Advanced Search

Hello Don,
you seem like you are an expert in servos field, so I will try my luck with this question:

I have both delta B3-L and B3-Ethercat drives
B3-L (step-dir) are currently hooked up to my machine using mesa card 7i95t and they are tuned via delta software to practically zero following error when checking scope in delta software (except for some minor when starting and stopping. like 0.006mm on 5mm ball screw).
However I am unable to accomplish the same results on EtherCAT drives, even though I compared all parameters. The filters are all off or minimum, but i have approximately 12000PUU following error on EtherCAT drives even though the motors are running freely without any load - they are only on the test bench.
The only difference is load inertia ratio which is ~5 when on machine and 0 in case of unconnected motors.
Might it be, that unconnected motors are unable to be tuned properly?

I am learning. I wonder why you would use the 7I76 for the spindle motor control; would the 7I97T not be enough for your needs? Also, 7I76 uses step direction, meaning your spindle uses step direction, not PWM. I am trying to learn as much as possible before starting my build. Please share your thoughts

PCW, that might have been a winner!

I am still running on my Live distribution because something seems screwed up with my installation and I can't re-install for some odd reason, however:

I tried:

cd / /sys/bus/pci/devices/0000:03.2
sudo su
echo -n "1" > enable

Then installed the driver with io=0xd010, io=0xd000 irq=18

I went into LinuxCNC and ordered a 5mm move on X, and when I monitor pin 2, I see a series of 5V pulses that are roughly 3.5 microseconds wide and spaced about 5.1ms apart.

I tried connecting the router (CNCEST 6040 Chinese, bog standard) and...

Z-axis seemed to work, albeit it seemed like it was working in the wrong direction (is -Z = up or down typically), very slow, and probably less movement than it should have been -- I presume thi is just configuration, no big deal.

Y-axis made noise when going in the + direction like the motor was doing something but there was no actual motion. I think I may have seen the leadscrew occasionally jump. When I went in the - direction there was a constant alarm sound which stopped if I hit the "+" after a second or two. Perhaps it was hitting an endstop?

X-axis didn't seem to do anything, at all, no noise, no motion, etc.

I couldn't get the spindle to work at all from LinuxCNC although it worked through the box.

You can go right to the directory and choose what to download. The Linuxcnc version may have been updated and the old file replaced. Same with the kernel.

http.us.debian.org/debian/pool/main/l/linuxcnc/

ftp.de.debian.org/debian/pool/main/l/linux-signed-amd64/

First thing I'd do then, is get X working if you haven't already, and run the board through it's paces, if you can't get this done or it feels really sluggish I wouldn't attempt trying Linuxcnc on this board.
In saying all that machinekit might be an option as they gave split a lot of it up, but I don't know if there's been any recent development. The mailing list is a bit slow, but answers do come. I think there maybe a few fellows hanging around there that have done some embedded stuff.

Fantastic, glad you found a solution.
If you could add a note in this topic it would be great
forum.linuxcnc.org/9-installing-linuxcnc...on-the-parallel-port

Trying to setup the Hitachi WJ200 with this fanuc spindle motor. I got the startup test to spin the motor in the Hitachi manual. As I look into interfacing the modbus with my 7i96, damn I'm in deeper than I can swim. I bought a few of the cheap usb to rs485 sticks as a stepping stone since there are more threads and youtube videos talking about that. Reading about the simple positioning (for tool change orientation) in the manual is not too simple either. I see the wiring for the encoder neglects the index and says channel a can be 5-24v but channel b can only be 24v.

Discouraging. 

... and finally synaptics screenshot, showing all linuxcnc relevant packages' versions

 
Greez
chris

Thanks gentlemen for the bump and the bit file!

Wow this may be what I just need, Thank you so much for that information

Thank you. I will look into the driver's cost

Attached my /etc/apt/sources.list
Was the same all the time and on all my linuxcnc pcs. Don't know, still the same
...sim/axis/remap/manual-toolchange-with-tool-length-switch$ head -n 3 manualtoolchange.ini
[EMC]
VERSION = 1.1
MACHINE = manualtoolchange.ini

...sim/axis/remap/manual-toolchange-with-tool-length-switch$ head -n 5 manualtoolchange.ui
<?xml version="1.0"?>
<interface>
<requires lib="gtk+" version="2.16"/>
<!-- interface-requires gladevcp 0.0 -->
<!-- interface-naming-policy project-wide -->

It's actually the very same as on github (github.com/LinuxCNC/linuxcnc/blob/master.../manualtoolchange.ui)
So, no joy, and I ask myself:

where do you guys get your " manualtoolchange.ui" from?
Could it be it is a library issue? Permissions? But if, where?
If you guys have it up running out of the box, it must be something very obvious.

Thx a ton for your help
Greez
chris

 

rodw and aciera

Thx for your support

Did as instructed, though I was already on "deb linuxcnc.org/ bookworm base 2.9-uspace ". Same result. So I chose a fresh install from linuxcnc.org/downloads/, using www.linuxcnc.org/iso/linuxcnc_2.9.2-amd64.hybrid.iso. Still same result.

[Got a problem with the editor, so this post is spread over several messages, sorry]

 

Thanks a lot! There was a connector labelled PAR on the motherboard - I used a generic cable and the BIOS didn't detect the port. Scratched around a bit and found that the HP cable connects pin 21 and 25 on the output side - I added a jumper across those - and voila! working 100%. Nothing else was needed to get it working with the latest linuxcnc version

That's a good advice, though the AI tool like chatgpt sometimes just made up the content, anyway it indeed speeds up the process of learning . Thank you .