Page MenuHomeAleph Objects Inc

Y to frame squaring jig
Closed, ResolvedPublic

Description

We need to have a way to accurately square the Y to the frame. If the y isnt square, the xy calibration will be off resulting in a messed up dual print

Event Timeline

west created this task.Oct 3 2018, 8:54 AM
kent claimed this task.Oct 3 2018, 12:58 PM
kent added subscribers: logan, kent.

@logan is out the rest of the week so I will get this started. @west what is the tolerance on the squareness spec and what is the best way to measure this critical dimension?

west added a comment.Oct 4 2018, 4:10 PM

@kent I'm not sure. We would have to look into it

logan added a comment.Oct 8 2018, 10:40 AM

@west @kent Do we have a method in mind for determining that tolerance? Would it involve setting the Y-axis at varying degrees and comparing print results?
When you say the X/Y calibration will be off, is that affecting the results of data collected from the X/Y calibration cube? My understanding was the cube was being used to identify backlash as well as nozzle positions, is that correct? If so, is there a data output from that calibration that could be used to compare different degrees of variance from square?

west added a comment.Oct 8 2018, 11:00 AM

@ logan, we could go with that. Start with a square machine and get the calibration info then start moving it out of square until we see a notable error

@logan @kent @west i think just measuring the distance from the frame on both extrusions of the y axis will suffice. 1 degree out of square over 300mm is about 5.25mm. which none of our beds are really ever that crooked since that would be very noticeable. Measuring this as an angle will leave us dealing with with very very small angles. 1 degree over 10mm (length of cal cube) is 0.35mm.

I think we should just measure from the frame and put some allowable tolerance on that.

west closed this task as Resolved.Oct 8 2018, 3:38 PM

I agree with @tutley, as long as the spacing is equal on the front and back, it should be square enough for proper calibration

logan reopened this task as Open.Oct 8 2018, 3:42 PM

@west Re-opening, this ticket is for jig creation, right? The tolerance was needed to determine jig dimensions.

west added a comment.Oct 8 2018, 3:43 PM

I guess the jig that needs to be created is a spacer from the vertical extrusion of the frame to the Y extrusion

logan added a comment.Oct 9 2018, 7:35 AM

And spacing front to back, between Y-corner and bed mount chassis

logan added a comment.Oct 9 2018, 9:33 AM

@west could you post those measurements here? or I can come up and find out, if you've got an assembled unit; ours is in pieces.

west added a comment.Oct 9 2018, 11:01 AM

@logan do you have a preference on the side of frame being measured form (left side, right side)

logan added a comment.Oct 9 2018, 3:41 PM

@west Off the top of my head I think the right side might be the most easily accessible. Both left/right for the front and back measurements should be the same

nickp triaged this task as Normal priority.Oct 24 2018, 9:27 AM
nickp changed the edit policy from "All Users" to "LulzBot Hardware Products (Project)".Oct 25 2018, 12:02 PM
logan claimed this task.Dec 7 2018, 7:59 AM

@west I pulled dimensions from an assembly that was in the repo previously but it sounds like those weren't right. Can you please post the needed dimensions here so these jigs can be made and tested?

west added a comment.Dec 10 2018, 8:15 AM

I measure 175mm from the inside of the left front extrusion to the left face of the left extrusion on the Y

logan added a comment.Dec 10 2018, 8:43 AM

Could you measure inside left front extrusion to left bed mount chassis?
Also rear Y-corner to bed mount table?

west added a comment.Dec 10 2018, 9:03 AM

The first measures 145
the second measures 109.5

Spacing jigs pushed to T3842 branch
Printing for testing now

@west Those aren't right. Misses the cal cube entirely. This is why it would've been helpful to update the assembly, rather than remove it. (T4505)

logan raised the priority of this task from Normal to High.Dec 10 2018, 1:14 PM

@west @tutley @kent
For these measurements I get:
left 123.95
rear 110.77

Please confirm

@logan the rear looks correct but im getting about 143ish on the left

@logan im noticing that this closely matches what @west measured above. Maybe im not measuring from the correct spot?

logan added a comment.Dec 10 2018, 1:53 PM


The above file contains two sketches showing placement of the Y-axis, based off entirely off of CAD models and the dimensions given to me by @tutley earlier. Those dimensions are as follows:
176mm from inside of right frame extrusion to right side of right Y-Axis frame extrusion
199mm from front of Y-Idler end plate to front of bottom front horizontal frame extrusion.
"Sketch" is along the X/Z plane
"Sketch001" is along the Y/Z plane
All other dimensions were pulled from current state of master branch

logan added a comment.Dec 10 2018, 1:54 PM

This resulted in the measurements given above:
Left 123.95
Rear: 110.77

west added a comment.Dec 10 2018, 2:08 PM

@tutley has your Y been adjusted to the new position?

