Finished parts ending like they were offset some how...

Actually I think this BOB does not have optos in the PUL & DIR outputs?

Update:

I think it is solved, running the stepper test on the X axis after 3000 times going back and for, it went back to the same position.
Due that I used a second xstep pin for the second X axis motor I forgot to do a reset on that pin example:
set parport.0.pin-02-out-reset 1
Tomorrow I will do a cutting test and let you guys know.

Yes I saw that!, this is a small machine and it never presented any problem just using one switch for each axis. I will post my machining results ASAP. Best regards.

BigJohnT wrote:

Yea! glad you solved it.

I don't want to muddy the water but if you have a dual motor gantry LinuxCNC Master (yet to be released 2.8) can properly home a dual motor gantry using a home switch for each motor.

Here is an old video showing it.

JT

This is a very wrong way of homing a gantry. I cannot imagine the damage this would do to a real, rigid machine with proper servomotors. If this is how it will be released it needs to be fixed asap!

That video was merely a visual demonstration, purposely set up out of wack to show how it can square a gantry.
If a real gantry were so floppy that it would rest that far out of square, the squaring move would not have hurt it anyway.

Todd wrote:

That video was merely a visual demonstration, purposely set up out of wack to show how it can square a gantry.
If a real gantry were so floppy that it would rest that far out of square, the squaring move would not have hurt it anyway.

It is not a question of how floppy the gantry is, but where the home switches (or a zero marks on encoders) are. The video shows that a first motor is stopped when the second still searches for a home position - this should never be allowed. The offset for "squarness" should be applied only after both motors find their home position, until then they should both follow one another.

I am not good with describing this, but I can provide a very good instruction manual from a real industrial CNC controller on setting a gantry drive.

That video looks like exactly how to home a gantry with 2 home switches,
that is, all motions either leave the racking the same or reduce it.

Waiting till both home switches are activated seems dangerous to me
(Going an unknown distance beyond the first limit/home switch activation is likely a bad idea)

fixer wrote:

This is a very wrong way of homing a gantry. I cannot imagine the damage this would do to a real, rigid machine with proper servomotors. If this is how it will be released it needs to be fixed asap!

This works far better than you think in the real world. Furthermore, it allows you to square the gantry in software very quickly and precisely by changing offsets based on actual measurements.. Once initial squaring is complete, it is very rare to see the gantry to become misaligned but on a light weight plasma gantry where speed and acceleration is the order of the day, it is possible to push one side out of alignment a few inches just to demonstrate this and it squares up flawlessly. In fact, you could remove the connecting gantry rail and push the gantry ends to opposite ends of the table and they would align perfectly when homed.

To say this is dangerous and needs to be fixed displays a lack of understanding of how this works.

When properly configured, misaligned motors move in unison (so nothing becomes worse) until the first one hits the home switch when it stops and waits for the other side to catch up. After the second home switch is triggered, the motors move in unison at a much slower speed off the triggers and rehomed to precisely locate the home positions and then the relative offsets are applied to ensure it is square. (In my case thats < 1mm) . The final position on my machine is 45mm away from the home switch position to maximise the table size as the gantry needs room to decelerate to a stop when homing without hitting the limit switches.

I said "homed", not "home switches activated".

Waiting till both home switches are activated seems dangerous to me

This is exactly what is happening in the video!

I will try to illustrate:
Imagine two joints with incremental linear scales with ABZ signals and home switches. Say one is labeled jointX1 and the other jointX2. The same can be applied to a different homing procedure, for example with encoder or even homing stepper to a single switch.
homing procedure steps:
1. X1 starts to look for home switch trigger with [jointX1]HOME_SEARCH_VEL. X2 follows.
2. X1 hits home switch - X1 is deccelerated to stop - X2 follows to a stop
3. X1 starts to look for a index pulse from a scale with [jointX1]HOME_LATCH_VEL. -X2 follows
4. X1 finds zero pulse and sets its counter value to a value [jointX1]HOME_OFFSET
5. X1 status is now homed, both motors decelerated to stop.
6. X2 now starts to look for a home switch with [jointX2]HOME_SEARCH_VEL - now X1 follows the motion
7. X2 hits home switch and is decelerated to stop - X1 follows
8. X2 looking for an index pulse with [jointX2]HOME_LATCH_VEL, X1 follows the motrion
9. X2 finds the index and sets its counter value to a value [jointX2]HOME_OFFSET from INI file.for X2
10. X1 and X2 are now homed and its counters should show the same values minus a little diference due to not being in square
11. X1 and X2 both move to a [AxisX]HOME and the gantry is now squared and axis X status is homed.

In step 10 there should be a limit in difference value - if to big difference should not continue to step 11. This is to prevent racking, this is also the only time the racking can occur if something goes wrong (defective switch, scale or wrong INI setting)

So first time you try to home a gantry, you will be stopped in step 10. Now you should modify INI setting [jointX2]HOME_OFFSET so the encoder values match. After that you home again and check for squareness. Modify [jointX2]HOME_OFFSET again to compensate the squareness error and home again.


Please forgive my bad English, I am not a native speaker... I hope you understand what I am saying above. Please note that I am writing from experience, I work professionally with different controllers, I am in CNC service business.