Well, we found the direct cause, now we need to find the root cause. We will get there!
Something is creating slop/backlash in the system. Just have to figure out what it is!
I just hope Murphy’s not settling in for summer vacation. 
Pardon me jumping in…
Are you sure that your Z lead screws are tight down to the ends of the motor shafts? If you are using the same flexible couplings as almost everyone else, and there is a slight “springiness” to the Z coupling, this can cause some backlash. It’s not much, but is about in line with what you’re seeing.
My printer, the X and Y axes are leadscrew driven, so it’s immediately obvious when something like this is happening, the ringing goes from “normal” to “OMFG!”
I seat the leadscrews all the way down in the coupler, and use 2 scrapers to stretch the springs a tiny bit, then tighten the grub screws. Then I seat the motor shaft in all the way with the coupler still slightly stretched, and tighten grub screws. This seems to eliminate backlash from the connections, though if you’re not really careful, it can cause some wobble to the leadscrew with respect to the motor axis. (I have a very slight wobble in my X axis screw, which gives me some “X wobble” – kind of like Z wobble, but at 90 degrees, lol.)
2 Likes
At this point I only checked the grub screws for tightness but Thanks!, I’ll definitely give your method of assembling the coupler/lead screw a go. It’s just had my head spinning that it’s ‘off’ yet so consistently ‘off’.
Well I did this with no joy. I’ve tried tightening/loosening the rollers with the same frustratingly consistent result, lower it a bit and the first commands to move .1mm up result in only .001" of movement, takes 3-4 cycles to yield the desired .004" of movement. I have the feeling I’m either missing something very simple or the solution will involve midnight, a graveyard, candles and mystical incantations.
A couple of questions to consider:
-
With the lead screws disconnected is the motion smooth and easy? Is there a mechanical issue they’re trying to overcome?
-
Is there a Z homing offset or a bed leveling mesh left somewhere (firmware or eeprom) that makes the machine think the Z is at a different height than it really is after homing?
-
I can’t detect any binding and the action is uniformly smooth.
-
Well, I changed nothing in the settings before this began happening and as I’ve said, it’s frustratingly consistent. Back it down a step or two and the first couple of commands to go up .1mm result in movement of only .001" before yielding the expected .004".
And to make it more maddening, I Tinkercadded a little test piece, basically half a thick walled tube that consistently prints very well from the first layer. That’s what really makes my head hurt, anything I can think of causing the issue should certainly be apparent in the test piece but isn’t.
I gotta kick Murphy in the butt, he could’ve at least brought some Guinness and Jameson if he was planning a long stay.
Are you saying that every time you move down then start back up you see this behavior? I was under the impression this only happened right next to the lower end of Z travel.
Yes I am, lower it a click or two and then send it upwards and the first click or two yield .001" upward movement each when the call is for .1mm(.004"), it may take 4 clicks before it goes up .004", and it repeats this behavior, consistently.
And as I’ve said, what really gets my head spinning is my little test piece prints consistently well from layer one and I haven’t a clue why. And there’s that nagging feeling, right or wrong, that I’m missing something simple and just can’t see it.
Many I3 style printers have the Z nuts loose, so that if you crash the head into the bed, they will just pop out, rather than stretch the couplers, or break the hotend, or bed. The only thing keeping those nuts in place in the towers is gravity. If something is wiggly there, the nuts might be coming down a little faster than the gantry, so when you reverse direction, it’s the nuts being forced back into the gantry, rather than the gantry lifting.
Generally this is a good safe design, but it can mean some weirdness if gravity is no longer keeping that tight.
Try this:
Move the gantry downwards, and attach your dial gauge. Leave it a little above the bed. Press downwards at each end of the gantry over the Z screws, and check if it moved after you release it. (Maybe wiggle it lightly.) If it does, this is where your backlash is coming from. You are basically checking for any play or a gap between the Z nut and the gantry mount. This can mean that there’s too much drag on the gantry when moving downwards, so maybe the wheels are too tight?
