Homing Issue "home switch inactive before backoff"

More
16 Jun 2022 07:08 - 16 Jun 2022 07:09 #245245 by little_sparky
Hi All,

I have been running PlasmaC for about 2 years now and my X-axis limit switch just randomly started playing up. When I home all axis, the Z homes fine but the X either goes slightly past/before the proximity limit switch and stops, where a "Home Switch Inactive before start of backoff move j=0" error message comes up, or it keeps going past the sensor and crashes into the hard stop. The light on the limit switch still comes on as if it is working. I can open and close PlasmaC about 2-10 times and keep re-homing and eventually it will work.

Does anyone have any ideas? Is it odd that this has randomly started after 2 years of it working fine?

As always, your help and knowledge is very much appreciated.

Thanks, Little_sparky

edit: I am using 12v proximity limit switches, no pull up resistor and no shileded cable. This is what I will be looking into, but wondered if there might be another way to identify/fix the problem.
 
Last edit: 16 Jun 2022 07:09 by little_sparky.

Please Log in or Create an account to join the conversation.

More
16 Jun 2022 07:29 #245248 by Clive S
It could be that the sensor appears to be working ( ie it lights up when triggered) but the wire going to it could be broken intermittent.

So check the wiring where it might be flexing.

Please Log in or Create an account to join the conversation.

More
16 Jun 2022 12:54 #245262 by tommylight
First, try lowering "home_search_velocity" to at least half of what you have now.

Please Log in or Create an account to join the conversation.

More
21 Jun 2022 03:55 - 21 Jun 2022 04:20 #245509 by little_sparky
Update:

So I did the above and I think it was a dodgy switch. I have purchased a new inductive proximity sensor, but this time I have a normally closed sensor instead of normally open. I have wired it up and I now need to invert the signal.

I tried to put -not at the end of the 

# - - - HOME - X - - - #
net home-x    <= hm2_7i76e.0.7i76.0.0.input-13-not

in the .Hal file however I now get an error when trying to open PlasmaC. Is there somewhere else I need to put -not? 

what is the best way to edit the base configuration? I tried to use pncconf but I had no luck.

thanks 

little_sparky



 
Last edit: 21 Jun 2022 04:20 by little_sparky.

Please Log in or Create an account to join the conversation.

More
21 Jun 2022 04:04 - 21 Jun 2022 04:16 #245510 by little_sparky
Also, I have wired the NC sensor exactly the same as the NO sensor, is this correct? Do I just need to invert the signal.

it is a PNP NC Inductive Proximity Limit Switxh
Last edit: 21 Jun 2022 04:16 by little_sparky.

Please Log in or Create an account to join the conversation.

More
22 Jun 2022 23:19 #245622 by andypugh
Appending -not is the usual way.

Though the docs:
linuxcnc.org/docs/stable/html/man/man9/h...eral%20Purpose%20I/O
Mention both in-not and in_not. That needs fixing.

I don't have a 7i76 but on my 5i25 the postfix is actually _not, so maybe try that?
 5  bit   OUT          TRUE  hm2_5i25.0.gpio.000.in
     5  bit   OUT         FALSE  hm2_5i25.0.gpio.000.in_not
     5  bit   OUT          TRUE  hm2_5i25.0.gpio.001.in
     5  bit   OUT         FALSE  hm2_5i25.0.gpio.001.in_not
     5  bit   OUT          TRUE  hm2_5i25.0.gpio.002.in
     5  bit   OUT         FALSE  hm2_5i25.0.gpio.002.in_not

Though the smart-serial driver (ie, 7i76 field IO) uses -not:
github.com/LinuxCNC/linuxcnc/blob/master...mot2/sserial.c#L1183

And it _should_ be -not according to the specifications...
linuxcnc.org/docs/2.6/html/hal/canonical-devices.html

But probably too late to make it consistent now.

Please Log in or Create an account to join the conversation.

More
23 Jun 2022 06:48 #245657 by strahlensauger
The NC and the NO are wired exactly the same,
you only need a pullup resistor if you have NPN sensors. For PNP
it is
brown is positive
blue is negative
and black is input goes to the Mesa
you just need to invert the signal in the hal file as you did.

What is the error message?
 

Please Log in or Create an account to join the conversation.

More
24 Jun 2022 12:36 - 24 Jun 2022 12:37 #245775 by little_sparky
Update: I was able to successfully wire the sensor and invert the signal, however... it appears the issue must lie deeper.

I wired the sensor directly to the terminals inside the control cabinet to bypass any issues with cables, noise from motors etc and I still get the same issue.I ran halscope to see what was going on and this is what I found.

1. The signal starts out flat 

 

but then it turns into this when I move it towards some metal, which I am not sure how to interpret.

 

Could this be something to do with the 7i76e board? Perhaps the terminal is fried or the ethernet cable is no good? I am unsure where to look next, however I am thinking I might purchase a new ethernet cable, even though the one I have now is not that old. Tomorrow I will play around with swapping some of the pins around to try and eliminate this issue.

Any help of guidance is very much appreciated.
Attachments:
Last edit: 24 Jun 2022 12:37 by little_sparky.

Please Log in or Create an account to join the conversation.

More
24 Jun 2022 13:23 #245781 by tommylight
Sparky, chasing issues where there is none is not helping you, leme explain:
-if the cable is not good, LinuxCNC might start but will loose connection at random, or not start at all
-if the Mesa board has no power or is dead, you will not be able to start LinuxCNC as it would not find the board,
-if the terminal is "fried" you would see it, it takes a lot of current to destroy the connectors Mesa uses, and there would be smoke
-the scope you provided shows Mesa and LinuxCNC working properly,
-
-change the sensor,
-check the sensor cable
-check the sensor power and power cable
-check the sensor top side for embedded metal shavings
-check the wires where they enter the connectors on the Mesa board
-if they are terminated with ferules, cut the ferules and try without them
-
---this last one is from experience, some people fold the thin wire and insert it into the ferule so it can grip better, then cut the top of the ferule when they find out it sticks out of the connector, loosing all contact. :)

Please Log in or Create an account to join the conversation.

More
24 Jun 2022 16:24 - 24 Jun 2022 16:25 #245798 by PCW
I would add that you should check the signal levels (with a voltmeter) at the
7I76E input to see if they are valid.

They should be close to 0V in one state and close to field voltage in the other state,
and never anywhere close to 1/2 of the field voltage (the input threshold)
Last edit: 24 Jun 2022 16:25 by PCW. Reason: sp

Please Log in or Create an account to join the conversation.

Moderators: snowgoer540
Time to create page: 0.130 seconds
Powered by Kunena Forum