Advanced Search

Thanks for your reply . Just curios if we need to loadrt something in hal before putting the new command.

Hi
G10 L20 P1 z0 doesnt work if the below program is stopped midway and rerun. Random z values come up and cut gets all wrong. Please guide as what is incorrect.
G10 L20 P1 x0 z0
G01 Z0.5
G01 Z-0.1 F5 ; cut here
G10 L20 P1  z0
Code here
M2

This very quickly is going to turn into a how to destroy some Fanuc encoders thread so stay tuned for that   

I've decided to replace the encoders... this is probably not the answer that future readers are going to want to read, but its the least expensive and most straight-forward approach I can come up with. Either I have a digital problem or a mechanical problem to solve, and the latter is much more manageable with my skill set. 

I'm trying to get better at knowing when to take a different path to the finish line. I once spent months trying to design a DC motor driver (90v 200amps) that ran on 600v 3phase for a Hardinge CHNC lathe based on open source electric vehicle schematics. Long story short, I didn't succeed and blew up a lot of expensive components along the way    Once I finally decided to swap the motor for an AC one and add a VFD, the whole thing was turning in in 2 days! 

More to come! 

timo wrote:

Unfortunately hollow shaft encoders are much more expensive than the required parts

 

One option might be an engine crank encoder:

Do not buy RPI5
www.youtube.com/shorts/qN-1nPGbS3I

If you end up doing ~91 full turns (@65536 counts per degree and 360 degrees per turn)
the 32 bit signed position number used in intermediate encoder calculations will wrap, I suspect
this may be whats happening. Are 91 turns possible?

running version 2.9.2

reason for a high count on C is because its tool path with just continues +C axis moves in it to form a swarf move in a hole,

in the ini for the c axis i have
MIN_LIMIT = -1e99
MAX_LIMIT = 1e99

ENCODER_SCALE = 65536
#1/5 first gear , 1/72 worm 2nd gear
#16384*4= 65536 pp/r
#gears 1/5+1/72=1/360
#65536*360/360=65536

One option would be a stmbl v5: github.com/freakontrol/stmbl
There is also a thread on the forum.
I have it running with smaller Mistubishi Motors, your encoders should be supported. The HA103 Motor is on the upper power limit, but it should be fine. It can run the 200V motors direct from mains voltage without a transformer.

i use M68 to control my rc-servo gripper's
so i can control it by gcode and for the gui i use mdi commands and a vcp gui with buttons to set some fixed positions

[HALUI]
MDI_COMMAND = M68 E0 Q-100
MDI_COMMAND = M68 E0 Q-75
MDI_COMMAND = M68 E0 Q0
MDI_COMMAND = M68 E0 Q75
MDI_COMMAND = M68 E0 Q100

www.youtube.com/shorts/Tp3savo-MGs

Very well prepared

Please a few questions.
I have a Scorbot-ER-3 (with original control box) working using LinuxCNC, But cannot activate the gripper using G-Code commands or the Logitech dual action F310 controller pendant. When Andy designed this for LinuxCNC in 2020 the only way to open or close the grippers is through a GUI pop-up box and the mouse.
Thank you for your time.

You can make or purchase a quadrature encoder with index that uses three detectors. With quadrature encoding the counter increments each time either the A or B detector transitions so you get 4X the number of teeth or holes. You also get direction and diagnostics if the pulses transition incorrectly. The index signal Z is generated once per rev and can be implemented with a deeper tooth or another hole offset from your main pattern. There is no need for excessive resolution for threading. I have no problems with 65 teeth/260 pulses per rev on my lathe or rigid tapping  using a 100ppr encoder on my mill.

With 55 holes and two detectors (A and B ) you would get 220ppm. To achieve this you'd want to get approx 50% duty on/off from each detector (controlled by size of hole or teeth and how the detector aligns. You may want more or slighty larger holes to get approx 50% duty. You would want the A and B signals offset by approx 25%. You should use halscope to display the signals as you adjust the detectors. The index detector can be adjusted independently as it is normally aligned to a different radius. 

If you have enough holes/teeth (55 should be fine) you can also use "missing tooth" detection to index.

I have implemented and run a homebuilt quad encoder with index on my lathe for several years and through a couple upgrades. I also use an Omron optical encoder to rigid tap on my mill. I suggest paying $20 for an omron 1000ppr encoder and 3d printing GT2 timing pulleys. You can use a GT2 belt or just use a rubber band. 

If you really want to get crazy you can take a old printer apart. Most old printers use plastic optical wheels with an easily accessible encoders all over the place. The problem I ran into was cutting the wheel to fit around the shaft. These will have crazy resolution but the mesa encoder can probably handle that. (check ppr for the rpm you want vs maximum mesa encoder frequency)

It depends on how you created your hal initial file, as the signal names may be different,
so a suggestion may not work. Best to create the connection with the configuration program
or read through the hal file to see if there is already a signal used for the desired connection.

If I made a new signal (probe) it would be something like:

net probe motion.probe-input  <= hm2_7i95.0.gpio.054.in

Then start linuxcnc with your selected configuration.
It should see your modified .ui file and use it.

If it does not, change the INI to debug mode (DISPLAY = qtvcp -d qtaxis)
run linuxcnc in a terminal and post the output.

I have completed the preliminary version of the documentation.
proma-elektronika.com/download/c2v_laser_sense_en.pdf