This’ll have to wait 'til tomorrow but one thing that still perplexes me is why does my little test piece consistently print well from layer one? It defies any mechanical logic I have. There’s only one Z screw and the nut’s tightly screwed to the gantry but I’ll fool with it tomorrow, I’ve been hoping for something to pop into my head when I’m not thinking about it, so far with no joy.
Well, that eliminates my idea of a Z offset issue.
Anything in the test object gcode compared to another job that would explain the difference in behavior? Maybe there’s something added or missing in a “start” or “filament” gcode section in the slicer that makes or skips the moves that get it working correctly.
I’ve got a few errands to run before I go do battle with Murphy once again. First thing I plan on doing is moving the indicator from the right upright to the left where the Z screw is, I think that should tell me if the issue is with the adjustment of the rollers on the right upright or not. The little test piece I’ve referred to consistently prints well with no changes having been made to the slicer settings so I don’t think that’s where the problem is.
Well, I’m still kinda lost. I moved the indicator to the left side where the Z screw is. If I back it down a bit and then try to go up, I consistently get .001-.002" on the first click, the same on the 2nd, and the 3rd yield .0035-.004". But, if I it back down and then load the left end of the gantry with my hand to a deflection between .005-.010", I feel almost as much as hear a dull sounding faint noise and it will then go up the desired .004" per click from there. It seems consistent in doing this but I can’t hear where the noise is coming from, the screws holding the leadscrew nut are tight. Of course I never tried any of these things while it was working OK so I have no idea if anything’s changed or not.
Definitely backlash in the system.
Those brass nuts don’t last forever. My main printer, all the axes are screw driven. I’m on my third set of nuts for the X and Y axes. Those are anti-backlash, so a little forgiving, but they wear faster. Also, X and Y get a lot more use than Z.
If there is something dragging on your Z axis (Wheels too tight) then it will also increase wear on that nut.
I don’t think so, I can move the gantry up/down by hand without applying much force and there’s not a whole bunch of time on the printer. It at least looks like I can get the nut(s) from Amazon, I’m guessing the added wear of anti-backlash shouldn’t be a big issue on the Z-axis?
One thing that still irks me is the little test piece I Tinkered consistently prints well with no issues.
Just curious if you think while I’m trying to correct this if a dual Z leadscrew upgrade would be a worthwhile modification?
Paul since you are still hitting your head against the wall I have a couple of long shots to try.
-
Check the z endstop position. If it comes loose your bed leveling is wacky and can cause issues.
-
Try turning up temp of hot end 10 degrees old PLA that has too much moisture can sometimes bubble and do this though I assume you have tried new filament right.
-
Preheat end to 250 and extrude 50mm or so of filament manually then do a cold pull. It’s possible your extruder has a clog.
1 Like
Thanks, I appreciate the input!
After seeing how I could press down on the left side of the gantry and correct the initial travel distance the idea of backlash in the system made sense to me so new leadscrew nuts on order. Being I’ve been satisfied with the printer up 'til now and the gantry being cantilevered off a single leadscrew seems a bit iffy, I decided the dual Z leadscrew upgrades I’d just became aware of made sense so also ordered one of those. Hopefully, be working well again soon.
The upside is Amazon seems to have shaken off its pandemic shipping woes. Ordered yesterday, initial ETA was Wed-Fri. Got an Email this AM and the order should be here Mon. I don’t have Prime, just ‘normal’ free shipping.
2 Likes
Well, seems I’m still lost in the ozone. I’d installed the additional Z leadscrew, seemed to work smoothly with nothing binding. I leveled the bed as I normally would, auto home and then adjust the height with a piece of paper between bed and nozzle. I attempted printing from a known file and it seems to have decided it wanted to print well below the level I’d set as zero, making some lovely gouges in the bed before I could kill it.
I’m clueless where to even begin looking, I’m far from expert in 3D printing but I certainly never had something like this happen before. Ideas appreciated while I’m waiting to get a new print surface.