Fine Homing on X doesn't work

Phaedrux

And it's definitely not a grub screw slipping on the motor pulley?

Stevtommo

Apologies I meant to add to this earlier. I've checked all the grubscrews, and they're rock-solid.

I'm frankly amazed how much torque the NEMA17's have in this COREXY configuration. I've tried to stall them by holding onto the carriage but they pull right through.

Phaedrux

Can you show some photos of the printer and/or a video of the homing in action?

deckingman

@Stevtommo said in Fine Homing on X doesn't work:

,,................I've been watching the 'object model' to see if/when the stops get triggered, and they (X,Y) DO for the first move, but only the Y gets triggered for the second.

That could be significant. If the G1 H1 X-n move gets executed, what should happen is that the carriage moves until the endstop triggers. If the endstop never triggers, then the carriage would continue to move after it had crashed into the frame. So logically, the only possible explanation for the endstop not trigggering and carriage not crashing into the frame, is that the G1 H1 X-n move does not get executed. Suggest you delete that G1 H1 X-n move and re-type it in case there is a hidden character which is screwing things up. Or you could add M400 after the Y axis move which would wait for that move to complete before doing the X axis move (although that shouldn't be necessary).

Stevtommo

@Phaedrux , good idea I've hosted them here, along with videos showing some wierd behaviour - more on that below... Google Drive

@deckingman - 100% agree. Why does executing the same command not elicit the same response? I 'feel like' the machine is holding onto a guess of where 0,0 is after rough homing and is trying to predict it rather than use the stops.

Retyping the line didn't make a difference, nor did the use of M400 between commands.

What I discovered this evening that does seem to 'work' is the following, and I can't make heads or tails of it...

Rough pass X,Y similtaneously (Always triggers the stops):

G1 H1 X-425 Y-365 F1800 ; move quickly to X or Y endstop and stop there (first pass)
G1 H1 X-425             ; home X axis
G1 H1 Y-365             ; home Y axis

Trying to fine tune the X position (Always fails):

G1 H1 X-100 F360        ; move slowly to X axis endstop once more (second pass)

Going for a Y home (Always triggers):

G1 H1 Y-365             ; then move slowly to Y axis endstop

Then doing X again (mostly triggers):

G1 H1 X-100 F360        ; move slowly to X axis endstop once more (second pass)

Doing X again-again (99% sure this is capturing all)

G1 H1 X-100 F360        ; move slowly to X axis endstop once more (second pass)

deckingman

@Stevtommo On a CoreXY, this command G1 H1 X-425 Y-365 F1800 will move both the X and Y axis until one or other end stop switch triggers (that's why it must always be followed by individual X and Y axis moves).

So it could be that the Y axis always triggers which stops further movement of both axes but the X axis does not trigger. i.e, it could be simply that there is something "flaky" with X axis endstop. You could test that theory by positioning the carriage such that it is a long way away from the Y end stop but fairly close to the X end stop, then doing that combined XY H1 move. If the X end stop doesn't trigger, then the carriage should hit the frame in the X direction and attempt to continue that move.

Do you have another microswitch that you could try?

Phaedrux

I don't see anything obvious from the videos.

Can you try this? Rename your current homeall.g to make a backup. Then create a new macro file and name it homeall.g. In there just have

G28 X
G28 Y
G28 Z

That will call your other homing macros.

Same behaviour?

If you watch the X axis endstop light on the board (if your board has the LED, not all do I think) does it light up when homing X?

A Former User

This post is deleted!

Stevtommo

@deckingman - good idea I did already swap in other X end-stop switches to no avail. Given the switches have a 100% success rate in catching the rough-passes, I can't see how it's anything electro-mechanical at this point.

To your other point, the system is failing to execute the latter X-move when both axes have just been successfully homed and positioned at X,Y:10,10. It fails when only one axis is being asked to move, namely

G1 H1 X-100 F360

So...what I'm finding is (and I fully appreciate that this looks odd) that if I repeat the request multiple times in a row, while splitting up the drive it's addressing, seems to be catching the error every time (I've done 30 home cycles and it's 'jerky' but it works.

G1 H1 X-100 F360
G1 H1 Y-100 F360
G1 H1 X-100 F360
G1 H1 Y-100 F360
G1 H1 X-100 F360
...

Stevtommo

Hey there @Herve_Smith,

Thanks for the suggestion, I tried this out. My printer fails exactly in the same place, the fine-homing move. It just stops somewhere along the axis (does not show activation of XStop in object-model, light on the board does not light up as the circuit remains closed, and it then merrily executes the next line...

G1 H1 X-455 F360     ; Slowly move to the X axis endstop once more (Fine Adjustment)

Stevtommo

@Phaedrux, same bad behavior... G28 X is the problem.

The below sometimes solves it, but it's not repeatable... Appreciate your thoughts.

Without lines 9,10,11 in there, the X-axis just refuses to engage the end-stop.

G91               ; relative positioning
G1 H2 Z5 F6000    ; lift Z relative to current position
G1 H1 X-325 F1800 ; move quickly to X axis endstop and stop there (first pass)

G1 X10 F6000       ; go back a few mm

G1 H1 X-325 F360  ;try fine-homing (fails)

G1 H1 Y-0.1 F360  ; jiggle the y-axis one way
G1 H1 X-325 F360  ; try to fine home X axis again (seems to always succeed)
G1 H1 Y0.1 F360  ;  jiggle the y-axis back

G1 H2 Z-5 F6000   ; lower Z again
G90               ; absolute positioning

deckingman

@Stevtommo God this is weird.... Clutching at straws here but is the carriage nice and free to move in the X direction? If your belts are too tight it can put too much side load on the motor bearings. Perhaps there is some sort of "sticktion" which a high feed rate can overcome but for some reason a slow feedrate cannot? Like I said, I'm clutching at straws but I note that your motor current is only 800mA which is on the low side. Are you able to increase that? If the motors are rated at say 1Amp, you could push the current up to say 950mA just to see if that helps. If it does, then maybe more powerful motors might be the answer and/or eliminate or reduce any "sticktion".
I might be useful to know if there is a sweet spot for feedrate for the fine homing move. You could try doing the slow homing at the same feedrate as the course homing, and if it homes reliably then progressively reduce it until you run into problems again.
I've no idea if this is significant but on a Core XY, both motors are employed for pure X or pure Y moves, but the direction of one motor is reversed. I can't think of any reason why that would be significant to your problem but I'll keep thinking about it.

A Former User

This post is deleted!

Phaedrux

Have you tried increasing the length of the second move from 100mm to at least the length of the axis?