Big Layer Shifting

[valid sequoia]

You can see in this picture a small shift only in the x then a big diagonal shift on the infill

[trim crown]

Is that the start of a 2x5 gridfinitiy box?

More relevant questions, what material are you printing in and is the print head dragging acroos the print making noise as it travels?

[valid sequoia]

Yeah it's some gridfinity boxes good eye

In this case it's ASA but I had it with ABS too

Haven't tried PLA yet

[trim crown]

"im new to core XY but i did notice when moving the tool head by hand in the x axis i get some tight spots that result in y movement. Is this normal? "
^^ I would say this is not normal.

[valid sequoia]

I figured as much. Seemed weird. I pushed the head to get access to the grub screws and felt it move in y despite me not putting any pressure on it

Doesn't move in x when I move y though

[trim crown]

I'm about ~1250 hours into my Voron 2.4 r2 350mm, so not an expert but layer shifts are probably the most common issue I've dealt with. My root causes so far:
-One linear rail had insufficient lube causing resistance and shifts
-PETG sticks to the Revo print nozzle and falls off in the print causing a blob that then repeatedly causes shifting
-Starting Z Offset too low

This printer seems very sensitive to the amount of material used so far.. I typically have to tune down the flow to help with layer heights, particularly on large flat objects.

I would try to fix that resistance you are feeling moving the print head first. Everything should move pretty smooth.
Also, I have heard it might be a good idea to disconnect the motors from the control board if you are going to be moving the head around a lot to avoid a short. I'm not sure how factual that is...

[valid sequoia]

I agree the resistance is the place I will start. I think I'll pull the belts and see if it's still hard to move in x. If it is i suspect it's the fault of the stupid lightweight x rail that came on it.

I bought it partially assembled so not my choice to use that rail

But maybe it's not flat causing the rail to bind on the carriage.

Steppers act as generators when moved so there's a chance they can damage things so I never move them fast or that far but honestly I've never had an issue with my other printers or my cnc router

[trim crown]

Just double checked on mine and the head moves smooth X and Y independently ignoring the motor resistance. Mine's a bone stock LDO Rev C.

[valid sequoia]

Okay then that's definitely a problem. Mine is a Frankensteins monster haha

Nothing stock or normal

Just loosened the belts all the way and it's still tight on x. Not as bad but still catches in spots

Also checked y again and this time with the belts slack it also moves x

[trim crown]

I want to say one potential cause of that movement you described could be the two X/Y belts that move the head not being the same length and tension.

going off memory from something I heard...

[hybrid island]

Maybe post photos of the belt routing all the way around, including on the back of the AB drives.

What you're describing where it's catching during x motion resulting in a Y move sounds a lot like a belt routing issue.

[valid sequoia]

Here's the front. The belts aren't tensioned here because I just loosened them.

The back will be an issue due to my umbilical. It will take some time for me to remove it without disconnecting everything. I didn't have much slack in the cables

[hybrid island]

Another common issue is these bolts here. They should be loose. It's the two farther forward ones on both xy joints.

[hybrid island]

Can you loosen it and pull the bottom of the panel away a few inches? Just enough to see the back of each drive.

[valid sequoia]

That's the plan

Here's the back

I wanted to re do the foam anyway so wrestled it off

[hybrid island]

Was the foam really thin? If so was the belt rubbing on the back panel? I don't see any foam or residue on the frame so is it on the panel?

[valid sequoia]

I'm going to load in some pla so I can print with the back off. It's on the panel

[hybrid island]

You'll also want to add a bit of tension and watch inside the AB drives as you move the toolhead around. Make sure the belts aren't climbing the pulley flanges at all.

[valid sequoia]

No it's not that thin. it's what the directions called for. I don't remember what it was but I bout it online.

I'll watch

Based on the fact the x was still a bit stiff even without belts tensioned I think there's a bigger issue with the xbeam . I want to replace that ultra light one anyway so I'm going to try and see if the new equal tension and adjusted pully screws help but if they don't I might table this until my replacement x beam comes in

[trim crown]

I watched this when building my gantry and setting up the belts:
https://www.youtube.com/watch?v=cOn6u9kXvy0

Seemed to have a lot of good test steps in there if you think its the X Beam and racking.

