Intermittant diagonal layer shift, how to diagnose?

Quote
the_digital_dentist
I think there's something wrong with the Z axis if the nozzle is banging into the print hard enough to cause layers to shift.

Indeed. The two lead screws are linked by a belt loop to a single stepper, the same type as the X/Y steppers, so plenty of power and the bed is not heavy.
But the thought that plagues me is, why? Why did everything work great for two weeks, and then, "gronk!"? I have been looking in the wrong place for two weeks though, maybe it will become obvious soon.

DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

Quote
the_digital_dentist
I think there's something wrong with the Z axis if the nozzle is banging into the print hard enough to cause layers to shift.

if the Z axis does not drop and the head hits the model it definitely could shift the XY axis. Mine has hit the plastic for other reasons but never moved the XY axis.
It may have something to do with the torque of your motors. I can't remember what you said they were, but when I read it I thought that mine were twice what yours are.

Edited 2 time(s). Last edit at 06/20/2020 11:56PM by ruggb.

If you think the issue may be that the Z axis isn't moving enough, have you tried reducing Z axis acceleration?

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dc42
If you think the issue may be that the Z axis isn't moving enough, have you tried reducing Z axis acceleration?

That is one thing that I am going to try. However, while watching the hot end bang on things as it moved I noticed that my Z carriage was not "flat". The bed was adjusted to be flat but there was a 2mm difference in the carriage from left to right. I balanced the carriage and leveled the bed again. I wonder if it being off that far would mean that one side would bind or move differently making an instantaneous height where the two sides would be different and cause a head strike. So we shall see.

Also, I swapped the X and Y steppers and now the shift is toward the other stepper, so clearly, one of the steppers is stronger than the other for some mechanical or electrical reason. As long as motion is not blocked, this should not matter much.

EDIT
Wow. After all that, on the second layer of the next test, when I wasn't looking, I heard a loud "bzzzt" kind of grindy, buzzy sound and the layer shifted a huge amount straight towards X0, no diagonal at all. The Z had already moved and about half of the layer had already been printed when this happened. While the motors were all powered I tried moving the X/Y carriage, and I could move the (now) Y motor, but not the X. It looks like I may have found my problem. I should not have been able to move that stepper pulley with a court order, but it moved when I turned it with my fingers, that shouldn't be. Since my shift moved when I swapped the steppers, that means that bind was not in either cable circuit, it is in the motor. I have a new one from the seller coming, I just don't know when. I may just go buy one of these, they are only about $12. Or, hmm, move my Z motor to Y and get another stepper from the junk box for Z, it does not have to match the other two.

EDIT 2
I moved the Z stepper to Y, replaced the Z with another decently powerful stepper and started another test. Well, everything moves more freely now, and it took longer before the shift occurred, but it still happened. Time to attach a camera or something and see if I can actually catch what occurs. I keep finding little things to fix, and they have all made the mechanism better, but still I get the shifts. I am going to let the whole print run and see how things pan out. See if the final "home" homes both X and Y properly.

DLC

Edited 2 time(s). Last edit at 06/21/2020 02:49PM by dlc60.

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

Well, weeks of effort on fixing this problem has gotten me, pretty much nowhere:
Swapped ports for both X and Y - no effect
Swapped X and Y steppers (changed the direction of the shift, and found one stepper that was weak and replaced it - didn't fix the problem)
Raised the current to 1.2 Amp (no effect), dropped it to .4 Amp(too low to run at all)
Changed acceleration, Jerk and even Stealth chop on all steppers - no effect.
Corrected bed plate level, then re-leveled the hot bed - no effect.
Oiled the linear bearings and the lead screws - no effect
Replaced the belts - nada
Replaced toothed idlers with smooth ones - runs quieter, but still shifts.
The X/Y carriage moves very smoothy in all directions and feels the same for all movements. In general I think that I have made the whole printer better, but it still will not print without the layer shift.

So, finally, and perhaps I should have started here, I slowed everything down by half. And I got one of the models that had shifted before to print fully. Not the biggest one, but a middle large model. Dropping my acceleration and jerk down to 1/4 and 1/3 respectively, did nothing, but dropping the speed down seemed to help. Again though, the first two weeks of this printer's life, it worked great at 60mm/s. Then it didn't and it didn't at 80mm/s (base speed), but at 40mm/s and 80mm/s travel speeds it seems to have started to work. Not optimal, but I can work with it. As long as it stays reliable! We will see after I do a few more prints.

DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

Mine works reliably at 60mm/s print speed but if I increase it 10% the steppers begin to overpower the belts and it slips.
Manual X or Y movement feels real tight because I am moving 2 motors and fighting magnetism.
Moving on a diagonal is expectedly about 1/2 that tight feeling since only 1 motor is moving.
Others have indicated being able to print at much higher speeds, but I am puzzled as to why.
However, If I want a great print, 60 is too fast.

What does it mean that "overpowers the belt" and it slips? Is the belt slipping on the drive pulley or is the motor slipping - i.e. missing steps?
Neither of those should happen, especially not at such a low speed. Are the motor and pulley mounts solid? Adequate tension on the belts?
Excessive jerk speed or acceleration? High voltage (12V) motors? Low power supply voltage (12V)?

I have 64 oz-in motors (relatively small, low torque motors) in XY and have no problems with layer shifts or anything slipping and can drive the mechanism at over 200 mm/sec even with acceleration set to 10k.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
What does it mean that "overpowers the belt" and it slips? Is the belt slipping on the drive pulley or is the motor slipping - i.e. missing steps?
Neither of those should happen, especially not at such a low speed. Are the motor and pulley mounts solid? Adequate tension on the belts?
Excessive jerk speed or acceleration? High voltage (12V) motors? Low power supply voltage (12V)?

I have 64 oz-in motors (relatively small, low torque motors) in XY and have no problems with layer shifts or anything slipping and can drive the mechanism at over 200 mm/sec even with acceleration set to 10k.

I feel the same way, I should be able to go faster. In fact, I was running at 60mm/s using PETG initially. However, I am not an ME and I can probably improve my design to make the idler mounts more robust (that is my plan). So it is possible that my current hardware wore out(?) and I need to beef it up for reliability - that is nearly a given I should say...
There are holes in my knowledge though. How tight should the belts be? At what point am I adding friction without adding robustness?

DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

"Tight, but not too tight"- if they can slip on a drive pulley, the belts are definitely too loose. Usually, in a corexy printer, if the belts are so loose they slip on the drive pulleys, there's a problem with the belt positioning relative to the X and Y axis guide rails or linear guides. Be sure the pulleys are positioned so that the belts are parallel to the guide rails. When they are positioned properly the belt tension won't change when you move the extruder carriage around from one place to another.

You don't have to be an ME. Just use what you know- thin things flex, thick things don't- and start from there. Pull on the belts, see what flexes, and fix those things first.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
What does it mean that "overpowers the belt" and it slips? Is the belt slipping on the drive pulley or is the motor slipping - i.e. missing steps?
Neither of those should happen, especially not at such a low speed. Are the motor and pulley mounts solid? Adequate tension on the belts?
Excessive jerk speed or acceleration? High voltage (12V) motors? Low power supply voltage (12V)?

I have 64 oz-in motors (relatively small, low torque motors) in XY and have no problems with layer shifts or anything slipping and can drive the mechanism at over 200 mm/sec even with acceleration set to 10k.

"Overpower" = The motor can drive it but the belt can't move that fast so the motor pulley skips over a few belt cogs.
The belt is about as tight as I can get them
The belts are very parallel. The mounts are pretty solid for the motors and pulleys.
The idlers are simple bearings and I think they are all turning freely, but I can't see a few of them.
I might try lowering the accel and see if it skips then. It is currently set at 9000 max.
PSU voltage is 12V with an overkill of 40A. From when I had a 12V hotbed, which is now 120V.

Quote
the_digital_dentist
"Tight, but not too tight"- if they can slip on a drive pulley, the belts are definitely too loose. Usually, in a corexy printer, if the belts are so loose they slip on the drive pulleys, there's a problem with the belt positioning relative to the X and Y axis guide rails or linear guides. Be sure the pulleys are positioned so that the belts are parallel to the guide rails. When they are positioned properly the belt tension won't change when you move the extruder carriage around from one place to another.

You don't have to be an ME. Just use what you know- thin things flex, thick things don't- and start from there. Pull on the belts, see what flexes, and fix those things first.

Makes sense. I note on this unit that the belts move differently in one direction than the other on the idlers, looks kind of like they are being pushed rather than pulled. I take it from what you are saying that means that one of the belts is tighter than the other so I need to balance.

As for the "if it bends, make it thicker" mantra. Totally makes sense.

BTW, at my slower speeds, all the models are working. I will look into the belts and bending next.

Thanks all,
DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

Maybe the range of adjustment for the belt tensioners is too small. It should be impossible for the motors to apply enough force to get the drive pulley to skip on the belt teeth.
Post a picture of the printer showing the XY belts from above, and maybe a few close-ups of the idler pulleys, motor mounts, and belt tensioners.
If you manually move the extruder carriage around, does the belt tension remain constant?

