Dynomotion

Group: DynoMotion Message: 12619 From: yanvrno@frontier.com Date: 12/27/2015
Subject: feed rate steadly reducing
I've been running some longer programs and notice the feed rates are gradually reducing, I have to restart the code to get it to reset. I've looked for any tell tale code that might be over riding the feeds but find nothing obvious.

Any idea where I might look, or if there is a way to get the feed override not to reduce on its own?

Thanks
Mac
Group: DynoMotion Message: 12620 From: Lindsay Whittle Date: 12/27/2015
Subject: Re: feed rate steadly reducing
I had this problem when I was running a machine that had very slow accelerations with a very high maximum velocity.

I believe Tom said that it's caused by the trajectory planner lookahead not being long enough to view an axis accelerating to max velocity (this is from memory, could be very wrong).

I fixed it by increasing my lookahead and acceleration.

Sent from my iPhone

On 28 Dec 2015, at 1:24 PM, fileyanvrno@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote
 

I've been running some longer programs and notice the feed rates are gradually reducing, I have to restart the code to get it to reset. I've looked for any tell tale code that might be over riding the feeds but find nothing obvious.

Any idea where I might look, or if there is a way to get the feed override not to reduce on its own?

Thanks
Mac
Group: DynoMotion Message: 12621 From: yanvrno@frontier.com Date: 12/28/2015
Subject: Re: feed rate steadly reducing
Thanks I'll check that today. I did increase the axis max velocity because the parts are very long and and I found I needed to speed my rapid (G0) up due to the next line (Z) moving before the (X) finished. (and i'm sure I never changed look ahead that was set to 3 seconds.)

Just so we are all on the same page I am using Dynomotion
And to verify, you are speaking of all the vel & acel settings found in the Tool Setup / Trajectory Planner Correct?

Thanks again
Mac
Group: DynoMotion Message: 12623 From: yanvrno@frontier.com Date: 12/28/2015
Subject: Re: feed rate steadly reducing
The best I can tell is it has something to do with the rotary axis radius settings

I am cutting a series 20 degree offset "V's" with 360 degree spiral on a tapered cone. I am seeing the larger the diameter the slower the feed. But for the longest time the feed would not reset for the next pass. Now it seems to be much better but the position of the Z axis (zero is top of part) affects the initial start feed. Rotary Radius values set to .00001 (Setting A radius to 0 slows the positioning to a crawl.)

Witch all makes a little sense, but now I'm a little confused by the settings needed to keep a constant surface footage based on the T-Planners settings.

And If I try to use rapid G0 back to X0 I get a following error quickly, (even though the error is set very high 9e+033) so I ended up using a very high feed rate, seems to be controllable and repeatable. Just not at all understandable.

Thanks
Mac
Group: DynoMotion Message: 12624 From: TK Date: 12/28/2015
Subject: Re: feed rate steadly reducing
Hi Mac,

The issue with Jerk/Acceleration/Velocity settings slowing motion has to do with settings in KFLOP not the Trajectory Planner.   With crazy settings KFLOP can get the idea that it can take a long time to stop (ie 100 seconds).  So it decides it should slow down if there is only a few seconds of motion in its buffer.  But since it was told it might take a long time to stop it must stop very gradually.  What are all your settings?  How long would it take to apply the accelerations based on your Jerk settings?  How long would it take to stop from your max Velocities using your max Accelerations?

I don't really understand your system or your G Code.  Maybe you could explain it more clearly and if possible include a small fragment of GCode that shows the unexpected feedrates.

Regards
TK

On 12/28/2015 11:24 AM, yanvrno@... [DynoMotion] wrote:
 

The best I can tell is it has something to do with the rotary axis radius settings

I am cutting a series 20 degree offset "V's" with 360 degree spiral on a tapered cone. I am seeing the larger the diameter the slower the feed. But for the longest time the feed would not reset for the next pass. Now it seems to be much better but the position of the Z axis (zero is top of part) affects the initial start feed. Rotary Radius values set to .00001 (Setting A radius to 0 slows the positioning to a crawl.)

Witch all makes a little sense, but now I'm a little confused by the settings needed to keep a constant surface footage based on the T-Planners settings.

And If I try to use rapid G0 back to X0 I get a following error quickly, (even though the error is set very high 9e+033) so I ended up using a very high feed rate, seems to be controllable and repeatable. Just not at all understandable.

Thanks
Mac
Group: DynoMotion Message: 12626 From: yanvrno@frontier.com Date: 12/28/2015
Subject: Re: feed rate steadly reducing
Attachments :
right now its cutting and doing well, but when it gets to a point i can copy the info I will include it,

Thanks Tom
   @@attachment@@
Group: DynoMotion Message: 12630 From: yanvrno@frontier.com Date: 12/28/2015
Subject: Re: feed rate steadly reducing
Tom
just pulled these off the machine and back to finish cuts
Right now it just changes feed rate from 9.5 at small end to 6 in the large end 17 inch X linear length and 2  inch Z diameter change

PS as a reminder: This is a 6 axis, msdos.BAT file run swapable machine lathe or mill has a separate but similar Data file for the other setup. Lot of additional scripting for spindles and homing.

