Advanced Search

I hadn't tried getting into the wiki in a few so I gave it a shot this morning and it came right up like a boss.

Don't know what changed but whoever it was gets a big thanks from me!

Hast Du in der maschinenname.ini bei JOINT_0, JOINT_1 usw die HOME_SEQUENCE auf 0 gestellt?
Nicht unter AXIS_x

Hello,

I'm struggling on this one...
You can find enclosed my conf files.
I have a 7i96s and a 7i73 and everything seems to work fine.
I have 2 limit switches on my X axes (lathe), They are both connected on the 7i73 and it seems OK. 
On the X+, my limit switch is also my Home switch.
The homing procedure goes well. But once it's homed, the offset of 10mm is applied, then when the lathe is idle, I get this joint 0 on limit switch error.

On HAL show, I don't see the pin changing color like if it was a noise or something. There is 10K pull up resistor on these pins if I remember well the documentation of the 7i73. So, what is going on? I'm a bit (totally) lost on this one.

Thank you for your help


 

Thanks Rod I will check that out.

I have started to set up LCNC with the RT kernel on my system (N3150 AMD630) running a fresh installation of LMDE6. I disabled UEFI and a list of other parameters as recommended eslsewhere, and used the download instructions at the top of this topic (the 5.4.279 versions).

At boot, if I select 'advanced option' I can select LMDE6 with Linux 5.4.279-rtai-amd64.

However, I get the following error:
bad shim signature
error: you need to load the kernel first.

I assume I have skipped something in the setup process. What should I do to correct this?

Update: This was fixed by (again) disabling secure boot - although it keeps saying it is enabled.
$ uname -a tells it is running rtai-amd64
There was a warning that the kernel was not compiled with ?? (message disappeared fast) but the system runs.
Now will try LCNC.

Update2: Switched Mainboard to A960D+ (6 core AMD), installed Debian12 and the rtai debs. Removed in BIOS anything not needed for a legacy setup. After setting isolcpu parameters (2, 4,5 - thanks for rt_setup.sh !!) got the best legacy timing ever on one of my systems: 6 - 8 uSec with 4 gears and a lot of browsing. Will test and apply the rtai version on my other systems as well.

Thanks and Regards,
Robert

 

Marius, have a look at package cloudthere is a free plan I used to host the Linuxcnc Ethercat driver until that was taken over by iGH and Scott Laird

looks perfect. Also seems terminal E is common to all of the positions. If so, connect 24v there. 

rodw wrote:

fully_defined wrote:

rodw wrote:

Not quite what I was looking for but something like this. SOme have stops that let you limit the number of positions
au.element14.com/lorlin/bcks1004/rotary-...MA-RECO-PDP-STM72189
 
In the US, this is all Digikey has that's coded for hexadecimal. My own personal lack of vocabulary for searching is also something I hope to improve as part of this conversation.

www.digikey.com/en/products/filter/rotar...AMqcAnAJYA7AOYgAvtKA

Me too! so I am hesitant to make a recommendation for that reason but I think that a 3 pole or 4 pole 12 position rotary switch also does what we want. eg decodes a position to a bit map on the poles.
 

I found the datasheet for Lorlin switches, and I like what I see: lorlinelectronics.co.uk/wp-content/uploa...-ROTARY-April-24.pdf.

This seems to be what's available at Mouser here in the US: www.mouser.com/c/?q=Lorlin%20BCK, and I think that based on what limited information I see, I want BCK1006. I am assuming I can cut down the shaft.

I ordered 5.

 

I have seen those, but I have until now dismissed them as something weird that I don't understand the purpose of.

According to the C&K switches datasheet: www.ckswitches.com/media/1349/arotary.pdf, the increase in the number of poles corresponds to a decrease in the number of positions by a factor of the number of poles, with no further explanation. The schematic just looks like three or four separate switches with three or four positions each, and I don't know the use case for that. I was hoping to get a better explanation here, and it looks like we might be on to something.

Is it possible that this is the exact same thing as hexadecimal coded, just without using those words? And that I don't even need diodes?

That is what I do mostly Rob but there are many users that are not capable of doing their own compile.
When I setup a machine for a customer or user, I use the buildbots to make it easy for them to stay current. MX Linux provides an easy way to update. This suits most users and therefore the buildbots are really necessary.

I am looking at setting up a private repository where I can place the compiled DEBS on a weekly basis. I will point their Synaptic Manager to the private repository to do automatic updates. I am just having a hard time to get the files signed in the correct manner for apt to recognize them as a proper repository.

Not quite as convenient but if this pull request was committed, making Debs is very trivial as it uses the Debian build environment to get the dependencies so it becomes a 3-4 line process.
github.com/LinuxCNC/linuxcnc/pull/2647
eg as I tested: in the comments.[/code] 

fully_defined wrote:

rodw wrote:

Not quite what I was looking for but something like this. SOme have stops that let you limit the number of positions
au.element14.com/lorlin/bcks1004/rotary-...MA-RECO-PDP-STM72189
 
In the US, this is all Digikey has that's coded for hexadecimal. My own personal lack of vocabulary for searching is also something I hope to improve as part of this conversation.

www.digikey.com/en/products/filter/rotar...AMqcAnAJYA7AOYgAvtKA

Me too! so I am hesitant to make a recommendation for that reason but I think that a 3 pole or 4 pole 12 position rotary switch also does what we want. eg decodes a position to a bit map on the poles.

PCW wrote:

As an aside, I am also a little confused why PCW suggested 00, 11, 01, 10, instead of 00, 01, 10, 11. It seems more logical to me to order it like legit binary, and it's not like it's harder to connect it this way.

The 4  --> 2 diode encoder circuit I posted is in binary order but you need to consider
that the 7I73 inputs have pullup resistors so are high when unconnected and low when
grounded through a diode so:

SWITCH POSITION
0  Grounds inputs 0,1 with CR1 and CR2 = 00b = 0d
1  Grounds input 1 with CR3 = 01b = 1d
2  Grounds input 0 with CR4 = 10b = 2d
3 No connect so = 11b = 3d
 

Can you recommend a source of information that can decode what you just said?

If this is happening at the end of cuts, it’s generally caused by the arc going out either because the slug dropped or the material was already cut while the machine still has some distance to move before reaching the M5.

If that’s the case, then you can set a lost arc delay as described here:

linuxcnc.org/docs/devel/html/plasma/qtpl...html#_lost_arc_delay