M19 and spindle-index-enable issues

Hey!
I am trying to tune up the orientation of the spindle to work with the M19 command. Before that, the spindle-at-speed signal was configured and the G33.1 command was tested. It works fine. However, when the M19 is running, it is constantly in error with the position. Observing the spindle-index-enable signal with the HAL, I saw that when the M19 was running, it was not exposed. Therefore, the value on the encoder is not reset (for each revolution it now increases by 1). As I understand it, the value at the input to orient.0.position should be between 0 and 1, and due to the fact that there is no spindle-index-enable, the number is not reset to zero when the index mark is reached. Also, I checked by manually setting the HAL command "setp spindle-index-enable true" that the encoder will reset the counter. Pin spindle.0.orient is activated normally when processing M19. Therefore, I tried to generate the activation of the spindle-index-enable signal by the presence of the spindle.0.orient signal, but it could not be done correctly.
Please, help me:
1) Should really orient.0.position be between 0 and 1?
2) Should linuxcnc set spindle-index-enable when processing M19?
3) What else should you check?
4) Does the spindle driver need to be changed from "speed control" to "position control"?

Thank you in advance,
Max

Also, I using MESA 7i76 , current hal file attached.

Hi

About homing spindle, imo using C axis allow to use built in hal pin for starting homing using the regular process and work very fine.

For orient (without c axis) i have do something in my revised code allowing to home spindle automatically after chosen axis is homed or before first velocity start or before first M19 move.
The code take care to home allways with same speed and direction for "i think" better accuracy, but maybe is useless to take care of this.

Br

Sorry to bring back a few month old thread but I was just tossing around this idea for my current build.

Basically the spindle needs to do a full rotation at start up before an M19 will work correctly. Seen quite a lot of ways to tackle this issue. But two seemed to be the easiest to implement.

1. Typically for orient the spindle gets a sperate position loop. If this loop was setup as a C axis could it be included in the homing sequence? What would be the pitfall of doing it that way?

2. Since I have to do a remap of the toolchange procedure anyway can my first lines of the NGC code for it be:

o<toolchange> sub

M3 S200
G4 P1
M19 R0 Q10 P0 ; align the spindle

That's an interesting idea. But I can't wrap my mind around how to just do it the first time.

Can do something like homing the spindle when turning on the machine? (By running the spindle for a few seconds).
And to control whether the index mark is found after that - if it is not found - an error!

andypugh wrote:

O100 IF [#5000 EQ 0]
M2 S100
G4 P2
#5000 = 10
O100 ENDIIF

(Untested, I am on a train)