Advanced Search

talking specifically about delta servos.

Having trouble with uploading the report file... trying again.

 

AttributeError: 'Info' object has no attribute 'MIN_SPINDLE_0_SPEED'. Did you mean: 'MIN_SPINDLE_SPEED'?

I've added instances of each setting in several places, so I don't understand.

I'm attaching my .ini and HAL file plus the error report. 

   
 

With this config for mb2hal i have a working system
And i have this config for mesa modbus.
Is that correct ?

That looks OK

Where things go bad with bare FPGA inputs is that say a VFD is running and an encoder wire connects
to frame ground accidentally, well unfortunately this frame ground has ~1 usec 1V negative high current
spikes  from the VFDs 350V square wave output and finite inductance of the frame ground to earth ground.
These spikes will easily short the input clamp diodes. ("Ground" is just a tapped inductor for high frequencies)

The encoder pin/parameter specifications are in the hostmot2 manual page:

man hostmot2

It may be that firmware that better matches your application can be made
(MPGs are better with MPG encoders rather than motion encoders because
MPG encoders support 1X mode which works better with MPGs having detents)

 

Forgot to say that the lcec.0.0.home-mode 34 should be changed to the home mode you want. 34 is just handy for testing without external switches connected (34 just returns to index).

After many hours of studying homing.c, cia402.comp, cia402_homecomp.comp and various forum posts, I finally found the last key to making LC10E drive homing work.  Here is my summary of all the changes needed to RodW's example custom homing configuration.  I can now click "Home All" in Axis and have it do a proper mode 6 home.  It still leaves all the axes with "H" in the preview window (rather than the round "survey marker" symbol), but MDI works and everything else seems to function normally.

Changes to get Lichuan LC10E homing working:

1. cia402.comp line 240

      change (statusword >> 10) to (statusword >> 12)

2. ethercat-conf.xml

      add to pdo 1600 section the lines:

           <pdoEntry idx="6098" subIdx="00" bitLen="8"  halPin="home-mode" halType="u32"/>
           <pdoEntry idx="6099" subIdx="01" bitLen="32" halPin="home-hi-speed" halType="u32"/>
           <pdoEntry idx="6099" subIdx="02" bitLen="32" halPin="home-lo-speed" halType="u32"/>
           <pdoEntry idx="609A" subIdx="00" bitLen="32" halPin="home-acc" halType="u32"/>
           <pdoEntry idx="2005" subIdx="24" bitLen="16" halPin="home-timeout" halType="u32"/>

3. cia402.hal

       after loadrt's add lines:

           setp lcec.0.0.home-mode 34
           setp lcec.0.0.home-hi-speed 50000
           setp lcec.0.0.home-lo-speed 50000
           setp lcec.0.0.home-acc 100000
           setp lcec.0.0.home-timeout 5000

      after # homing signals change nets x-home* to:

           net x-home-request joint.0.request-cia402-homing   cia402.0.home joint.0.index-enable
           net x-homing       joint.0.is-cia402-homing       cia402.0.stat-homing
           net x-homed       joint.0.cia402-homing-finished  cia402.0.stat-homed

4. cia402.ini

       in [joint_0] be sure:

           HOME_SEQUENCE = 0 (or whatever sequence number is appropriate)
           HOME_USE_INDEX = YES
           HOME_SEARCH_VEL = 0
           HOME_SEARCH_VEL = 1 (or any non-zero value)

I hope this helps somebody else.
-- Ralph

Thanks Guys,

Changed settings to reflect e-stop as a button and i'm able to select it now.

Still have a blank or other shadow in the box where you would normally see what you have loaded. Not sure why.

An now that everything has been double and triple checked, i can't form any connection to the motors.

sharp_reaper wrote:

When I turn on QTPlasmaC the value of Arc Voltage is 145.
The encoder velocity pin seems to be working fine on halshow.
 

Sounds like you have not correctly calibrated qtplasmac for your THCAD settings
The THCAD 0 volt frequency should be what you observe on the encoder velocity pin give or take some bouncing around.
This offset needs to be entered in qtplasmac.

If you read the QTplasmac docs, there is a link to a calculator to determine your settings.

PCW wrote:

The first P1 encoder is  encoder 1 so the position would be:

hm2_7i96s.0.encoder.01.position

and the most important parameter for that encoder would be:

hm2_7i96s.0.encoder.01.scale

 

After your warning (thanks) I was going to solder up a little proto-board with some protection diodes for the negative rail. Then decided to just do a quick and dirty actual PCB. I will obviously be careful when connecting, and to your point chassis ground is not zero volts. I inserted a 100r series resistor and TVS, please see attached and let me know if you think this is adequate?

I'm ready to press the send button for a free PCB, but I can tweak if you have any suggestions.
Re: hm2_7i96s.0.encoder.01.scale, can I assume that's a float value? Is there documentation available for the finer details of the 7i96s_p1encd.bin?
 

langdons wrote:

You say LinuxCNC is way way better than Mach 3.

How so?

A work in progress by a user in NZ but try commanding a pallet changer on a horizontal 5 axis machines like this in Mach3.

This machine fills a shed.
Many people use Linuxcnc to retrofit big commercial machines because of its flexibility and extensibility.
Here the user had hit the limits of using classic ladder and wrote a custom Linuxcnc C component to do this. A week earlier, he knew nothing about programming and is now loving it!

ihavenofish wrote:

cool.

downside to the ethercat is its ONLY ethercat. i wish it has pulse as well. would be more versatile.

It does if you use an Ethercat stepper driver. There are many. I use rtelligent ECT86

From google translatedoes not make much sense....so
VFD or Servo?
With VFD, yes, 7i96S can be used as it has the 0-10V control and has encoder inputs.

Hi, the outputs on the NVEM are not an output voltage like your thinking. They are as shown in the diagram attached. Basically a switch sinking to ground.

These are Linux kernel patches, so I guess where the kernel devs hangout.