Advanced Search

I got this error. I reduced the acceleration slightly, and after that I produced 20 parts without any errors. Today the same error appeared again. I then restarted the same program, and so far there have been no errors.

 

 

Certainly looks like your home pose is off.

1. The lower arm is not vertical so your homing parameters for joint2 must be off.

2. Joint 4 is off as well it has to be either straight down or horizontal 

3. Also joint0 is not oriented correctly, the upper arm is not aligned with the base (Although this is less problematic than the misalignment of joint2 and joint4 )

4. As you noticed the pose in the first drawing and the most recent ones are different. You can choose either one but you will need to adjust your 'side offset' accordingly. In your picture the wrist is obviously offset in the positive y-direction but the coordinate values show a _negative_ Y value, that needs to be fixed in the DH parameters:

 

So, in order to go forward you need to:
1. choose which home pose you want to use and derive the DH parameters accordingly
2. set up your homing parameters so the arm joints are actually in the chosen home pose when all joints are at 0°
 

Holzwurm56 wrote:

Ich will meine Drehmaschine auf eine modernere Steuerung umrüsten. Jetzt läuft sie unter Mach3 und 2 Printerports. Ich habe aus alten
Zeiten noch eine Mesa 7I96 und habe mir aus dem Forum von Tommylight LM22.1 mit LCNC2.10 runtergeladen. Wichtig sind für mich 2 Handräder damit ich auch quasi-konventionell drehen kann. Hat jemand so etwas schonmal gebaut und kann mir ein paar Tipps geben? Oder gibt
es unter LCNC eine bessere Lösung? Die 7I92 scheint ja eben sehr aktuell und beliebt zu sein?
Ich habe die Frage im englischsprachigem Forum gestellt, weil mir das mit den Google-Übersetzer
doch etwas umständlich ist, stelle ich die Frage auch nochmal hier.
MfG
Hans
 

also gemacht hab ich das schonmal, zwar nicht mit mesa aber funktionieren tut es auch mit mesa und mehreren handrädern:



musst nur schauen das du 2 encoder eingänge auf deiner mesa karte aktivieren kannst, da kann dir im nofall PCW helfen.

 

I did finally get my spindle to work. I removed everthing from the spindle control cabinet and replaced them with a KBSI-240D signal isolator paired w/ a KBMM-125 DC drive. Relatively simple wiring, I can put together a simple diagram if needed. Now I'm working on trying to get RPM feedback from the existing tach.

Yes, this is very interesting.

and after reading

forum.linuxcnc.org/ethercat/58227-how-go...ion-control?start=20

From: NWE

this Make me feel that 1,000,000ns thread is slow for ethercat
But as the same time I start to fill myself in the place of look ahead reading paradigm where I can make interpolated circles as fast as my control system can respond on time without lose the path

Thinking in real world G0 at 200IPM and I never machine a part beyond 100 IPM, and high precision, this is not my goal, I guess I will need a controlled temp room for this propose and this is not my field. +-0.003inch it is fine.

My question is Ethercat is enough at 1khz thread to keep the path? ,or should I look for a 500,000 servo thread. Or less?

I have a system full LCNC and Ethercat control, my concern is, how to reach a 200IPM max speed and 0.0001 min resolution in a repeatable and stable motion system, I think now about what is on my control and what is not, what variables I can candle and what I cannot handle, and what is on my budget and what is not.
And as well I fill lack of experience and acknowledge on a perhaps non mature control method. Compared with analog or pulse and direction methods

compared with pulse and direction method mesa card at 100 or 200 kHz is enough for a 200 or 300 ipm at 0.0001inch min resolution. Where even with my parallel port 10kHz I was not far from my needs. And under this method I feel that I can handle all variables that the system have, and I can keep them under control.

But in the other hand I feel ethercat is the future, clean wiring, a lot information from servos and VFDs, closed loop etc. and every day you can find more and more equipment that offering this control method.

Someone has been working with smaller servo thread than 1,000,000ns.
someone on a multi thread CPU has been isolated the ethercat servo thread In one CPU thread, as the same is done with LCNC, Is that possible? Can be helpful?

Please share your experiences.

I use UHU DC servo controllers. They compare the actual position with the target position. If the deviation is too large—for example, if I crank faster than the axis moves—the controller goes into a fault state and stop the axismoving.

tuxcnc wrote:

There are many reasons why you should not use MPG for turning, but I will give you just one - you can turn the handle faster than the axis is moving, in other words, when you are done turning, the axis will continue to move until it reaches the commanded position.

