[solved] Teacup unreliable endstops

I recently built my first prusa i3, using Gen7 Electronics 2.0 ARM and the Teacup firmware. Unfortunately my endstops show very strange behaviour. They are simple mechanical endstops in a normally closed configuration. As can be seen on the pcb layout (Link), they are in series, so they should all behave the same.
In the closed state all work as expected, but as soon as one is pressed/the circuit is opened the behaviour starts to get unpredictable.

z min is the only reliable one, the other two show completely random and fast-changing states with M119 as soon as they are triggered. I suspected noise from the motor wires, but the endstop wire pairs are twisted, and checking the voltage of PIO1_8 with an oscilloscope doesn't show much noise. The voltages are ~3.3 V when closed and 0.5-0.6 when opened, regardless of endstop used.

This is my config.h:

[pastebin.com]



Any suggestions?

Edited 1 time(s). Last edit at 10/01/2016 06:49AM by Fivefold.

Maybe try to enable internal pull ups ? Extract from your config.h :

/** \def USE_INTERNAL_PULLUPS
  The ATmega has internal pullup resistors on it's input pins which are
  counterproductive with the commonly used eletronic endstops, so they should
  be switched off. For other endstops, like mechanical ones, you may want to
  uncomment this.
*/
//#define USE_INTERNAL_PULLUPS

---

Most of my technical comments should be correct, but is THIS one ?
Anyway, as a rule of thumb, always double check what people write.

You might also maybe want to try to play with that part - sorry if you did already, just call me captain obvious then :

/** \def ENDSTOP_STEPS
  Number of steps to run into the endstops intentionally. As endstops trigger
  false alarm sometimes, Teacup debounces them by counting a number of
  consecutive positives.
 
  Use 4 or less for reliable endstops, 8 or even more for flaky ones.
 
    Valid range: 1...255.
*/
#define ENDSTOP_STEPS            5

---

Most of my technical comments should be correct, but is THIS one ?
Anyway, as a rule of thumb, always double check what people write.

Quote
Fivefold
z min is the only reliable one, the other two show completely random and fast-changing states with M119 as soon as they are triggered.

My fault. With a triggered endstop there's no pull resistor at all, so the signal indeed picks up noise floating around.

The trick is to turn on pulldown resistors on the signal pin. And here's the fault: I forgot to push the commit which allows to do this. ATmegas have no pulldown resistors, so this requires a few lines of new code. I hope I get it done later today, currently it's a quick hack.

Doing this in hardware is also possible, a 10k or 100k resistor from the leftmost of the 6 connector pins to GND.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Done, just pushed to branch experimental. Printer configuration files now have this new switch:

/** \def USE_INTERNAL_PULLDOWNS

  Some controller chips feature internal pulldown resistors on their input
  pins, which get used for endstops by turning on this switch. Don't turn it
  on when using endstops which need no pull resistor, e.g. optical endstops,
  because pull resistors are counterproductive there.

  One can't use USE_INTERNAL_PULLDOWNS and USE_INTERNAL_PULLUPS at the same
  time, of course.
*/
//#define USE_INTERNAL_PULLDOWNS

If you don't want to start over with your printer configuration, just add this #define on the previous one.

Let's cross fingers it works, because I didn't test too much to get this into the public quickly.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

The internal pulldowns did the trick, everything works as expected, now! For anyone having the same problem and trying to keep the old config by just adding the pulldown#define: there's been a small change in experimental to #define TEMP_EWMA, too. Just compare the printer.mendel90.h file with your printer.h/config.h if you get errors about temp_ewma when compiling.

Thank you Traumflug for the fast response and thanks to DeuxVis too!