GRBL v1.1e and f and endstops, I have a CNC shield its V3 and from the pin assignment it seems after grbl v0.9 I have to move the Z axis endsops from the Z-/+ pins to the spindir header ( Ive tired in both locations and neither work ) seeing as that’s where the pin 12 from the arduino connects to.
When I run the home command the Z lifts but the switch doesn’t trigger it just keeps going.
I frequently see error9
Ive gone through every document I can find reflashed it tried different gcode senders. about the only thing I haven’t done is gone back to a old version of grbl.
What I can see is looks like many people struggle with endstops and it kinda shouldn’t be a hard thing to get right.
Any insight if you’ve had this and came right would be appreciated
Looking at the code suggests you might try pin 12 instead of pin 11. Especially if you have variable spindle enabled as pin 11 has PWM to allow for the variable spindle.
Doh! I guess I should take reading comprehension so I would see that you have tried 11 and 12. I don’t know what to suggest.
Wow gerbil hit v1? So that’s it then? It’s done now? When you downgrade does your machine work then? Doing that would narrow it down to a software issue. Which logic would imply it was, but there is no guarantee that is the case. So I’d still like an A B comparison, just to make sure. Troubleshooting is the process of elimination through testing. Going from does not work to works is a powerful indicator in the process too. Does not work to does not work proves nothing at all. Well, I suppose it proves you have yet to find the cause? Lately I’ve been having great luck just hitting stuff, and getting it to work again. Have you tried that? Although whenever I do that I never feel terribly confident about the fix. Percussion tune up is an advanced skill though. It takes a long time to learn just how hard to strike things in order to effect a repair, without causing further damage.
@Kyle_Kerr years ago I worked in a gas station and our point of sale machine stopped working. We had a number we could call for service to get it fixed. So we call and the person tells us to pick the machine up about 6 inches, and drop it. OK. So we do it, and lo and behold, it works! It turned out the PCB they used warped in changing environmental conditions and that caused the socketed ICs on the board to walk out. So dropping it made them all drive back in again. After that whenever the console would fritz out we’d just drop it, to get it going again. Ain’t nobody got time to fix stuff.
Connect to it using a terminal (e.g. PuTTY) and send it the ? command. It should send a status block back. Look for |Pn:. Within that field will be X, Y, and Z values if it sees any switches active.
Repeat when you’re holding down each switch to make sure they register properly.
I too have had major problems with end stops, one solution I’m working on is with debounce. I have added a second arduino that debounce all the endstops before passing on to GRBL.
So far works flawlessly on Ports 9, 10 and 12. Appears that the motors on the Z-axis created so much noise and vibration that it interferes with the signals.
I say so far as I have yet to fully test it but so far no problems since I installedit. It also has the advantage of using all six endstop and reducing to the three on the interface. And currently expanding to provide debounce to the Start, Hold and Stop buttons.
Looking promising.
@Kevin_Worsfold I know Arduinos are cheap but there are certainly simpler and more elegant debounce circuits you can use. You could probably use my optical isolation circuit to clean the signal up too. It is tits on a Ritz when it comes to control signal conditioning.
Let me find the schematic for it.
This, or some variation of it is the ticket for conditioning logic signals around noisy gear. The important take away is the opto being squared up by the Schmitt inverter. Though the observant can see a bit more going on in this particular snippet of circuitry. It only took me about a year to hit upon this so I suggest anyone ponder it for a bit of time themselves. Because this works, and it works rather well. http://i.imgur.com/wmbrCVI.png
Hi I thought I would share this here I am applying a universal solution to this problem for my own machine. In additìon to the proximity switches I am placing nc micro switches 5 mm (almost 1/4 inch) past the proximity switches. This is a safety backup in case my normal sensors fail. If tripped they cut the power to that particular stepper driver. Cheers
