Dynomotion

Group: DynoMotion Message: 11744 From: Hardy Family Date: 6/22/2015
Subject: Problem with g0 a... x... y...
My colleague just noticed a problem, which I can also reproduce:

Using version 433L.

When coding a rapid move with A moving about 45 degrees, and with X and Y moving a small amount, the A axis will attempt to overspeed and lose steps, even though we have the max A velocity at a reasonable value.

This does not seem to occur if we just have one linear axis in the move, or move A by itself then XY on the next block.

As far as I understand, the motion planner should always ensure that the most time-consuming axis will limit the overall move velocity.  It seems to be underestimating the total move time in this case, or maybe ignoring the A axis max velocity in its calculation.

Regards,
SJH
Group: DynoMotion Message: 11746 From: Tom Kerekes Date: 6/22/2015
Subject: Re: Problem with g0 a... x... y...
Hi SJH,

It isn't clear if you are saying the problem onlu occurs when all three X,Y,A move together?  Are you doing Inverse Time or normal Specified Feed Rate?

Could you send me all your Trajectory Planner Settings and a Gcode Example?

Thanks
TK

Group: DynoMotion Message: 11747 From: Hardy Family Date: 6/22/2015
Subject: Re: Problem with g0 a... x... y...

Hi Tom,

It's a g0 so feed rate doesn't apply. Yes, x,y,a must move together to show the problem. Ill send the setup files tomorrow. (Not at my pc right now).
Regards,
Sjh

On Jun 22, 2015 9:10 PM, "Tom Kerekes tk@... [DynoMotion]" <DynoMotion@yahoogroups.com> wrote:
 

Hi SJH,

It isn't clear if you are saying the problem onlu occurs when all three X,Y,A move together?  Are you doing Inverse Time or normal Specified Feed Rate?

Could you send me all your Trajectory Planner Settings and a Gcode Example?

Thanks
TK

Group: DynoMotion Message: 11748 From: Tom Kerekes Date: 6/22/2015
Subject: Re: Problem with g0 a... x... y...
Hi SJH,

For Rapids I will need your KFLOP V, A, J settings also please.

Regards
TK

From: "Hardy Family hardy.woodland.cypress@... [DynoMotion]" <DynoMotion@yahoogroups.com>
To: DynoMotion@yahoogroups.com
Sent: Monday, June 22, 2015 9:18 PM
Subject: Re: [DynoMotion] Problem with g0 a... x... y...

 
Hi Tom,
It's a g0 so feed rate doesn't apply. Yes, x,y,a must move together to show the problem. Ill send the setup files tomorrow. (Not at my pc right now).
Regards,
Sjh


On Jun 22, 2015 9:10 PM, "Tom Kerekes tk@... [DynoMotion]" <DynoMotion@yahoogroups.com> wrote:
 
Hi SJH,

It isn't clear if you are saying the problem onlu occurs when all three X,Y,A move together?  Are you doing Inverse Time or normal Specified Feed Rate?

Could you send me all your Trajectory Planner Settings and a Gcode Example?

Thanks
TK

Group: DynoMotion Message: 11757 From: Hardy Family Date: 6/23/2015
Subject: Re: Problem with g0 a... x... y...
Hi Tom,

Here is sample g-code:

g21
g0 a45 x1 y1
a0 x0 y0
m2


The KMotionCNC settings file is attached, plus c code for setting up kflop axis parameters.

Regards,
SJH


On Mon, Jun 22, 2015 at 9:35 PM, Tom Kerekes tk@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:
 

Hi SJH,

For Rapids I will need your KFLOP V, A, J settings also please.

Regards
TK

Group: DynoMotion Message: 11759 From: Tom Kerekes Date: 6/23/2015
Subject: Re: Problem with g0 a... x... y... [2 Attachments]
Hi SJH,

I was able to duplicate it.  We found two bugs.  Hard to explain but basically where angles are treated as distances at the specified radius the conversion was being performed twice.

Here is a patched GCodeInterpreter.dll for Version 4.33l that should fix it.


Changed sources are in the directory if you need them.

BTW I noticed your configuration specified radii of 1 inch.  Is this correct?  For specified Feed Rates the A Axis will be factored in as generating feed at the specified radius.  To have the A Axis operate as pure degrees you must set the radius to zero.

Please let us know if this solves the problem.

Regards
TK


Group: DynoMotion Message: 11767 From: Hardy Family Date: 6/24/2015
Subject: Re: Problem with g0 a... x... y...
Hi Tom,

Thanks very much for working so late to fix that.  It's working for me (I had to rebuild from source since we are using VS2013), so I'm handing it over to my colleague for further testing.

Subsequent to the initial post, we noticed that moves with two rotary axes (A,B) would also occasionally lose steps, although the effect was subtle and only noticeable because we are trying to get our rotary table alignment procedure as accurate as possible.  Separating axis moves into separate g0 blocks would work around the issue, so it did seem like a related coordinated motion problem.

Regards,
SJH


On Tue, Jun 23, 2015 at 11:25 PM, Tom Kerekes tk@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:
 

Hi SJH,

I was able to duplicate it.  We found two bugs.  Hard to explain but basically where angles are treated as distances at the specified radius the conversion was being performed twice.

Here is a patched GCodeInterpreter.dll for Version 4.33l that should fix it.


Changed sources are in the directory if you need them.

BTW I noticed your configuration specified radii of 1 inch.  Is this correct?  For specified Feed Rates the A Axis will be factored in as generating feed at the specified radius.  To have the A Axis operate as pure degrees you must set the radius to zero.

Please let us know if this solves the problem.

Regards
TK