 to 1 ratio on the drives to driven.
ball screws with oil groves milled on all surfaces
(I don't believe it is an issue with power/torque)
At high speeds(above 200 mm/min) it doesn't seem to worry it. Max is 700mm/Min
low speeds 50-100 really make it appear.
last night I altered the step pulse to 1uS from 3uS and it appears to have corrected it!

if i change the feed rate it appears to fix it or make it worse depending on where I change the Shuttle numbers or step Pulse width too.
Any ideas on removing this bug in my system would be greatly appreciated.
I have access to a CRO if anyone can suggest where to look if it may be a noise issue
This is my Profile File below if you want to check it.

Also if any one knows, how can you get the job path to reflect the actual directions of the machine. Mines about face.

Offline BR549

*
  •  6,871 6,871
    • View Profile
Re: Machine Missing Steps? Direction on Corners
« Reply #1 on: April 18, 2015, 09:58:45 AM »
Mach3 uses Standard toolpath orientation for toolpath display . If you think it is backwards then your Axis directions are backwards. It is based on Tool travel. not neccesarily on  Machine travel.  For example bed mills run backwards compared to knee mills.

The drives sound like they are running into midband resonance. You can add dampners or change drives to ones that do midband corrections for resonance.

OR you acceleration setting it too high for your system. Reduce by 25% then retest.

Just a thought, (;-) TP

(;-) TP
Re: Machine Missing Steps? Direction on Corners
« Reply #2 on: April 18, 2015, 05:30:17 PM »
Hi TP, I have sent the accel. Up and down however as I change it I find I have to alter the shuttle numbers and if it's to low it appears worse.
Could you explain the dampners idea? Or should I just use google?

Offline BR549

*
  •  6,871 6,871
    • View Profile
Re: Machine Missing Steps? Direction on Corners
« Reply #3 on: April 18, 2015, 09:41:08 PM »
I am not sure WHY you are fiddling with the shuttle settings as that ONLY deals with backlash responce.   You should be able to do a seacrh on this site for dampers.

I am not sure of your steps per settings but with 3:1 gearing and possibly micro stepping and then Screw gearing it may be too low of a final gear ratio for  steppers ??  You want to keep the motors inside of their torque band.

Just a thought.
Re: Machine Missing Steps? Direction on Corners
« Reply #4 on: April 18, 2015, 11:40:35 PM »
Hi BR549,
As it occurs at the corner this is where the backlash switches in after a 90 degree prior turn and the shuttle setting affects this point e.g. Cutting a square or rectangle etc etc. No micro stepping, only using full steps, and if the speed was an issue I would expect it at a high speed as opposed to a lower speed.

Offline BR549

*
  •  6,871 6,871
    • View Profile
Re: Machine Missing Steps? Direction on Corners
« Reply #5 on: April 19, 2015, 01:45:41 PM »
Full steping is where most steppers have a midband resonance problem . I t is not always SPEED(Velocity ) where a problem shows up it is Accelleration that causes the most problems. I believe there is some info on resonance on the Gecko website.

Does your drive have the option to microstep ?? 

(;-) TP
Re: Machine Missing Steps? Direction on Corners
« Reply #6 on: April 19, 2015, 05:49:30 PM »
The resolution I have setup is to set the step and dir to 1uS (down from 3uS) and this seems to have resolved the issue. Why? I don't know. I am still getting tool walk on cutting shapes with multiple cuts in z axis, however I think this is related to backlash, and resolution of stepping being set to full steps. I have  0.0083mm (0.000328")resolution and the the backlash is between 0.001" and 0.0013" , this however leaves me with a mismatch on compensation(division of resolution into backlash) and I believe this accumulates in multiple cuts around the periphery of a square for example. I also have discovered I may have a little play in the tool head and I will being trying to eliminate this today if it is the case. I will also try half steps to see if this will bring the accumulation error down further. And I will have a look at resonance issues BR549.