Dynomotion

Group: DynoMotion Message: 14557 From: cnc_machines Date: 4/5/2017
Subject: Kmotion Buffer

Greetings,


I have a small router running a K-Step with encoder feedback. We use it to profile parts with a small ball mill. This is a lot of very small straight line movements, and the programs are quite long some are 500,000+ lines. I am trying to squeeze every second of cycle time out of the process. When profiling parts that have lots of motion in the Z axis, it seems that I never hit the target velocity. My feed rate is 400 IPM, I am lucky to get up to 60 IPM. I thought this was due to low acceleration in KMotionCNC. I Increased it from 100 in/sec^2 to 200, 300, and 400. This doesnt appear to have any effect. The total cycle time remains the same. I thought that maybe I needed to change the look ahead because of all the tiny little moves (buffering issue??)- I moved it to 15 seconds. Still no change in cycle time. 


I designed the machine to be very fast with 1G of accelleration. Could this be a buffer issue? Will the KFlop struggle doing coordinated motion on 3 axis, 400 IPM, and 1G (365 in/sec^2)?


Thanks,


Scott

Group: DynoMotion Message: 14558 From: Tom Kerekes Date: 4/5/2017
Subject: Re: Kmotion Buffer

Hi Scott,

It isn't clear to me what you are doing.  Are you moving along a smooth path consisting of small segments?  Or stopping to probe at every segment?

Can you post a short GCode fragment that demonstrates the issue?

Also please post all your Trajectory Planner parameters.

What Version are you running?

Regards
TK


On 4/5/2017 12:59 PM, cnc_machines@... [DynoMotion] wrote:
 

Greetings,


I have a small router running a K-Step with encoder feedback. We use it to profile parts with a small ball mill. This is a lot of very small straight line movements, and the programs are quite long some are 500,000+ lines. I am trying to squeeze every second of cycle time out of the process. When profiling parts that have lots of motion in the Z axis, it seems that I never hit the target velocity. My feed rate is 400 IPM, I am lucky to get up to 60 IPM. I thought this was due to low acceleration in KMotionCNC. I Increased it from 100 in/sec^2 to 200, 300, and 400. This doesnt appear to have any effect. The total cycle time remains the same. I thought that maybe I needed to change the look ahead because of all the tiny little moves (buffering issue??)- I moved it to 15 seconds. Still no change in cycle time. 


I designed the machine to be very fast with 1G of accelleration. Could this be a buffer issue? Will the KFlop struggle doing coordinated motion on 3 axis, 400 IPM, and 1G (365 in/sec^2)?


Thanks,


Scott


Group: DynoMotion Message: 14559 From: Russ Larson Date: 4/5/2017
Subject: Re: Kmotion Buffer

Scott,

Did you check your gcode?  Were the moves using G0?  There is acceleration parameters in two places one is in your init.c file along with Velocity.  The second area is in the trajectory planner under the Tool tab in KmotionCNC.  If your gcode is specifying a feedrate other than G0 that is as high as it can go.  You said your feedrate was 400 IPM.  But in Kmotioncnc you said you are doing 100 inches/sec that is very fast, meaning 6000 inches per minute.  Please look at the unit in the trajectory planner they are in inchs/second.  I know I made this same mistake and TK pointed out my issue.

 

Russ

 

 

From: DynoMotion@yahoogroups.com [mailto:DynoMotion@yahoogroups.com]
Sent: Wednesday, April 05, 2017 4:00 PM
To: DynoMotion@yahoogroups.com
Subject: [DynoMotion] Kmotion Buffer

 

 

Greetings,

 

I have a small router running a K-Step with encoder feedback. We use it to profile parts with a small ball mill. This is a lot of very small straight line movements, and the programs are quite long some are 500,000+ lines. I am trying to squeeze every second of cycle time out of the process. When profiling parts that have lots of motion in the Z axis, it seems that I never hit the target velocity. My feed rate is 400 IPM, I am lucky to get up to 60 IPM. I thought this was due to low acceleration in KMotionCNC. I Increased it from 100 in/sec^2 to 200, 300, and 400. This doesnt appear to have any effect. The total cycle time remains the same. I thought that maybe I needed to change the look ahead because of all the tiny little moves (buffering issue??)- I moved it to 15 seconds. Still no change in cycle time. 

 

I designed the machine to be very fast with 1G of accelleration. Could this be a buffer issue? Will the KFlop struggle doing coordinated motion on 3 axis, 400 IPM, and 1G (365 in/sec^2)?

 

Thanks,

 

Scott




Avast logo

This email has been checked for viruses by Avast antivirus software.
www.avast.com


Group: DynoMotion Message: 14560 From: cnc_machines Date: 4/5/2017
Subject: Re: Kmotion Buffer
Thanks Russ. I am going moves of G01, KmotionCNC trajectory Planner has the following values:

Vel in/sec - 20
Accel in/sec2 - 365

When I lower the acceleration the total cycle time stays the same.
Group: DynoMotion Message: 14561 From: Dan W Date: 4/5/2017
Subject: Re: Kmotion Buffer
Scott,
I had a similar issue I recently. Tom thought it was just a instantaneous feed vs averaged feed rate display in the gui, when I was doing a helical bore at high speeds. In my case it was a tuning issue. But my setup is a bit different than yours in that I am using AC servo motors in speed mode with a Kanalog, so u am not sure any of this will help. 

The underlying issue in my case was the inertia ratio setting within my servo drives. I had set the drive to think the ballscrew was only 30% of the inertia of the motors rotor. When I calculated the inertia ratio (93%) and set the drive correctly my move speeds picked up and now seldom drop below the commanded speed. What was happening is the Kflop was slowing down the system because I was hitting my set error limit.

Now I had tuned my machine and my plots were good. I cranked the acceleration way up like I was recommended when tuning. I yhink the problem is that using very high accelerations and some feed forward can be used as a crude way to compensate for inertia mismatches, sometimes. My problem was being covered up by doing long fast movements when tuning. Doing short movements with a lower acceleration brought the problem out for me. 

I am now surfacing at 250ipm on my machine with programs pushing over 500k lines at times.

I tried to change my servos to torque mode recently but had a ringing show up during tuning and without a inertia ratio setting in kmotion and no idea how to fix it at the time, I gave up and went back to speed mode.

Dan


Sent via the Samsung Galaxy S®6 active, an AT&T 4G LTE smartphone