Z Axis Homing and Movement Issue

Hey Guys,

during the setup of the last axis (Z) I copied the section from X and Y Axis in the Ini file and adjusted the Limits of movement for Z and the changed the Sign for the Homing Search and Latch. Z will be homed in High Position (Z+).

Two Issues I can not resolve:
1) During Homing (Z Axis moving Up in direction Z+ to Ref-Sensor) the UnitCount in LinuxCNC is about 1000 mm (and then linux cnc Stops and prompt a message "J2 Bewegungsende im Referenzstatus 7"). If I press again Z-Ref the the Homing proceeds well until Z-Ref Sensor is reached.

So the travel of such long way (like the display) says is crazy. I just have a travel of about 500 mm on Z.

2) After Homing processed discribed in "1)" I can move the axis DOWN with the right measurement count.
BUT if I like to move the Axis UP again it didn´t works. Just a few µm and then linux CNC says "0.000" position of Z.

It seems like it rehomed during the DOWN move. Can´t understand.
I rechecked Positions with measurement tools and in the Servodrive. Everything is real, exept the value on LinuxCNC Z Position.

 

I created a video for issue 2)

As tommylight suggests I would try lengthening the step pulses
just to eliminate that as a possible issue (say to 1000 ns)

Yeah, you could be right, that the problem is outside LinuxCNC.
My though was it´s linuxCNC (config) because of I copied the Z Axis by hand in the Ini-File (X and Y which are working are created in PNC-Conf) AND moving the Z Axis DOWN there seems to be no difference in way of moving.

I also have on X, Y and Z the same Servo Drive (Kollmorgen S346 with Firmware 3.89). Maybe the drives have different Hardware configuration - I will recheck. Motors are slightly different on X/Y and Z. Motors have Multiturn-Encoders connected to the Drive and they are in closed loop (drive and motor). So if something should be wrong it only can only be the step / dir signal between Mesa 7i76e and Servo Drive.

Servo Drive is rated at 1,5 MHz frequence for the step / dir interface. Cable is 30 cm.
I will recalc everything and give it a try on the weekend to change the timing for test, even if this timing is working fine on X and Y Axis...

I also noticed the resistor in the rs485 line - at the moment I havn´t insert this end-resistor. Could be an issue too.

Today I increased the step timing from 0,5 µS to 1µs each (high and low) but there was no change.

Then I was thinking again about that down seems to work (maybe only better) and up is more crazy.... So I recheckt my cabeling from Mesa to Drive - specially the DIR signal.

Here there was a missing contakt on DIR - (I think I crimped without stripping or someting like this). So crimped again and everything seems to work fine! Because of the floating DIR- the step count was sometime Up and sometime Down...

Thank you guys!