[hybrid island]

Yeah if the X movement alone is rough that's not a good sign.
Also I'm glad to hear you're replacing that x beam. Those wire frame ones cause a bunch of issues.

[valid sequoia]

That's all everyone tells me. Wasn't my choice since it came with the machine. In my testing it does work as is but cant use input shaping. The prints I've been getting off it have some ghosting but actually not that bad. I think if they just beefed it up some and stiffened it then it could work. But there's no point anyway since the bottle neck is my nozzle size not how fast I can move.

[valid sequoia]

Failed again so I guess I'll wait for my new x beam then see if I still have the problem.

[hybrid island]

What speed and accel were you running? Can you share your printer.cfg file

[valid sequoia]

basically left values stock. Just changed pins for NH and added what beacon needed

[hybrid island]

Can you share the gcode file you printed that caused the shifts?

[valid sequoia]

I can, it's just a gridfinity box. But the shifts have happened on other prints with different speeds and settings.

this is the asa print.

[hybrid island]

I was curious if maybe your slicer was setting speeds or accels too high for your motor current but it all looks fine to me. I guess we will see if the x axis swap changes anything.

[valid sequoia]

I removed the rail and beam already. without being attached the rail moves very smoothly so my guess is the xbeam is warped slightly causing the carrage to bind.

[valid sequoia]

okay i got the x beam swapped over. I tensioned the belts and it shifted again. put down two really good layers then i heard the steppers skip.

printing pla

i also re did my zoffset just to be sure thats not the problem.

seems like for no reason in the middle of printing the steppers are skipping steps. it didnt catch on anything or hit anything. It was just printing then they skipped and it shifted over then kept going like nothing happened.

i ran a bunch of quad levels and bed mesh to simulate it moving but it never skipped when doing that.

[hybrid island]

I just looked at the gcode file again and I see you are printing sparse infill at 500mm/s. I think klipper should be limiting that speed to 300 as per the config file but maybe there is something else bumping that up. What hotend setup do you have? Going back and looking there might be some extrusion issues here which can cause layer shifts. Maybe the speed is also just too high.

[valid sequoia]

i have a volcano

its shifing before it even gets to the infill now

i printed the same print diagnally and it shifted on the skirt

i can try it slower but i would be very unhappy if thats the problem its already slower than i want

[valid sequoia]

Dropped it down to 300 and it shifted again. Staring the third layer

[hybrid island]

Could you show me how tight your belts are? A video like this.

[valid sequoia]

i did a few more tests. I slackened the belts some. Re lubed the x rail and slowed it down some more. made it to the infill then it shifted. same as before, it was working fine then the steppers just skipped steps.

[hybrid island]

Would it be hard to catch a layer shift on video?

Those belts were definitely too tight. That sounded like 150hz to me.

Not sure if I've already shared this with you. #slicers_and_print_help message

[valid sequoia]

i loosened them some more aned still got the shift. I will try and film it tomorrow. shouldnt be that hard since it happens early on in the prints.

[valid sequoia]

looking at the list the only thing i noticed is i still get some movement of the y when i push in x manually and same in reverse. I was hoping that would go away with the new x beam.

[hybrid island]

Can you show that in a video? Does a y only move also cause X movement?

[valid sequoia]

Clearly moving in x as I pull on y

[hybrid island]

Ah ok. Thats clearly an issue. Looks like binding on the upper belt. You'll need to check every idler and bearing stack for resistance. Using a short length of belt and "flossing" each one independently might help.

[valid sequoia]

I'll have to order some belt or find a stand in. Since the belts came pre attached. I did have to remove them because they were routed wrong but I didn't change the length so I don't have any cut offs.

[valid sequoia]

How do you know it's the upper belt and not the lower?

[hybrid island]

Watch the motor pulleys when you do that to see which one spins and which one stays till.

[valid sequoia]

both move as far as i can tell. I checked all the pullys and non seemed to bind. I loosened the belts and spun them by hand.

