Advanced Search

zippoffs wrote:

I have only ever experienced issues when the VFD is on.

I tried disconnecting everything from the BOB except for the 24v coming in and the parallel port and USB from the PC (I switched the VFD to manual control). LinuxCNC ran the program with out any trips even though the VFD was running, so the BOB itself can't be the issue. I tried reconnecting one of the stepper driver pulse/direction signal wires (the steppers were completely off) and ran the program again, and it did trip, even though that wire is shielded and grounded.


 

Sounds like effects of a ground loop. You have more than one ground connection from point A to point B at one or more points anywhere on the system. You want exactly one ground circuit from each 'groundable' piece to one single point. It is called single point ground aka star connection. Redundant ground paths create ground loops which powerfully amplify interference effects from the vfd, they behave as you described above.

Potential sources of redundant grounds:
1. shield grounded at both ends
2. motor cord have ground wire
3. additional frame ground to the machine when the motor is already in electrical contact with the machine
4. the machine may be in physical conductive contact with another machine that is also grounded
5. is the machine on a steel workbench that has a grounded outlet?

Interference mitigation can be a bit of an art because of the numerous variables involved. Usually not everything has to be perfect, you can cheat some, but if the BOB is within inches of any AC when there is also a VFD nearby then I say you will have trouble at least randomly.

Does LinuxCNC see the card?
 

Good. EoE has to go.

But you see automata's posts here? Do not be surprised if it still does not work.
Thankfully, automata has kindly posted the solution here.
Please report back how it goes and if there are problems with lcec
I hope automata will guide you to use his solution.

Do you already know Andy's lathe macros?
forum.linuxcnc.org/41-guis/26550-lathe-macros
The graphical interface is fantastic.

Rather than reinventing the wheel, have you considered what it is which you are adding with you macros? Maybe you can coordinate with Andy to make them even more awesome.
Cheers,
Mark

Blocking in the "make yer own" accessories. side arm ISO20 ATC, Stand, stainless chip pan, enclosure... Fun

I just switched to default jumper positions:

Result is still the same:

dmesg and lspci outputs are provided.

QtPlasmaC updates - Master (v2.10):

• fix tabbing order for recent GUI additions
• make all spinboxes consistent for Auto Activate THC

That will likely not work, (not meet timing) you must jumper the card for 5I25 mode to use 5I25 firmware (the default)

But it is 6I25, not 5I25. 



Jumpers are:
And it reports 66MHz.

I forgot to mention, VFDs radiate interference out of both incoming and outgoing power wires. Outgoing is by far the worst, but incoming side also quite bad. Thanks for the BOB picture. I circled the power wires going past BOB. Those certainly radiate interference when the VFD is running, everything all the way back to the breaker panel and beyond is affected to a certain extent. I prefer to keep all that AC wiring a minimum of 12" from the BOB.

How long is your parallel port cable? Is it shielded? The longer it is, the more interference it can collect. If it is shielded then I would think no worry.

Best you make sure that your drive is indeed an Ethercat drive first of all.

If so, I think you can try this small example.
github.com/dbraun1981/hal-cia402/tree/main/example
I wouldn't be surprised if it works.
Some drives need some additional settings, if you are lucky yours will run directly.
Anyway, this little example shows the structure of how to connect an Ethercat servo drive to LinuxCNC.

The encoder position is used internally in the drive for feedback.
The encoder position is feed back to linuxcnc also, see the example hal file.
After you get it going, you should tune the drive's PID parameters,
just like is needed for every servo drive.



 

Just a few thoughts.

If for professional production, take it with a lot of caution. I only toyed around for hobby use.

300 rpm is fairly quick for a rotary axis.
How fast can the hob run during the cut?
The machine can tilt the spindle, can it?
You do not really need the spindle position, under the condition that the spindle is brought up to speed, before you start the cut and not stopped before it is finished.
A simple incremental encoder on the spindle probably works fine. More than 100 counts per rev and it might already work.
I run the rotary open loop, as long as the motor can keep up with torque and speed, encoder on the rotary is "nice to have", but not essential.

You can even get away with a semi electronic gearing for straight gears this will get rid of the computer altogether.


There is (was) an instruction on the Linux Wiki where the gear hobbing was explained.
wiki.linuxcnc.org/cgi-bin/wiki.pl?Hobbing

The spindle syncronisation is independent from x-y-z axis, as long as you do not. a) want to cut helical gears b) want to run hob shifting c) want to try hobbing wormgears with tangential hobbing.

Thank you so much, Mr.Hakan.

I installed "go" package and then everything seems to be successful.

I appreciate your kindness.

I will try to connec YASKAWA Servo tomorrow.

I have thought of it but i havent the skills yet to build a new driver ..
the drive is
cdn.delta-emea.com/cs/download/file/4641...2_UM_EN_20201209.pdf

Ok is CSP mode only motor encoder ? i am not sure at all now..

when i started the build i planned to use
6.2.1 Position Command in PT Mode
Becurse of the feedback from the scale

you have to adjust the resolution over SDOs during start up and then to reduce it over float factor in the SMs ... if you are scaling 1 / something big it will just create you crazy numbers then ... 

share config at first place ... then we can help you well