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
Maybe the range of adjustment for the belt tensioners is too small. It should be impossible for the motors to apply enough force to get the drive pulley to skip on the belt teeth.
Post a picture of the printer showing the XY belts from above, and maybe a few close-ups of the idler pulleys, motor mounts, and belt tensioners.
If you manually move the extruder carriage around, does the belt tension remain constant?

It does now, there were "catches" in it before, but I refactored one of my carriage plates (what are those called anyway?) and found a binding idler, then re-tensioned the whole lot and it is the same in all similar directions (diagonals are obviously easier to move than the straight lines.)

I dropped the speed and the especially the jerk and acceleration and things are reliable, just not as fast as I would like. I also changed how I secure the belts to the end effector. Initially I used the same clamp topology on each end, but that meant zip tying smooth-to-smooth on half of them. I think those slipped, so I created a set that were 3mm "shorter" so that I could run the belts through the other way and zip tie them tooth-to-tooth for a more secure connection. The combination of all those things smoothed out the motion and eliminated what looked like "pushing" a belt on one idler when going a certain direction.

I can probably speed it up now, but I need to get some stuff out, so I'll wait until I refactor the carriage sliders and widen the machine 100mm to change those values.

DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

I noticed that de42 had suggested dropping your z acceleration speed and am wondering if you had tried that and by how much? I had a bit of an issue with the Tronxy x5sa pro when I increased the overall acceleration speed (M8008) from 100 to 700 and the Jerk (M8007) from 20 to 10 and that resulted in a total stop when the head went from front left to rear right and the z started. The fix for this eventually was to bring the maximum speed of the Z (M8013) from 20 down to 5. That worked and the prints are now smoother (Less jerky)and faster and as yet I have not seen any degradation in the prints.
I am not by any means very knowledgeable with this stuff but I reasoned out that the Z does not need all that current draw at the same time as all the others are sucking it up and the board seems to agree with me so far. If you are doing anything on your Tronxy like this you need to do your extruder speeds at the same rate as the others, (8007 I10 and 8007 E10 and 8008 I700 and 8008 E700. I hope this may be of some help. Regards Brian

Quote
RockyAussie
I noticed that de42 had suggested dropping your z acceleration speed and am wondering if you had tried that and by how much? I had a bit of an issue with the Tronxy x5sa pro when I increased the overall acceleration speed (M8008) from 100 to 700 and the Jerk (M8007) from 20 to 10 and that resulted in a total stop when the head went from front left to rear right and the z started. The fix for this eventually was to bring the maximum speed of the Z (M8013) from 20 down to 5. That worked and the prints are now smoother (Less jerky)and faster and as yet I have not seen any degradation in the prints.
I am not by any means very knowledgeable with this stuff but I reasoned out that the Z does not need all that current draw at the same time as all the others are sucking it up and the board seems to agree with me so far. If you are doing anything on your Tronxy like this you need to do your extruder speeds at the same rate as the others, (8007 I10 and 8007 E10 and 8008 I700 and 8008 E700. I hope this may be of some help. Regards Brian

Dropped accel, dropped jerk, increased stepper current. Only slowing the whole thing down does any good. Even then, occasionally I get the binding and the layer shift. I am totally out of ideas.
I can avoid the layer shift on a long piece if I orient it along the Y axis direction, along the X axis direction, I'll occasionally get the shift. If I have a model that is large in both axis, then I am out of luck until unless I really slow the printer down. Kind of frustrating, the only clue that I have is the potential for a not perfectly parallel belt path along the X axis. It is really darn close, as close as I have on my TronXY x5sa CoreXY, which does not have a layer shift issue.

I am quizzing the Exoslide guy about this, I am using Exoslides for my X and Y axis.

DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

To Quote - I am quizzing the Exoslide guy about this, I am using Exoslides for my X and Y axis
I have looked at the design of these on the tube over and over and I have to say I would be quite concerned about having so many rollers continually running along soft and sometimes dirty aluminium. I have seen some applications where running 3 rollers has worked better than 4 let alone 16 as shown in their intro video. That is a LOT of surface area waiting for the rolllers to pick up any little nick or dust particles. Just another observation over time to consider.....many people try and make the top as solid and inflexible as possible and I have noticed way less problems when the hot end is able to move/flex upward just enough to not get knocked out of line when it hits some cold parts of the print. It generally all smooths out fine in the end.

---

Brian

If the nozzle bangs into cold parts of the print, either the Z axis isn't working properly and needs to be fixed, or the print is curling up and that needs to be fixed or both. It is not normal for the nozzle to bang into the print, and if it does, the resulting print will be trash regardless of whether it is a Z axis problem or a print curling problem.