Mac
   @@attachment@@
Group: DynoMotion Message: 12631 From: TK Date: 12/28/2015
Subject: Re: feed rate steadly reducing [2 Attachments]
Hi Mac,

Yes it seems you have crazy Velocity, Acceleration, and Jerk Settings.  For Example:

    ch4->Vel=1.1e+010;
    ch4->Accel=1e+007;
    ch4->Jerk=1.5e+012;

I don't think your system can move at a velocity of 11 Billion counts/sec!

KFLOP's encoders can only count at 1 million counts/sec.

And at the specified Acceleration it would take 1.1e10/1e7 =  1100 seconds to stop from that speed.  Most systems should be able to stop in less than 1 second.

This is sort of like thinking you have a car that can go 1 million MPH but you have only tested the car on a road 50 ft long.  You don't notice an issue until you use the max speed for a safe following distance while driving.  Or find a really long stretch of road.

This is probably why you get following errors with G0 moves in some cases.

You might use the Step Response Screen to that all the Velocities and Accelerations of each axis is achievable.   If you don't understand what Jerk is you can just set it to 1000X the acceleration number and the result should be reasonable.

You might read:

http://www.dynomotion.com/wiki/index.php?title=Main_Page#Velocity.2C_Acceleration.2C_and_Jerk

HTH
Regards
TK






On 12/28/2015 2:34 PM, yanvrno@... [DynoMotion] wrote:
 

Tom
just pulled these off the machine and back to finish cuts
Right now it just changes feed rate from 9.5 at small end to 6 in the large end 17 inch X linear length and 2  inch Z diameter change

PS as a reminder: This is a 6 axis, msdos.BAT file run swapable machine lathe or mill has a separate but similar Data file for the other setup. Lot of additional scripting for spindles and homing.

Mac
Group: DynoMotion Message: 12632 From: yanvrno@frontier.com Date: 12/28/2015
Subject: Re: feed rate steadly reducing [2 Attachments]
odd dont know what happened to that one, but it is my mill head spindle and it is only used in this case as a rotary cutter spindle. So it gets turned off in the spindle call. but it does orient fine.  Thou guess i better check out what happened. Hate computers, might have grabbed the wrong file too. Hell to get old.

How can the spindle drive have an effect on X axis? or are you saying my axis 2 (X) needs a longer move and step test than around 1 inch? I try to keep the profiles tests around 1 to 1.5 seconds should I test closer to 2 to 2.5 seconds?

Thanks for the tips
Mac
Group: DynoMotion Message: 12633 From: TK Date: 12/28/2015
Subject: Re: feed rate steadly reducing
Hi Mac,

Axis 4 is being included in the Coordinate System with:
DefineCoordSystem6(2,0,1,5,6,4);// z,x,y,c,a,b (mill = x, y, z, a, b, c)

Other axes have nearly as crazy settings.

Definitely hell to get old :}

Regards
TK



On 12/28/2015 3:40 PM, yanvrno@... [DynoMotion] wrote:
 

odd dont know what happened to that one, but it is my mill head spindle and it is only used in this case as a rotary cutter spindle. So it gets turned off in the spindle call. but it does orient fine.  Thou guess i better check out what happened. Hate computers, might have grabbed the wrong file too. Hell to get old.

How can the spindle drive have an effect on X axis? or are you saying my axis 2 (X) needs a longer move and step test than around 1 inch? I try to keep the profiles tests around 1 to 1.5 seconds should I test closer to 2 to 2.5 seconds?

Thanks for the tips
Mac
Group: DynoMotion Message: 12634 From: yanvrno@frontier.com Date: 12/28/2015
Subject: Re: feed rate steadly reducing
profiles always looked and felt smooth in both step and move profiles in initial testing/setup.

DefineCoordSystem6(2,0,1,5,6,-1); // axis 4 gets turned off in the spindle script

then gets turned back on when spindle stops with M5

I will re-address all and see if i can find a better profile setup,

Thanks again
Mac
Group: DynoMotion Message: 12639 From: yanvrno@frontier.com Date: 12/29/2015
Subject: Re: feed rate steadly reducing
revisited all my servo settings for a much longer test. Seems to be much better but not quite perfect. As the X has not quite returned before the next gcode movement is started.

Thanks again Tom

here is a link to a quick video of it running.

https://youtu.be/Kj-H9qzd8Y8

Mac
Group: DynoMotion Message: 12642 From: Tom Kerekes Date: 12/29/2015
Subject: Re: feed rate steadly reducing
Hi Mac,

You didn't post your settings.

I'm not sure I understand what you are saying.

But a common problem is to have the axis configured for a speed faster than what the axis is capable of going.  This is like tying to follow someone going 70MPH in a car that can only go 65MPH.  The following error will grow to a large value.  Finally when they stop to make a turn you might cut the corner to catch up to them.

If your following errors are set to small values the problem will be detected immediately with a following error and you will know your settings are wrong.  If you set the following error to a huge value to ignore problems you will get the result described above.

HTH
Regards
TK


On 12/29/2015 12:02 PM, yanvrno@... [DynoMotion] wrote:
 

revisited all my servo settings for a much longer test. Seems to be much better but not quite perfect. As the X has not quite returned before the next gcode movement is started.

Thanks again Tom

here is a link to a quick video of it running.

https://youtu.be/Kj-H9qzd8Y8

Mac