I tightened the belts back up to the point they make the tool head barly move and im going to try printing. The belts are fairly loose now but i want to see what happenes. im filming the whole print

[hybrid island]

Maybe next we need more photos of the belt path. You can just spam the channel with them.

[valid sequoia]

Now I'm having issues with under extrusion. I didn't change anything related to that and I'm using the same roll of pla this whole time.

[hybrid island]

Take a look through this list. #slicers_and_print_help message

The reverse bowden is the first thing that stands out to me.

[valid sequoia]

Why would it work previously then just stop working? At this point I have a lot of hours on this machine without extruding being an issue

As you said there was a little under extrusion but I'd chock that up to tuning. I didn't really dial it in yet. Now hardly anything coming out

I think it's more likely my nozzle is the problem. I don't like it and it's been nothing but problems

[valid sequoia]

Well in this case your definitely correct. Idk why but it's way too tight. Wasn't like that before

[hybrid island]

Over time Fep reverse bowden can become sticky. Especially in warm chambers. That's why we recommend real PTFE which is alway milky white. It's also important to make sure it has a 3mm ID.

[valid sequoia]

I ran the filament in with no tube and the extruder is skipping too. Might need to go back through my toolhead. Might have multiple issues

[hybrid island]

There might be something else going on but running without a reverse bowden can also cause skipping.

[valid sequoia]

I'm not getting any skipping at 5mms feed rate so I'm going to try a print and see what happens. It's definitely better

Before I was up at 8mms feed rate so it's definitely not as good

[hybrid island]

Ok then yeah it does sound like something is off. Do go through that list and see if anything stands out.

A volcano should do well over 10mm/s linear feedrate.

[valid sequoia]

I never got that. So maybe somethings been wrong in my sb this whole time

I went with a volcano because the toolhead mount this came with is aluminum. With a standard nozzle it puts the beacon inside the metal keep out zone but the volcano pushes it down far enough. But I really don't like my nozzle. I still think it's an issue

[valid sequoia]

I caught it this time! You can hear the steppers when they skip. Then it just continues printing.

[hybrid island]

Oh that's a weird one. It skips in the middle of a path, not during an acceleration.

[valid sequoia]

Yup

[hybrid island]

Can you get another video of this? I'd like to see movement all the way from the back to the front and then back again. With a wide camera angle as in the most recent video.

We need to figure out that binding in the A belt path.

[valid sequoia]

You mean just get a video of it moving forwards and backwards? I edited that video to be shorter so I have the beginning of the print if you want to watch it move farther

[hybrid island]

I want to see you moving the gantry by hand again, as in the video I linked.

[valid sequoia]

Okay

[valid sequoia]

Interestingly it's now moving opposite. When I pull it moves right and when I push it moves left

I also noticed it jumps. Sometimes it doesn't move then it does. If I loosen the belts all the way it twitches but doesn't move

Also before I changed the x beam when I pushed the head in x it was stiff and would catch in spots. Thankfully that's gone with the new beam and it's nice and smooth

[hybrid island]

Oh that's good.

So unfortunately we just need to figure out what's binding on the Y/A (upper) belt path.
The layer shift was actually on the other motor for some reason so there might be multiple things going on but I think we need to start with the binding.

I can't really offer advice other than checking the belt routing and bearing/idler rotation. like I said earlier "flossing" each bearing with a belt length might help so you can put some tension on it.

Also check that both motors feel the same when turning them by hand with the belts off the pulleys.

I just looked at the config file again and I think it might be a good idea to turn up the X and Y motor currents to 1. Those cnc parts are heavier and stiffer so it might need more juice than normal. You also don't have to worry about melting the motor mounts.

[valid sequoia]

I'm all for trying that before I take the belts off. Unfortunately whoever built this originally cut all the excess off the belts so once I remove them it will be a lot harder to re install them

[hybrid island]

That's no problem. You might just need to remove this bolt when installing the belts into the x carriage. Then once they are in the carriage you can put this bolt and toothed idler back in and then tension it.

[valid sequoia]

Thanks