As a result,  www.forum.linuxcnc.org/38-general-linuxc...lock?start=30#344768

you could try a shutdown.hal file that setp's it to what you want. You may have to unlink the pin as well

Looking at your first post, its not network related. Note those previous settings do not persist. Running out of ideas. As PCW mentioned ensure the NIC interrupt is pinned to isolated core 5

You mentioned graphics so you probably have an Intel i915 GPU which hangs on my PC. Add these to grub to lock it downi915.enable_dc=0 i915.enable_fbc=0 i915.enable_psr=0 

I've got a SSR in the latching loop of my spindle contactor being controlled by one of the outputs from the 7i76
net machine-is-enabled hm2_5i25.0.7i76.0.0.output-07

Short of re-wiring, is there a way to temporarily leave that output turned ON when linuxcnc is being closed and re-started to pick up new config changes as I mess around inside the HAL and ini files setting up my new VFD and gear selection code? I can't help but think having the 220v power ripped out from the vfd over and over again isn't doing it any good
M

 

RotarySMP wrote:

I am not sure what to do about motion to a tool change position. Given the excessive size of my carousel, relative to the lathe envelope, and wide variety of tool geometries, tail stock etc I suspect it is safer to not automatically move to a tool change position. Do you have thoughts on that?
Cheers
Mark
 

The capstain turret lathes I've seen move to a safe location - generally max X & max Z - before tool changes (Hardinge, Haas CL1) but none of those have tailstocks.

Do you have a tailstock detector in the carriage?  If so, you could make a toolchange macro conditional so you aren't right up against the tailstock, potentially swinging a long drill in to the tailstock snout.

If not, you might think about installing such a detector.

00:1f.6 Ethernet controller: Intel Corporation Ethernet Connection (7) I219-V (rev 10)
DeviceName: Onboard - Ethernet
Subsystem: Lenovo Ethernet Connection (7) I219-V
Flags: bus master, fast devsel, latency 0, IRQ 120
Memory at b1000000 (32-bit, non-prefetchable) [size=128K]
Capabilities: <access denied>
Kernel driver in use: e1000e
Kernel modules: e1000e

eee was already off, not sure about gro, lro or tso but I will try with them disabled also.

thanks, 6 core PC with one isolated = 5 available so that's right. i5 should be fine
What network card? is it Realtek?

I've been researching this today. Here are a few more obscure settings that may help where eno1 is the NOC name from ip a

Its an odd looking voltage plot.
I would look at the time after arc ok the THC is activated. To me it looks like it needs to be 2 seconds. I just can't see the setting I'm looking for.
Early activation could be disabled by the big fluctuations in voltage.

The discussion should occur on the PR. its up to the dev team, not forum users so you neet to convince the gatekeepers. Personally 80 files in one commit is too much to clearly understand and debug build errors.  Perhaps the component is in one PR and the sim in another.

No, it didn't

mfg
Juergen
 

What aspects of Andy's carousel.comp have you identified as needing improvment, and/or what other features/funtions are you adding with this comp?

Do you have a github repository, or just keeping the current version of code uploaded to the initial post in this thread?

My Schaublin tool changer is extremely simple, with just a single input to open the tool change pneumatic valve, for a fixed duration, and then checking the next feedback pin went high. Cycling through until the target position is reached. I don't need much of the code like tool unlock, tool lock as that is all pneumatically sequenced.

I am not sure what to do about motion to a tool change position. Given the excessive size of my carousel, relative to the lathe envelope, and wide variety of tool geometries, tail stock etc I suspect it is safer to not automatically move to a tool change position. Do you have thoughts on that?
Cheers
Mark

PCW: Thanks for that, I'll consider changing kernel, or maybe trying with another pc.

rodw: processor: Intel(R) Core(TM) i5-8400T CPU @ 1.70GHz, echo $(nproc) prints 5.

pos is of course the number of rotations since zeroing on index. This is the main driver for
the synchronized move, the z-axis is being forced to follow pos.

Rotational speed isn't mandatory, but convenient to have.
You use rotational speed for the at-speed pin and is displayed in the gui.
I guess enc-frequency-scale should be set to 1/2000, so 0.0005, for rps.
For rpm 1/(2000*60) = 8.333333e-6.

Hello,
Is the Forum where discussion of new hal comp takes place?

On Github, on my fork there is a Discussion button
On Github in LinuxCNC.linuxcnc there is no button.
Google said the Forum was the main location for discussion.

Where is proper place to discuss?

My case is
  problems with package creation and a new comp.
 My comp builds and runs fine locally,
but a PR fails during package creation.
I hope a new push will fix the packahe creation error.