Inconsistent z probe reading

Hey there,

I got the problem that everytime I do auto bed levelling I will have a slightley different detection range of my z probe
(but within the leveling procedure it seems like it stays the same).

It behaves like this:
1. I start the gcode from SD
2. the printer homes all axis and starts auto leveling (before heating either nozzle or bed)
3. the printer will return to X30 Y30 Z0 in order to prevent ooze when heating the nozzle
4. now I can see that the nozzle is a little too high above the bed so I abort the print and adjust the z offset to a more negative value so the nozzle is about .1mm above the bed and store settings
then I repeat steps 1 to 3 but now the nozzle scratches the bed so I adjust the z offset to a less negative value so the nozzle is .1mm above the bed again but after auto bed leveling the nozzle will scratch the bed again

so it seems like the the detection range of my z probe is decreasing the longer the printer runs.

Things I have tried so far:
z probe was running on 5V (from the z endstop port) I set it up to run on 12V
(but the Endstopp pin on the board will only get 5V with the help of a schottky diode which is ultra fast switching and has no thermal dependencies etc.)
but this did not help


Has anyone of you a clue why it behaves like it does? Auto bed leveling doesnt make sense when you always have to adjust the offset again till you are lucky.

What type of Z probe is it? Some types give different trigger heights depending on temperature and/or moisture and/or supply voltage.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

It is one of these: [www.google.de]
But mine is blue.
Its a SN04-n NPN NO

I have two of those. They are both junk.
I had to replace the first one as it completely stopped working. The second does not consistently trigger on the bed. I could improve the results by feeding it higher voltages such as 12 or 24 but 5V was horrible.
I am trying to get @dc42's IR sensor working but for now and the past several weeks I just went back to using limit switches.

I had a Green SN04-N NPN NO 8-24volts it worked perfectly for the 6 months that I used it. Though I wanted some thing that did not hang 65mm behind the nozzle.
So decided to look around and found these low profiles Flat Inductive Proximity Sensors by Omron TL-W3Mc2.

Had to invert the End_stop switch but other wise it sits within 25mm of the nozzle, could be closer though did not see why. working now for 3 months on 5Vdc.
Heck just noticed the supply I am feeding it with is only 5vdc operating voltage, might want to change that someday.