Biesse Rover 346 Retrofit

I think the most trustworthy thing to do would be to add a proximity switch to the machine. But I think this could be done in hal.
Find a hal pin that consistently gives the axis position, regardless of how the machine was homed or offsets...
I don't have a Linuxcnc pc near me right now to check the behavior of some pin candidates.
Then use another near, one input connected to the axis pin and the other setp to your trigger position.

Todd wrote:

I think the most trustworthy thing to do would be to add a proximity switch to the machine. But I think this could be done in hal.
Find a hal pin that consistently gives the axis position, regardless of how the machine was homed or offsets...
I don't have a Linuxcnc pc near me right now to check the behavior of some pin candidates.
Then use another near, one input connected to the axis pin and the other setp to your trigger position.

I like the idea of a prox sensor. I have some also..... I think I will implement that option.

I need two sections so when I do mirror jobs small enough to fit on one side. I can disable the sensors on one side so I can setup the job on the other side while the machine is cutting. If it crosses over that line while I am working on the table it
pause the program that is running. Then there are buttons with lights that tell me which ones I need to press that will resume if you are out of the sensor range. That is working now. I did it in classicladder.

So prox switch it is.

I'll use the rising edge one way and trailing edge the other so it will never get mixed up.

As to doing this in software, the pin axis.N.joint-pos-cmd, looks like the one to use, and should work very well.

Todd wrote:

As to doing this in software, the pin axis.N.joint-pos-cmd, looks like the one to use, and should work very well.

That's commanded position.....

Would this be better as this is actual position, so it is absolute and where the joint is actually.

axis.N.motor-pos-fb

axis.N.motor-pos-fb and axis.N.motor-pos-cmd, are kind of goofy. They can end up being odd numbers that vary from one Linuxcnc session to the next depending on where the machine was parked when it whas shut down and restarted and how the machine gets homed. They are set to 0 when Linuxcnc is first started and don't change for the entirety of that session. Unless you can guarantee the machine will always be in the same position when Linuxcnc is started or if you are using absolute encoders, I'd suggest not using them.

The axis.N.joint-pos-fb and axis.N.joint-pos-cmd, are absolute numbers and equivalent to the G53 position numbers on a Cartesian machine.

Todd wrote:

axis.N.motor-pos-fb and axis.N.motor-pos-cmd, are kind of goofy. They can end up being odd numbers that vary from one Linuxcnc session to the next depending on where the machine was parked when it whas shut down and restarted and how the machine gets homed. They are set to 0 when Linuxcnc is first started and don't change for the entirety of that session. Unless you can guarantee the machine will always be in the same position when Linuxcnc is started or if you are using absolute encoders, I'd suggest not using them.

The axis.N.joint-pos-fb and axis.N.joint-pos-cmd, are absolute numbers and equivalent to the G53 position numbers on a Cartesian machine.

In other words, when you home the machine, these do not get zeroed. Linuxcnc references the machine zero everytime you startup to whatever value these pins are. So if I pick an number, if the axis has moved when power was off my point will be different.

Thats unfortunate. There are ways to get this via software but I think I'd rather have a switch at the point of my reference....

Thanks for all your responses and help Todd. It is really appreciated.....

Test panel. Got tired of messing with one physical button. Now I have 14 and 14 output lites.

Cut the first pieces on the machine tonight. 64th of an inch off on the X @ 110", guess will have to fix that.... Y and Z are dead on.

Found some things while doing this process. The emer rope switch is defective. The program starts before the spindle is at speed.

All in all i am very pleased.

Thanks all for the help, especially Todd and JT and Tommy. Still more to do nut i have a small contract to do before EOB wednesday with plastic 2" thick.