west added a comment.Dec 10 2018, 2:10 PM

@logan you need to be measuring the x position based off of the left side of the frame since the toolhead homes on the left side

logan added a comment.Dec 10 2018, 2:12 PM

@west That is the dimension I have been asking for this whole time. I was given other dimensions or wrong dimensions.

west added a comment.EditedDec 10 2018, 2:20 PM

@logan the dimensions I gave you were from the right face of the left frame to the left face of the chassis mount. If you have the correct FW and measure from the correct position then it will work. I stand by my measurements.
The image shows it at 165 but that is because I am measuring form the outside face of the extrusion, I subtract 20mm due to the thickness of the extrusion 165-20=145

The first measures 145

logan added a comment.Dec 10 2018, 2:23 PM

I will flash FW again to double check but I flashed this unit this morning

logan added a comment.Dec 10 2018, 2:27 PM

@west maybe I flashed from the wrong directory or something? Apologies, it appears to be working now. But that means the measurements we had on our board were wrong, @tutley might want to double check and make sure your's is updated to reflect Josh's measurements

logan added a comment.Dec 10 2018, 2:29 PM

@kent jigs passed testing, ready for review

west added a subscriber: marcio.Dec 10 2018, 3:34 PM

For the Y @marcio has his machine at 117mm and the newest FW is .36. Those two things put the calibration in position

logan added a comment.Dec 10 2018, 3:34 PM

Well idk now actually, could be something up with FW, or Cura (start gcode), or the measurements given because it misses with the probe point or the cal cube.
@west @tutley @marcio Let me know what y'all figure out, thanks!

west added a comment.Dec 10 2018, 3:35 PM

@logan see above. we posted at the same time. I think that will solve your issues

logan added a comment.Dec 10 2018, 3:40 PM

@west still landing to the left of the cube with spacing set to 145 from left frame to bed mount chassis

west added a comment.Dec 10 2018, 3:41 PM

What FW version?

logan added a comment.Dec 10 2018, 3:45 PM

hits wiper pad but towards the right/
gets first probe point, misses second
rewipe misses to the left of the pad
machine crashes into itself
print stops

logan added a comment.Dec 10 2018, 3:46 PM

next rewipe missed on the right

karrad added a subscriber: karrad.Dec 10 2018, 3:48 PM

@logan @west We did have an update to start gcode this morning on T1868, it may be affecting your results

logan added a comment.Dec 10 2018, 3:49 PM

So far it misses on bits written into FW, calibration cube and G29

The latest FW is hitting the top of the calibration cube correctly during nozzle offset calibration.
@karrad do you have an ETA on when the latest changes to start gcode will make it to CuraLE testing repo? If not, could you link us to the most recent start and end gcode for quiver?

@logan They will not be in the testing repo for a few days, we are currently merging in upstream 3.6

You can run the master build locally in order to get the latest changes

7ff1ba93babe contains the latest start gcode; for reference

@karrad Latest start gcode hits the wiper pad in the center for both nozzles so that is good
But I am still unable to complete a probe sequence. It gets the first probe point and misses the second

logan added a comment.Dec 11 2018, 1:04 PM

Looks like I was getting some interference from the E2 harness; it must rest on the right side of the motor; not behind it. Otherwise it interferes with the X-End Idler at X-Max.
I'm getting the probe point now, but just barely, you can still hear the motor skip steps as it hits X-max.
Regardless I don't think this issue is related to this ticket.

In regards to the jigs, the jigs currently residing in the T3842 branch are correct and have passed testing.

@logan Ahh I am not sure on the gcode missing the probe. @marcio @west may be able to address that more directly

logan added a comment.Dec 12 2018, 7:21 AM

@karrad Possibly related to T4866 but still working towards a solution. Doesn't appear to be related to the positioning of the Y-Axis so far.

logan added a comment.Dec 12 2018, 7:55 AM

After discussing this with @west I was wondering about shifting the Y-Axis to the left a few mm. This would give us more contact with the right bed leveling washer at X-max travel and bring us a bit closer to center on the cal cube the first time. This would require a change to start gcode however.
Thoughts?

logan reassigned this task from logan to Steven.Dec 12 2018, 11:28 AM
logan reassigned this task from Steven to kent.Fri, Jan 4, 11:43 AM
logan added a subscriber: Steven.

The jigs existing in T3842 branch of taz-quiver measure 145mm from left and 109.5mm at rear. These were used for all of the alpha printers and is currently functioning well.
@kent As far as I am aware, these jigs are ready for review/merge into master

logan lowered the priority of this task from High to Normal.Fri, Jan 4, 11:43 AM
kent closed this task as Resolved.Wed, Jan 9, 10:57 AM

Looks good, merged.