Problem with jerky cutting

Androcles · Posted at 2018-5-6 17:32:07

I just recently got my DDCS V2.1 and have been setting it up and doing some test cuts.

I am having some problems with jerky moves when cutting

I have uploaded a video of it here: https://streamable.com/89y5h

I have tried with very low acceleration (100mm/s^2) and high (1000mm/s^2) and some values in between.

It seems as it is en exact stop mode G61 but inputting G64 (CV mode) does not change anything?

Any ideas?

Thanks

Best Regards
Andreas Landsfeldt

PPRINT · Posted at 2018-5-8 10:26:48

I just ran the code on my machine and did not see any reduction in speed.

I will link a video below.  I ran the first 2:40 minutes at 300% override (est 7200mm/min) until the program made it to the last z height and started it's inward morphing.  At that point it was pounding my machine pretty bad so I slowed it back down to 100% override.  I didn't see any reduction of speed.

If I ran this sort of code all the time I would have to spend time tuning the parameters in the control and my drivers to get that pounding out.  Since I am not worried about speed on this hobby machine I will leave that for another day.

https://youtu.be/RGxIPwSjoWE

Turn the sound down unless you love the scream of a cheap air-cooled router motor.  It does start sounding like a scene from Jurassic Park once it gets to pounding in the bottom.  LOL

71taa · Posted at 2018-5-6 20:22:10

What CAM/postprocessor is being used? Helical moves are disabled?

Had severe stepper shakes before in case of different/wrong settings of pulses per mm for X and Y. Haven't tried to solve it - as it was test setup with random settings for drives not attached to ball screws.

Also it would be helpful if you are able to share some of your test files (smallest ones).

Androcles · Posted at 2018-5-7 20:05:24

Last edited by Androcles In 2018-5-7 20:08 Editor

I am using Solidworks with HSMworks ie. Fusion 360. I have tried a lot of different post-processors including the one listed on here http://www.brainright.com/Projects/CNCController/

I have tried with different pulses/mm - 200, 400 and 800 pulses/mm, and oddly enough at 400pulses/mm one of the points where it "stopped" disappeared but the other ones remained.

Here is a video of the machine cutting the file that I have attached.
https://streamable.com/kaax1

It is very odd how it is in the middle on both sides but not in any other places.
I have disabled circular motion in more than one plane in the postprocessor along with helical moves.

EDIT
I can't see what the problem could be with the toolpath from HSM
I attached an image of it also.

PPRINT · Posted at 2018-5-8 01:16:28

Last edited by PPRINT In 2018-5-7 11:20 Editor

I have seen similar problems with a project that I have worked on with my DDCSV1.1.

I believe I have tracked the problem down to micro moves. Meaning tons of little moves on the order of magnitude of .01mm. I am guessing there is some sort of issue to do with the accel/decel settings.

I normally post my code with a very tight linearization tolerance to give me very accurate results. Problem is that means tons of small moves. I am normally cutting foam or wood so the tolerance was not necessary. Once I relaxed my tolerances I now get code that will run smooth.

I know that is not an answer to your question, but maybe it will help in some way.

Rob
Edit:
I will run your code in my machine later today and see what I get on my system.

71taa · Posted at 2018-5-8 07:40:22

Last edited by 71taa In 2018-5-8 07:59 Editor

I've runned the attached code on my test rig (ddcsv2.1 + closed loop stepper on X). Spent some time to rule out problems with interference (unshielded cables laying everywhere) and received similar picture. Not full stops though, but clear reduction of the speed (sometimes till the zero), esp. when it comes to the middle of the programm.
As PPRINT earlier, I've suspected problems with micromoves and arc interpolation alghoritms of the controller. So I've played with settings, and for 8000 subdivision of steps on driver, 1280 pulses per mm and 300mm/s2 in controller I've received stop-free execution, but still with clearly seen heavy decelerations while in problem spots. 100mm/s2 had much more stops/decelerations, for example.BTW while I had problems with interferences I had a lot more shakes and stops on the stepper, probably controller is sensitive to them.

Androcles · Posted at 2018-5-8 21:10:31

Hey
Thanks for all the feedback and the time you guys spent testing it!

So i tried a lot of different things.

Rotated the workplane so that the long side was along the X-axis (Y-axis before). And oddly enough the jerky motion now came at det endpoints instead of in the middle like in the video...

Seems very odd, so I adjusted the closed loop stepper drives and even did a factory reset, still the same.
But I really dont think that something is wrong with the motor or drives but I could be wrong..

I then saw the Feedrate display on the DDCS, and it made a noticeable slowdown when reaching the problemarea.

With low acceleration
https://streamable.com/75l0t

With high acceleration
https://streamable.com/3ntr1

It is almost as if there is som slow feedrate command in the GCode.

I also tried flashing an older firmware, did nothing.

What firmware are you guys running?

Thanks in advance again.

enjoy_cnc · Posted at 2018-5-8 22:07:44

Androcles Posted at 2018-5-8 21:10
Hey
Thanks for all the feedback and the time you guys spent testing it!

Hello Androcles

Can you upload video processing files and screenshots of all parameters?

Thank you!

71taa · Posted at 2018-5-8 23:39:37

Last edited by 71taa In 2018-5-9 00:22 Editor

enjoy_cnc Posted at 2018-5-8 22:07
Hello Androcles

In case if it will help - added a video of my test setup as well: https://youtu.be/9QCczdX6cAs
Timing (video):
3.51-3.52
4.01
5.34
Hiccups before/aftter the vertical center of the scallop in the part there is no significant changes of X coordinate. That's the most strange thing here. As there is no mechanics - the tonality of the stepper sound could be used to locate those spots easily.

8.00
8.08
8.23
... and so on (the closer to the end of the programm - the more prominent slowdowns). Here we have clear corellation with the moment then the X axis steps betwen minimum negative coordinates and minimum posotive coordinates. Probably it's expected, but maybe not:

X-0.117 Y-20.354
X-0.003 Y-20.86
X0.002
X0.171 Y-20.101
X0.332 Y-19.343

And later it starts to slow down in different parts of the programm.

The settings are:LCDA357H+LC57H3100 on the X axis only.
24V/5A dedicated for the controller, 36V/12A for the stepper driver. No limit switches/probe/etc
Switching off the encoder on the stepper doesn't change the picture (checked as well)/
The driver is 5000 pulses/rev. Tried different subdivisions on driver/pulses per mm in the controller - slowdowns are present, but change in scale/time. Tried 91 and 93 builds, tried enabling of the soft arc - no difference

Androcles · Posted at 2018-5-9 02:29:02

71taa

Thanks a lot for your detailed post. It seems identical to what I have tried also our settings a very alike.

I tried installing the newest firmware again, and at first it seemed like it worked a lot better, but the after short time the decelerations came back...

It really seems like it can't handle the code and being able to time the pulses correctly, but it should be more than capable of that with my settings (currently turned the pulses/mm down to 160 as I have been using that always with mach 3 and 4.

Tried bumping up the feedrate, and this is how it behaved
https://streamable.com/n0v3n
A lot more irratic motion

Problem with jerky cutting

TEL

0755-87654321