Advanced Search

I was thinking that but If I watch X and Y or j0 and J1 Position and Position feedback pins in Hal with the following error cranked up I can see it actually home and the values stop but the physical motors just keep rolling coal so to say. It's like Linux cnc think everything is fine but the stepper modules never got the message to stop. Even looking at the Ethercat pins, lcec.0.D2.srv-cmd pin stops when home is set but motors still spinning. Also swapping Kins back out to trivkins shows the same behavior during homing. That's why I was wondering if it's something to do with the ethercat setup perhaps?

Do you get the same if you use a program rather then MDI?

So as far as I know the kinematics need to be created so I have been taking a stab at that and have something to test.

Are you sure it's not your kinematic that creates the runaway?

Hey all,

I'm working on a 5 bar parallel robot setup.  Something similar to this.   5Link Robot

I cant seem to find if that person had a repo or config they shared. So as far as I know the kinematics need to be created so I have been taking a stab at that and have something to test.  At the same time the hardware I'm testing with are the Beckhoff 7041 stepper cards.  I have two nema 17 motors connected to the 7041's and have LCNC connecting to them with no issue.   The problem I'm having is homing.   I don't have any home switches setup right now.  Because of the kinematics I cant home at 0,0  I need to home in a position that fits in allowable space.  I have the joints setup So they home immediately. Hove Search and Latch Velocities = 0 along with Home = 0 I did set the Home offset to the positions I want the joints to be considered at.  When i fire up axis and home the joints Joint zero starts running away limited to the value I set in Ferror. I'm confused if i have a hal problem with the Ethercat setup or if something else is at play as my understanding is that the joint should not cause motion and be set as home with an offset of 60.  What May I be missing?   

Not quite sure I understand. the position feedback should work whenever LinuxCNC is running
so is seems you should just set the position to retract the head at the end of the Gcode.
In addition the initial Z position should be in the retracted position.


 

This probably doesn't belong to this specific place on the internet, but I'm posting it here anyway, 'cause it's in my head and I want to get it out:

As I have said many times before to anyone willing to listen, I fail to fully grasp how forums work. In my head, the right way to present a question, problem or dilemma is to give as much backstory and context to the problem as I can, so that the people that would know the answer would also know why I am asking and can see my full logic behind the problem. Ultimately, it's because I want to learn something, more than just solving my immediate problem, hoping the answers will have as much context as my question.

But that's not the way forums work. There are basically two outcomes:

1) Someone replies to the question they think I asked, in their own way. I usually have to clarify that the response doesn't apply to my question. Then they get offended and the rest of the thread is an argument about my attitude instead of an answer to my question. Onlookers pile on for predictable reasons, I disengage, and then a year later someone answers my question.

2) Someone replies to a small part of the subtext of the question, and not the question itself. Ten pages later, I still don't have an answer. Turns out, all the experts in the forum never thought of that. Over the next three years, people DM me instead because they didn't want to be on the wrong side in public.

I understand that I cannot demand better answers; I have no control over how people respond. I have control over:

1) How I ask questions;
2) Where I ask questions.

I need to get better at both.

How to configure the 7i77 electronic board to always start with a positive output of 10 volts before home referencing?

Good afternoon everyone, please, I need help with the following situation:

Just to give an example of a stepper motor drive device, when we connect the machine is enabled or amplifier enable input, it physically stops the motor with power to the coils and only moves when there is a Step and Dir pulse, that is, the axis does not move while the machine is on.
Unfortunately, what I am controlling is not a simple stepper motor, and neither is a Step and Dir pulse used. In my case, I am analogically controlling a hydraulic cylinder in the z-axis movement with the 7i77 electronic board.

This hydraulic cylinder, when the machine's hydraulic pump is turned on, slowly descends on its own, due to the pressure and the force of gravity itself. To keep it high, it is necessary to press jog + on the Z axis, which injects positive voltage and causes the hydraulic cylinder to rise, but as the analog output returns to zero, the cylinder descends gradually, unlike a stepper motor that remains locked until further notice. In other words, you have to keep pressing the jog + button on the z axis to keep it high.

below illustration:

1 - initial situation before axis home

 

2- jog +Z situation

 

3- jog situation -Z

 

below video demonstration



This situation in the video may show that when moving the tool wheel, it will collide with the Z axis.

 

To avoid such a collision, I need the Z axis in the case of the 7i77 board output to always have a positive output of +10 volts, this before the home position and before running the g-code program, but when running the g-code program, it will have to obey the commands, when stopping the g-code, it will return to +10 volts. I accept any help or suggestion, I thank everyone who can contribute.

How to configure the 7i77 electronic board to always start with a positive output of 10 volts before home referencing?
 

tommylight wrote:

Are you trying to run LinuxCNC in sim mode inside a VM?
 

No.

Now, what was your question?

I need a Parallels LinuxCNC VM created in Parallels on an Intel Mac.

JT wrote:

Your screen shot looks correct. I added multi state labels on the 20th, how are you updating?
JT
 

I'm downloading the deb release from the git page. It's still at 1.1.1 but is says it's been updated since then.

Looks like apt didn't update it right this time since it's still version 1.1.1, I used gdebi and now it works. I did get a wierd error about a different label already having that hal pin name but that wasn't the case. I'm not sure if that stems from the multi-state label having the same widget name as it's hal pin name or the hal pin name having underscores. I had other regular labels with a single hal pin with underscore in the name and it was fine. I just changed the underscores in the hal pin name to dashes and the error went away.

See network can't connect to to them. What am I missing? 

Are you trying to run LinuxCNC in sim mode inside a VM?
Then it all pertains to it, several options:
Using the official ISO, should work as any other distro in VM, but might have issues with RT kernel and VM not getting along nicely
Using Debian or LMDE6, if those work in VM, then adding LinuxCNC is easy, and should work as you can use them without booting into the RT kernel.
Now, what was your question?

JT wrote:

Take a look at this thread for the coordination of an external E-Stop with LinuxCNC
forum.linuxcnc.org/47-hal-examples/25861-external-e-stop
 

Nah estop-loop doesn't work right in practice without extra logic components either. The way it's shown in the other thread it does block the UI Estop, but the reset won't trigger until you toggle the UI button afterwards. I just didn't notice I still had iocontrol.0.emc-enable-in connected to iocontrol.0.emc-enable-in. I just needed to break that with a NC circuit. Maybe one day somebody will rework the mess that exists between halui and iocontrol estop pins. It would probably be as simple as adding 2 pins to halui to break the iocontrol loop when is-activated=true. If that loop is unused it would behave as it does now, like a toggle off a momentary switch.....That's besides the point, I figured it out.

I have not tested an external E-Stop so the colors may not stay in sync.

If they followed the same logic as the actual E-stop and Power states like the button labels do they'd be fine. You can test it easily by setting the halui.estop activate, reset and halui.machine.on pins

Have a look at mdi_command as in this example (instead of 'net coolant-flood => halui.mdi-command-00' you would connect it to the button hal pin):
forum.linuxcnc.org/qtvcp/54583-call-cust...om-gui-button#315402

Ok, thank you PCW, I will do that and try it out these days. I'm just getting a cold and my garage has 9°C, thats not very comfortable.
I will report the results asap!

Kind regards

I would make some changes in the hal file.

Change:

setp hm2_7i76e.0.dpll.01.timer-us -50

to:

setp hm2_7i76e.0.dpll.01.timer-us -200

and change all joint PID "P" terms to 666