Re: [CAD_CAM_EDM_DRO] Emc Help! Help! Help!
Posted by
Jon Elson
on 2001-08-05 20:53:29 UTC
Stephen Goldsmith wrote:
a number of people. There is a fix for the stack size setting, which is
defined
in a couple of places in the files .../emc/src/emcmot/emcmot.c and
.../emc/src/emcmot/emcstepmot.c
These are normally 2048 in the original code, and some people have reported
that a number of problems, including the hang on startup are fixed with this
change. After changing these define lines, you have to recompile EMC.
That requires you to have installed the development version rather than the
user version of the BDI, if that is what you started with.
Otherwise, you can also turn down the max velocity. I think the problem
might be if you need to produce more than 1 step pulse per cycle of the
steppermod.o code, it will go haywire. if you turn up the INPUT_SCALE,
that means more pulses/second to go the same speed. I seem to remember
the fix for this problem is to turn down max velocity, try that first before
recompiling.
simulating a
servo system. You can put in a small amount of deadband, which should stop
the jittering. .001 or so should work. If the motors really are spinning
away,
then there must be something wrong with the parallel port connection, as
freqmod is usually pretty stable in this regard.
interval is too short, then the CPU can't complete the required computations
before the next interrupt occurs. KaBOOM! Once you have found the
minimum safe value, just leave it there. 16 microseconds is not a lot of
time for a fair bit of calculation.
Jon
> Firstly Thanks to everyone who has help thus far.....I don't understand the details, but this behavior has been described by
>
> When I enter a high number for INPUT_SCALE say 10000 the machine locks.
> It seems to work better with a low number say 800!
> But I need 4 turns of motor to 1mm which = a value of 1016 for
> INPUT_SCALE ( I think)
a number of people. There is a fix for the stack size setting, which is
defined
in a couple of places in the files .../emc/src/emcmot/emcmot.c and
.../emc/src/emcmot/emcstepmot.c
These are normally 2048 in the original code, and some people have reported
that a number of problems, including the hang on startup are fixed with this
change. After changing these define lines, you have to recompile EMC.
That requires you to have installed the development version rather than the
user version of the BDI, if that is what you started with.
Otherwise, you can also turn down the max velocity. I think the problem
might be if you need to produce more than 1 step pulse per cycle of the
steppermod.o code, it will go haywire. if you turn up the INPUT_SCALE,
that means more pulses/second to go the same speed. I seem to remember
the fix for this problem is to turn down max velocity, try that first before
recompiling.
>It keeps spinning, or just jitters? the jittering is normal, as it is
> So I tried the freqmod this is running on a pentium pro 200 which has a
> advanced method for calculations I am led to believe.
>
> The problem now is that the position number in the gui fluctuates up and
>
> down and the motor either caries on spinning or sometimes makes strange
> noises.
simulating a
servo system. You can put in a small amount of deadband, which should stop
the jittering. .001 or so should work. If the motors really are spinning
away,
then there must be something wrong with the parallel port connection, as
freqmod is usually pretty stable in this regard.
>This sets the time between interrupts that wake up the stepper task. If the
> Could you give me any pointers with regards to freqmod and also could
> you let me know of a fixed steppermod.o I am currently using the one
> which I installed from BDI which I downloaded two nights ago not sure
> what version.
>
> Oh I also enabled PERIOD (emc.ini) to see if this would sort the freqmod
> problem.
> It did not but it did hang my machine if the number was to small ie
> .000016 etc...
interval is too short, then the CPU can't complete the required computations
before the next interrupt occurs. KaBOOM! Once you have found the
minimum safe value, just leave it there. 16 microseconds is not a lot of
time for a fair bit of calculation.
Jon
Discussion Thread
Stephen Goldsmith
2001-08-05 16:03:26 UTC
Emc Help! Help! Help!
Paul
2001-08-05 16:49:51 UTC
Re: [CAD_CAM_EDM_DRO] Emc Help! Help! Help!
Jon Elson
2001-08-05 20:53:29 UTC
Re: [CAD_CAM_EDM_DRO] Emc Help! Help! Help!