Dynomotion

Group: DynoMotion Message: 11203 From: Wcarrothers Yahoo Date: 3/12/2015
Subject: Follow error
Hello

On occasion my analog drives controlled by my kflop kanalog setup go off in a direction and get stopped by the follow error shitting down the system. With this machine I do thousands of small parts with small moves and does not happen very often

Not sure if the occasional move is noise or drive issue or something else. I do not think this is the k's fault.

But think the k could trap for it a little quicker

For these parts I have a rapid move to location then plunge followed by small slow moves and retract and rapid to next location

The rapids course limit how low I can drop my follow error number

So is there any way perhaps have a lower follow implemented while it is working g1 and g3 moves vs g00?

Thanks

B
Group: DynoMotion Message: 11204 From: Tom Kerekes Date: 3/12/2015
Subject: Re: Follow error
Hi Bill,

There aren't separate following error limits for Rapids vs Feeds.  The only way I can think of to do this is with a User Program to monitor CS0_DoingRapid (added in the latest Test Program to facilitate Rapid FRO) and switch max following error settings.  Something like:

#include "KMotionDef.h"

main()
{
    for(;;)
    {
        WaitNextTimeSlice();
       
        // Rapid in progress or not running a Job (ie Jogging)??
        if (CS0_DoingRapid || !JOB_ACTIVE) 
        {
            ch0->MaxFollowingError=500;
            ch1->MaxFollowingError=500;
            ch2->MaxFollowingError=500;
        }
        else
        {
            ch0->MaxFollowingError=50;
            ch1->MaxFollowingError=50;
            ch2->MaxFollowingError=50;
        }
    }
}


How big are your typical following errors for Rapid moves?  With a properly tuned system they shouldn't be that big.

Regards
TK

Group: DynoMotion Message: 11205 From: Wcarrothers Yahoo Date: 3/12/2015
Subject: Re: Follow error
Hahaha.  Ya actually I have run the follow errors down a little.  But you know I'm the guy running my system by the seat of my pants for years.  Actually I'm on v429. Don't like to change much on this system seeing it runs mostly the same family of parts all the time and does it really well as it sits.  I still have not tracked down why it runs off in a direction on occasion.  We are taking it makes it maybe 1/4" before the system shuts down. My follow errors are at 180 counts.  Long ago they were 500 and then dropped to 250

System for these axis are 9126 counts per " and 8691 counts per ".  
So Technicly that is a small distance of follow error 
So when it does it I'm not convinced actually the drive reports the move fast enough to catch it till after the fact.  Perhaps why these drives were removed from the system at Wilson way back when seeing I would bet something like this would result in the golf ball loader robot to send balls all over.  Ha

Anyway. All that said my follow error is pretty low as is so while this command is cool looking perhaps would not trap for it seeing I think the fault is in the drive

Some day I'll get machine 3 running parts (k flop snap amp) and redo machine 2 with new drives and servos 

B 

On Mar 12, 2015, at 1:06 PM, "Tom Kerekes tk@... [DynoMotion]" <DynoMotion@yahoogroups.com> wrote:

 

Hi Bill,

There aren't separate following error limits for Rapids vs Feeds.  The only way I can think of to do this is with a User Program to monitor CS0_DoingRapid (added in the latest Test Program to facilitate Rapid FRO) and switch max following error settings.  Something like:

#include "KMotionDef.h"

main()
{
    for(;;)
    {
        WaitNextTimeSlice();
       
        // Rapid in progress or not running a Job (ie Jogging)??
        if (CS0_DoingRapid || !JOB_ACTIVE) 
        {
            ch0->MaxFollowingError=500;
            ch1->MaxFollowingError=500;
            ch2->MaxFollowingError=500;
        }
        else
        {
            ch0->MaxFollowingError=50;
            ch1->MaxFollowingError=50;
            ch2->MaxFollowingError=50;
        }
    }
}


How big are your typical following errors for Rapid moves?  With a properly tuned system they shouldn't be that big.

Regards
TK