If the print curls, you need to adjust temperatures, try the print cooling fan (depending on the type of filament), print slower, or print a sacrificial object so the main print has a chance to cool a bit before the hot nozzle comes along and dumps more hot plastic on top of the previous layer.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
RockyAussie
To Quote - I am quizzing the Exoslide guy about this, I am using Exoslides for my X and Y axis
I have looked at the design of these on the tube over and over and I have to say I would be quite concerned about having so many rollers continually running along soft and sometimes dirty aluminium. I have seen some applications where running 3 rollers has worked better than 4 let alone 16 as shown in their intro video. That is a LOT of surface area waiting for the rolllers to pick up any little nick or dust particles. Just another observation over time to consider.....many people try and make the top as solid and inflexible as possible and I have noticed way less problems when the hot end is able to move/flex upward just enough to not get knocked out of line when it hits some cold parts of the print. It generally all smooths out fine in the end.

I have watched it happen, the nozzle isn't colliding with the print. "Something" else is happening, and since the belt is skipping, my bet is on a bind because it happens in my X axis, which is the axis that spans the printer rather than the Y which is just one carriage going back and forth. BUT, my belts are parallel to their travel - I keep wondering if I do not have a perfect tension/alignment, I am just not sure how to balance the two belts and KNOW that they are balanced. If you get my meaning. I have another CoreXY that does not do this, but it isn't using Exoslides either, just a"V-belt wheels on bars" sorta linear slide (TronXY x5sa Pro).

DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

The way you know if the belt tensions are correct relative to one another is that the X axis is square with the Y axis. The absolute tension is another matter- it should be tight but not too tight. In my printer the belt tension is adjusted by sliding the motor mounts back so I get a direct feel for how hard I am pulling on them. If your belt tension adjustment uses a screw it might be harder to get a feel for it. Try plucking a long span of belt - it should ring at a very low frequency. If you suspect the mechanism is binding, try loosening the belts a bit.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
The way you know if the belt tensions are correct relative to one another is that the X axis is square with the Y axis. The absolute tension is another matter- it should be tight but not too tight. In my printer the belt tension is adjusted by sliding the motor mounts back so I get a direct feel for how hard I am pulling on them. If your belt tension adjustment uses a screw it might be harder to get a feel for it. Try plucking a long span of belt - it should ring at a very low frequency. If you suspect the mechanism is binding, try loosening the belts a bit.

Hmm. Those are good ideas. Figuring out if the entire rig is perfectly square is a challenge. I am a programmer, not an ME. What is a good way of determining this? Can it be done without a fancy tool?

Thanks,
DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

On my Tronxy I put the hot end into the middle first as that gets all of the belts positioned equally and then on the front belts I pluck them like a guitar. By ear you should be able to hear if there is much of a difference and if so tighten or loosen a tooth at a time untill they sound right. There is an app called spectroid that can be installed on a mobile phone if you want to have a look into that. It shows you a graphical representation of frequencies.

Quote
dlc60

Hmm. Those are good ideas. Figuring out if the entire rig is perfectly square is a challenge. I am a programmer, not an ME. What is a good way of determining this? Can it be done without a fancy tool?

Thanks,
DLC

I am trying to figure out how mine can be anything but square and I can't see it. However, If it is far enough off it could cause binding in the movements. Also if you print a square, it will be a parallelogram but not square. A perfect square will have equal diagonals. So print a square in a few lines/layers at the limits of your bed and measure those.

Edited 1 time(s). Last edit at 09/16/2020 10:10AM by dustinoff.

If the frame is square, and the Y axis rails are parallel to the frame rails, all you need is a ruler to check if the axes are square- move the X axis to one end of the Y axis and measure the distance from the end of the X axis to the edge of the frame at both ends of the X axis. if it's square the distances will be the same. Verify by printing a rectangular object and measure diagonals.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
If the frame is square, and the Y axis rails are parallel to the frame rails, all you need is a ruler to check if the axes are square- move the X axis to one end of the Y axis and measure the distance from the end of the X axis to the edge of the frame at both ends of the X axis. if it's square the distances will be the same. Verify by printing a rectangular object and measure diagonals.

Beg to differ with you on that DD. Unless I don't understand what you are saying, high possibility, you are measuring 2 sides of a square. If you tip the square into a parallelogram those two sides will still be = but the diagonals will not and therefore the square is no longer square. Printing a rectangle or square and measuring the diagonals definitely works.

If the frame is square it works fine. if your frame isn't, all bets are off.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
If the frame is square it works fine. if your frame isn't, all bets are off.

