Fanuc 0M P/S 90 alarm

[botha.y]

Hi all,

I have an issue with reference return for a Novar machine with a Fanuc 0M control.

Here is the quick summary of it:
- This machine has 3 axis: X, Y, Z
- X and Z zero out properly
- All software overtravel parameters are correctly set and work properly for all axis
- Y axis always gets a P/S 090 alarm for an incomplete reference return when zeroing the axis. It goes in the right direction and used to work.
**This has been happening since the Y-axis was actually stuck in one end. The motor had to be dismounted to manually unstuck the Y-axis and put back.
- The deceleration switch works properly and properly decelerates the Y-axis. The P/S 090 alarm happens exactly when the deceleration signal deactivates. For other axis, the machine keep going a certain small distance before zeroing.
- The machine has a Heidenhain LS 107 linear encoder for position and a Fanuc encoder in the back of the Y-axis motor for motor speed control.
- Speed (rapid and FL speeds) are the same for all axis, and all axis have the same type of connections (Heidenhain LS 107 linear encoder for position control with Fanuc encoder for speed control)

What I tried:
- I tried all positions for deceleration switch, no use.
- I tried all positions for magnet of linear encoder, no use
- I tested the reference switch of the linear encoder that should be activated by the magnet, which does switch properly with a magnet close to it.
- Parameters are as they were when it was working
- I tried switching the Fanuc card, same result

Now I know this is extremely specific, but any ideas would be appreciated.
If possible, if anyone knows the exact procedure to install this linear encoder for the first time? The linear encoder keeps it machine position and only resets its absolute and relative position values upon power off / power on. So how does Fanuc 0M control sets this machine position for the first time, the machine zero?

Thanks!

[underthetire]

Well, last know reference does not agree with the new reference position, in a nut shell. If the servo encoder was disconnected, the first thing you need to do is initialize the encoder. This can be done in MPG, you simply need to get the motor to spin at a relatively fast jog. Power down, hold CAN and P on power up, then zero all the axis. If it still doesn't work, you can then simply change the grid shift parameter by 1 and power down and back up and zero return.

[botha.y]

Hi,

Thanks for the reply. Actually I forgot to mention that indeed the machine zero was lost. I manually displaced the encoder of the linear encoder by unscrewing it (causing a servo excess error of course) and power off before putting it back in place to have the zero machine approximatively the same place as it used to be, plus or minus a little distance.

I tried setting parameter 509 GRDSY to 1 and restarted the machine, it did the same thing. I tried it in combo with 399 bit 3 set to 1, same result. I tried parameter 399 bit 7 to 1 by itself, same result.

I think grid shift is only used to displace the machine zero? There are no macros with machine coordinates like ATCs and the like in this machine, so as long as the reference would be done without that p/s 090 alarm it would work fine.

Thanks.

[botha.y]

The problem was solved.

Here was the solution of my problem:

The Heidenhain linear encoder has a small switch activated by the magnet. it has connections with the two wires. One of them was broken, so I soldered it on the old solder. There was still an issue, so I cut the other end (the input voltage wire) and soldered it directly on the end of the switch to which the board connects it to (which means instead of going through the board, went directly to the switch). And that worked.

Cheers.