Rough Motion on Circular Toolpaths

[Fordmasterkw]

I'm dealing with same issue/different servo though. I have been trying to correct it but have been unsuccessful so far. Just started all of the sudden.
So I will be following this closely.

What power do the servo drives have? What does the surface look like when you machine a square component? If the surface then looks better, the servo drives are not set correctly or their holding torque is too weak

I'm using IPC5 PS.
So I got a handle on this. Back story was I installed a 4th axis and changed fusion post. That's when problems started. So I put old post back in. Still same issue. I noted that only the y axis would make the glass on the enclosure vibrate when homing. All others glass smooth.
Yesterday I decided to re-tune y axis.
I noticed the vibration was gone when I re homed machine. Not sure what happened but the machine is cutting better than ever. Still waiting for spindle from PRECISION MATTHEWS. The spindle taper was actually ground wrong. So I can't wait to see it cutting after.

[Houseman303]

looks good

[avp]

I believe it is most likely a servo tuning issue. I am using clearpath SDSK servos and was having similar issues in certain situations. Especially at lower feed rates 100 ipm and under. Over the past few months I have tried autotuning with different drive mechanics and got it pretty good but not perfect. I finally just started over with the stock unloaded motor profile and just adjusted the RAS and fine tuning toward the stiff side a bit. Did not do any autotuning. Problem gone. 1000 ipm rapids smooth as well. I always struggled with the idea that I had to use the autotuning and not leave the stock profile. So far so good. I would love to learn how to manually tune the clearpath motors.
Bill

[cnckeith]

Not seeing any tiny .002" moves... Unless I don't understand how these arc moves work.

Keith any ideas on this?

as will almost all issues the symptoms can be caused by a number of different factors, if you have eliminated excessively short vectors and are using full size lines and arcs then the issue is back to motor tuning and machine design.
g code smoothing is not designed to fix bad tuning or bad machine mechanical or be used with full size lines and arcs.
g code smoothing is designed to be used on ratty short vector g code with properly tuned axis motors and a machine operating in the kinematic sweet spot with good vibration reducing design (triangulation, mass etc ) so now your focus should be on those items.

[Salted_Pork]

g code smoothing is not [to] be used with full size lines and arcs.

Thanks for the response Keith! Gear reducers are on their way.

However, as you mentioned, I should not have to use g code smoothing for simple lines and arcs. As in the previous video, for some reason, the system can't handle these simple moves without smoothing turned on.

So while I am going to continue down the mechanical troubleshooting path, there is also something else fishy in the system.

This video is of the machine doing the same 3 inch diameter circle, 4 separate times, at different feedrates. Smoothing is off.
1. Programed: 50ipm // Actual: 50ipm
2. Programed: 100ipm // Actual: 100ipm
3. Programed: 200ipm // Actual: very rough noisy motion (much worse than original problem), 50-150 ipm
4. Programed: 400ipm // Actual: very rough noisy motion (much worse than original problem), 50-150 ipm

When smoothing is turned on, it can run the full speeds.

https://photos.app.goo.gl/5S3rchaDDALgXdwdA

[cnckeith]

please install and configure the latest version of CNC12 v5.1. so we are all testing and working with the same software.
when posting a video , please always include a copy of the actual g code file that is being used in the video.
video looks like classic data starvation.
while gearing down will help , i've run plenty of machines at 1 turn per inch that don't do this.

[Salted_Pork]

Hi Keith,

Okay I just finished the painstaking process of installing the latest version of CNC12.

I ran the G Code program again, and it does the same.

See attached gcode, report, and Video

https://photos.app.goo.gl/tkD7kjjtNf1CQW2F8

[cnckeith]

great.
now i would try a different PC.
run the PC tuner on it and simply copy the CNCM directory from the existing PC to the new one (no need to reinstall v5.1)
and run your test file again.

if problem still exists, they evidence is pointing more and more to motor/drive related issues (not tuned properly?, under sized? etc..)
remind me what make and model you are using again.

[avp]

I ran your G-code on my machine. Without smoothing on my machine got up to about 250 ipm. With smoothing on using a custom profile that I use most of the time I got up to 395 ipm. It is a small circle and my clearpath motors RAS setting of 99ms may affect that as well.
Bill

[Salted_Pork]

I'll see if I can find different PC (I don't have another one), but this one is relatively new and seems to be well equipped to run CNC12. Can you think of any setting that I have wrong in CNC12? Before I go buy a new PC, do you know why an unfit PC would cause this problem?

I have the Leadshine Hybrid Servos:

Screenshot 2024-05-29 at 12.06.18 PM.png (208.89 KiB) Viewed 314 times

Im not ruling out issues with the motor/drives, but I find it unlikely (for the data starving issue) since they don't do it when smoothing is on. They are not undersized/underpowered as far as I can tell. Ive done 800ipm on 3/4 plywood with no issue at all. They do stall out if the acceleration is set too fast, but that seems normal to me.