Well, isn't the question about how to determine squareness. The only way to do that, other than a lab instrument to measure 90°, is to measure the diagonals. But on printer hardware, measuring two diagonals accurately seems to me to be pretty difficult. The easiest thing one can do is to print a big square/rectangle and measure its diagonals. Then you can relate that back to your hardware. Otherwise, you will just be hand grenade close.

Quote
dustinoff

Quote
the_digital_dentist
If the frame is square it works fine. if your frame isn't, all bets are off.

Well, isn't the question about how to determine squareness. The only way to do that, other than a lab instrument to measure 90°, is to measure the diagonals. But on printer hardware, measuring two diagonals accurately seems to me to be pretty difficult. The easiest thing one can do is to print a big square/rectangle and measure its diagonals. Then you can relate that back to your hardware. Otherwise, you will just be hand grenade close.

I printed a 40mm cube, and it was square, and very close to 40mm on each axis.
I moved the carriage close to one end and measured the distance from the slider edge to the end of the printer. There is a 2mm difference. A triangle check tells me that my printer is as square and I can tell from a triangle, in all directions. I spent an hour or so tweaking belts to change that 2mm difference, and was unable to budge it. Moving the carriage back and forth at a variety of speeds gives me no binding. Moving the carriage by pulling on the belts at various speeds didn't bind up either. Grrr. Something is clearly NOT right, but I haven't a clue what. I really don't think it is the controller either, I have a cooling fan running across the electronics and the PCB temperature barely moves during a print. The Duet 2 Maestro is a pretty reliable controller, so I am doubting that I have a driver issue.
I get really good prints from this printer as long as I keep the movement speeds down.
It isn't as fast as I would like, I have plenty of extruder margin in that regard, I guess that I should just be happy that it works well, even if not as fast - I can always send the job to my Delta printer, that will print plenty fast...

DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

Quote
dlc60
I printed a 40mm cube, and it was square, and very close to 40mm on each axis.

DLC

measuring each side does not tell you it is square. You must measure the diagonals.

Quote
dustinoff

Quote
dlc60
I printed a 40mm cube, and it was square, and very close to 40mm on each axis.

DLC

measuring each side does not tell you it is square. You must measure the diagonals.

"square, and very close to 40mm on each axis"
means it is square (56.56mm on the diagonal, pretty much dead on) AND 40mm on a side.
Commas matter.

DLC

---

Kits: Folgertech Kossel 2020 upgraded E3Dv6, Anet A8 upgraded E3Dv6, Tevo Tarantula enhanced parts and dual-head, TronXY X5SA Pro(E3DHemera).
Scratch: Large bed Cartesian, exchangeable heads, Linear slide Delta, Maker-Beam XL Micro Delta, 220x220CoreXY.

Quote
dustinoff

Quote
the_digital_dentist
If the frame is square it works fine. if your frame isn't, all bets are off.

Well, isn't the question about how to determine squareness. The only way to do that, other than a lab instrument to measure 90°, is to measure the diagonals. But on printer hardware, measuring two diagonals accurately seems to me to be pretty difficult. The easiest thing one can do is to print a big square/rectangle and measure its diagonals. Then you can relate that back to your hardware. Otherwise, you will just be hand grenade close.

You can always check using a machinist's or carpenter's square. It's usually pretty easy to measure the frame's diagonals, too.

If the ends of the frame pieces are cut square and the pieces are matching lengths, when you bolt them together they will be square because they can't be anything else. OTOH, if you cut the frame members with a hacksaw and built the frame using a bunch of corner plates to hopefully hold everything square, you never know what you'll end up with. You can check squareness of the cuts by standing the frame pieces up on a flat surface next to each other. If they aren't parallel, the ends aren't square. It will be obvious if they aren't matched in length, too.

It's usually pretty easy to line the Y axis guide rails up parallel to the frame members they attach to by using spacer(s). Line one up first, using a small metal spacer to set the distance between the edge of the rail and the edge of the frame member (don't use a series of printed plastic parts- their dimensions are well controlled and they will flex) and bolt it down tightly and it will be the reference. Align the other rail parallel to the reference by bolting a spacer to one of the bearing blocks. Move the spacer bar to one end of the second rail, lightly screw down that end of the rail, then move the spacer to the other end of the rail and screw it down. Finally, screw the first one down tight and run the spacer bar back and forth to check it. Now the frame is square and the Y axis rails are parallel to the frame members and each other.

Building a square frame makes alignment of guide rails easy compared to trying to align them on an unsquare frame...

DLC60, if it's printing square you're done. The X and Y axes are square. The error may be in the frame which doesn't really matter.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]