Advanced Search

Hello,

that looks like a lot of work to do and I hope that the OTK-post processor will be
helpful for you as it just adds the rotation for the knife and, if necessary, the
retractions.
FYI: The stepcraft OTK-3 is a little bit "special", because it requires an unusual 30V power supply and I also had to build a small circuit to connect it to my homemade CNC router.(running at 48V) and to handle the homing and safety stop signals.
Just let me know if you encounter any problems with the post processor.. Maybe
I can fix or improve it.

I rechecked it multiple times myself. According to --readhmid, it should be fine. In the end,

I gave up and purchased a 7i97T. That should solve all the current problems and all it cost is money
That is the card I use in my mill. It has one Smart Serial port, and I will need two (for the 7i73 and one 7i85). I plan to build this small board from Talla83 (github.com/talla83/RS422-SmartSerial-Adapter
), but for starting out, the I/O on the 7i97T itself should be sufficient to get most of the machine running.

A way to debug this is to plot (with halscope) the joint 2 commanded and feedback positions
(setting the following error limits higher may help debugging as well)
 

Here is the connector layout. We have the P value set to 0 because we aren't able to get the axis to move when it is set at 1000. I thought we would try to get the motion working better before worrying about it. If we need to figure that out first, I can work on that first. 

Thanks,

-jblanscett

Is it possible that you have ended up with the wrong firmware flashed? I looked back and the last one I saw you mention was 7i92 plus 2x 7i85. Sounds okay, but maybe Peter can confirm that?

Lower the jog speed a lot, like something really slow and try to jog again.

It's a collaborative arm. I haven't seen an earlier LCNC project that uses the dual encoder joints used in collaborative arms.

They typically use Cycloidal drives with encoders on both the input and output to measure the strain. The difference between the two encoders gives you the strain. This sounds like I need to develop a new HAL component.

For example:
www.celeramotion.com/mechatronics-soluti...ollaborative-robots/

The tool path needs to be plotted relative to the work piece. (tool -> world -> work)

The problem is that your tool position is plotted in World coordinates (tool -> world) which works if only the tool moves because the Work coordinate system remains identical to the World coordinates.
However on a machine like yours where the tool AND the work move in space the plotter needs to track both the tool tip as well as the work piece.
This is why the 'vismach' tool needs the capture() function for the tooltip as well as for work.

Not sure but maybe the backplot in ProbeBasic only plots tool position in world coordinates instead of plotting tool position in Work coordinates?

Hi, What piece You need?

If you don't mind updating LinuxCNC, you might be better off with more current replacements for the 5I20:

PCIE --> 6I24-16
Ethernet --> 7I80HDT

The 5I20 is a 20 year old card...

Do you have the step/dir pinout information for the drive?

It does look like it may be single ended from the pictures of the
step/dir modes.

Also in your .ini file, you have the PID P term set to 0.
This should be 1000 (for a 1 ms servo thread period)
A 0 P term will cause position drift

See GIF file I have uploaded. The tool marker from view port is moving in opposite direction. I want it should be below spindle tooltip. As it moves while jogging in X or Z direction. But Y jogging time, table is moving correctly. But tool marker is moving in opposite direction.

How can I resolve this?

The problem is that the backplot needs to move with the table.

I suspect that is already in the works.