LinuxCNC MPG inputs can be configured to stop moving as soon as the MPG stops. (Because this is such an obvious problem)

joint.N.jog-vel-mode

at

linuxcnc.org/docs/stable/html/man/man9/motion.9.html

It might be worth looking through the PRs as I think that there is other work ongoing in this area.

github.com/LinuxCNC/linuxcnc/pull/3718

After homing in joint mode (idintity) i add second screenshot from world! but i moved B a little bit to go away from 0.0
in my ini:
 
RS274NGC_STARTUP_CODE = G21 G90 G94
#RS274NGC_STARTUP_CODE = G21 G10L2P0 x19.2 y5.5 z9.4 a0 b90 c0 (debug, ini: startup offsets set)

ALPHA-0 = 0
A-0 = 0
#D-0 = 330.2
D-0 = 328.93

ALPHA-1 = 1.570796
A-1 = 0
#D-1 = 121.92
D-1 = 0

ALPHA-2 = 0
A-2 = 203.2
#A-2 = 198.12
D-2 = 121.92
#D-2 = 0

ALPHA-3 = 1.570796
A-3 = 0
D-3 = 203.2

ALPHA-4 = -1.570796
A-4 = 0
D-4 = 0

ALPHA-5 = 1.570796
A-5 = 0
D-5 = 55.88


thanks for your help!

Lets verify your home pose:

1. Home the arm
2. Activate 'Identity' mode
3. Remove all offsets (MDI: g10 l2 p0 x0 y0 z0 a0 b0 c0)
4. Move all joints except joint_4 to zero (MDI: g0 x0 y0 z0 a0 b0 c0)
5. Take a screenshot of the axis gui window
6. Take a picture from the side (ie the Y direction) of your robot
7. Post the two images

Hello everyone,

I'm an engineer based in Switzerland, French and German speaking, doing
research before building my first LinuxCNC machine. I have a background
with other CNC and PLC systems (FANUC, Beckhoff, Siemens...), and I'll be using EtherCAT hardware.

I started with a week of plain reading — docs, forum, source. Coming fromother controllers, what I really wanted was a picture of how a G-code lineactually becomes motion. At the same time I was experimenting with AI for
my job, so I combined the two: I used the machine-learning work as a way
into the codebase, and the codebase as a serious test of the tooling.

Two things came out of it.

First, the Code Notes block diagram, redrawn and brought up to date:

   zia-research.github.io/linuxcnc-audit/sh...de-notes-errata.html

   The original diagram as published, side by side with a corrected
   version reflecting the current source.

Second, a generated command-flow sheet:

   zia-research.github.io/linuxcnc-audit/sh...nc-command-flow.html

   An animated sheet following one command all the way through: .ngc file
   -> Interp -> canon -> interp_list -> task -> the emcmot shared segment
   -> motion-command-handler -> the trajectory queue -> tpRunCycle ->
   inverse kinematics -> HAL -> pid/stepgen -> hardware, and the feedback
   path climbing back up. Hovering any block shows its role and the source
   file behind it. It also carries three detail panels: the servo cycle as
   a ring, the seven position representations, and every buffer drawn to
   scale.

I know the reputation AI-generated content has, so the rule throughout was
that nothing is claimed without being verified in the source, every
statement carries a file:line citation, and a script re-checks all 111
citations against the tree so anyone can validate them independently.

That verification work turned up a handful of places where the
documentation and the source have drifted apart over the years — the Code
Notes say so themselves in their opening paragraph. Nothing dramatic, but
a few would bite someone: an example in the HAL manual that names a funct
which doesn't exist, and one behaviour around spindle-synchronized motion
that isn't documented anywhere.

After that I let the AI implement the corrections as well. Small doc
patches are ready if there's interest, but that's secondary — the sheets
are what I mainly wanted to offer.

Everything targets master (2.10.0~pre1); I'm aware some details differ on
2.9.x.

The most important part:

   Could someone with real experience of the codebase take a look and tell
   me whether this makes sense? I'm new here, and a diagram that is
   confidently wrong is worse than no diagram at all.

Sources, findings and patches:
github.com/Zia-research/linuxcnc-audit

Many thanks. I'll give it a try.

ah i get it...i think..

this means... its important that if i jog in +.. the + the direction is of the joint i move... and this is the direction of the arrows +.

this means for the alphas from the drawing: 0,90,90,0,-90,0 would be tool to the table?!

my setup is alpha 0,90,0,90,-90,90

i throwed again my hand with the right hand rule.. thumb up is Z..
i get 0,90,0,90,-90,90 BUT the toolplate is parallel to the base/table. different is that the toolplate is in another angle?!

its so complex...

thank you for your help....

Routerworks wrote:

Any suggestions?
 

camotics.org/gcode.html

Regarding the joint rotation direction: It does not really matter if you follow the drawing or not the important thing is that the actual joint rotation directions on your arm match those in your DH-parameter model. Again, the DH  parameter values describe the sequence of translations and rotations used to transform the world coordinate system in the center of the base to the hand of your robot.
The positive rotation sense of each individual joint motor on your arm is determined by how you choose to manipulate the coordinate system on its path to the hand when you derive the DH parameter values. That is why you need to understand what the DH parameter model actually is. But I repeat myself.

The side offset (ie the one in the y direction) is the distance between the rotational axis of joint0 (likely the center of the vertical column) and the center of the wrist.