Nozzle crashing into bed during first layer

[rsloan76]

After reassembling my metal plus after installing an Ubis13S, and after setting up the z-probe and making sure the gantry was level (using calipers on each end of the gantry), I noticed that at a certain point during the first print layer, the nozzel would start scratching the print bed.

It puts down the skirt layers perfectly as well as the outline layers of the actual print. Then it starts filling in the first layer. After a few passes, the nozzle begins to get lower and lower until it starts scratching the bed. -At which point I flip the switch off.

I've checked the gantry for level after each one of these failures and what I'm seeing is that it's falling out of level! Looks to be around 0.75mm-1.0mm difference on each end of the gantry.
The way I'm leveling the gantry is by turning the threaded rods individually, by hand until both ends measure the. same height from the print bed.

I'm sure how to remedy this. Never had this issue before, and I've leveled the gantry in that manner multiple times over the last 4 years.

[Mooselake]

Did you change the Z screws? i swapped the ones on my Kickstarter plywood bot with TR8-8, and found they had so little friction they would drift down during printing. The older (and perhaps newer...) Printrbot firmware turns the motors off between moves, which means they have no holding torque, for inexplicable reasons. I had to update the firmware to keep the motors powered up and prevent this.

Could this be your problem? If the motors are off (it's a compilation option PB picked) the carriage will drift down. I reflashed to fix it, got the image for my original printrboard from another forum member but these days I'd see if Marlin 2.x would build and run on it.

cxbiker61 has . He still stops by occasionally if you want to ask about the motor power issue.

Kirk

[rsloan76]

No I haven't changed the Z screws.

It occurred to me earlier today that maybe I was manually turning each z-screw incorrectly. I would turn each one independently (to change the height of one end at a time). But I was NOT holding the other z-screw in place when I did this. I will try this when I get home this evening. Fingers crossed, but honestly I'm not really hopeful since when I turned each screw individually without holding the other one still, I was still able to check for level with the calipers and I was able to get level. We'll see I guess.

[RetireeJay]

The original Printrbots came with 5/16 threaded rods that had 18 turns per inch; it was pretty much impossible to force those rods to turn by exerting vertical pressure on the gantry. But then they started using Acme threaded rods for the Z axis, even ones with two "starts", and these rods have such steep thread pitch that it becomes very easy to make them rotate just from gravity acting on the gantry.

A little side discussion about how stepper motors work: even when stationary they can be (and should be) fully powered up (and therefore hold their position very firmly). But when powered down, they don't have much resistance at all to being rotated.

Some versions of Marlin allowed the Z motor(s) to be powered down automatically after a certain period of inactivity. This really never made any sense; the power savings in electricity are negligible, and if the motors were getting too hot then the proper solution was to reduce the driving current being fed to the motors. But anyway, somebody somewhere thought it was a clever idea - and it didn't cause a problem in the old days with the shallow pitch screws. Now, with the very steep screws, it does cause a problem.

You could check this theory yourself; anytime you can turn the threaded rods by hand, the motor is off. For example, with the power off completely, adjust the rods as you describe to level the gantry, and then try pushing down on the gantry just a little bit; does the gantry move and make the screws turn? We assume that they hold position when totally static (power off), but during operation there are vibrations that loosen up the grip and make it more likely for things to slip.

I guess since you moved from a ceramic hot end to a Ubis 13, you added a fan (along with supporting bracket etc) so you've made the gantry heavier and more likely to cause that undesired gravity-induced rotation of the Z rods.

And it's not surprising that one side would "drift" down a little more than the other; no mechanical structure is ever totally and absolutely symmetrical with regard to geometry or friction.

So Mooselake's proposal to find a firmware that keeps the motors powered up sounds to me like your best solution. Be sure that you find one that's compiled to work with the Metal Plus and compatible with the particular board that's installed in your printer.

[rsloan76]

Ok. I tried pressing down on the gantry and it didn't budge.

I re-leveled the gantry making sure to hold the "other" z-rod in place.

I started a print (3mm cube)
The first layer printed fine. So I figured I would just let it print the rest. No sooner did I turn around from the printer did I hear the nozzle scraping the bed! On the second layer!? How does that happen?
Nothing seems to have been jostled. Nothing seems out of place. Checked for gantry level = fine. Checked that extruder was seated correctly.= fine.

I have no idea what's going on. Could something be wrong with the z-probe? Might it be failing to detect the metal which is causing the extruder to move closer to the bed?

[RetireeJay]

After the printer establishes "zero" on all three axes, it then ignores any inputs from axis-limit probes or switches.

I'm wondering if something has caused your Z motor(s) to be running in the wrong direction. This could happen with incorrect firmware, or it could happen if you unplugged the Z motors from the Printrboard and then plugged them back in with the connector reversed end-for-end. It wouldn't hurt to take another look at ALL your connectors to make sure they are all installed correctly. However, the easiest way to check for Z direction is to physically move the gantry up by 5 or 10mm when the power is off. Then you'll know for sure if it moves in the right direction when you turn on the power and tell it to zero the Z axis. (If you're always starting from a point that's within the sensing range of the probe, you might not discover if the direction is wrong.)

Are you running the printer from a computer (via USB) or from an SD card? If you are running it from a computer, you probably have a User Interface (UI) of some sort (e.g. a "Pronterface" type UI built into Cura, or maybe you're using Repetier or some other software).

Using whatever UI you have, see if you can just issue commands (not during a print, just a commands by themselves) to zero the Z axis and then raise the Z axis by 1mm or 10mm. Does the gantry move up or does it move down? Does the readout of position agree with the actual position?

[rsloan76]

Im using Simplify3D connected to PC via USB. Every (movement) command I send to the printer works as expected. Up is up, down is down, etc, etc.

After some more leveling of the gantry, and setting the z-offset to where the filament was laying down nicely, I let the print continue (a 2.5cm cube).
During the 3rd layer, the nozzle started dragging through the previous layers and I shut it off before it got down the bed.

So basically what's happening here:
Printer runs G29
Head moves to center and prints the outline skirt (4x outlines)
Starts to print cube.
Lays down outer wall and first layer nicely.
At some point after this, the print head starts to get lower and lower. I can't actually see it lowering, I just see the nozzle start to drag through the previous layer more and more with each pass.

[RetireeJay]

Does Simplify3D show you a live indication of current X, Y, and Z? I assume it does. So if you watch the reported on-screen Z that it "thinks" the printer is at vs the actual Z that you observe, do they agree or do they disagree? If they disagree, then the motors are not doing what the program is telling them to do.

Possible cause #1: Bad connection. Stepper motors have two independent coils of magnet wire inside; to sequence the magnets properly and make the motors move correctly requires all four wires to be making good contact; if even one wire is not making good contact then the direction of motion can get scrambled. So if your plug from the motors is not fully seated on the board contacts, maybe one of the wires is going "open circuit" some of the time (e.g. due to vibration).

Possible cause #2: Insufficient drive current. Your board probably has "digipots" that set the current drive to the motors. Depending on your version of firmware, you may be able to adjust the current drive using M commands. The motors are probably not in the M500 series of settings, if I recall correctly they are somewhere in the M900's. You'll have to do a little research to locate the exact numbers.

[rsloan76]

I'll have to check the on-screen readouts when I get home. Thanks!

[RetireeJay]

[rsloan76]

Hmm. Thanks for the info! It's been a busy few days. I'll have to look into all this this weekend.
It's hard to believe these issue might be cropping up all of a sudden just because I took the printer apart and installed a new hotend! ...Isn't it?

[rsloan76]

Imputing an M503 command gave me this:

Steps per unit:
READ: echo: M92 X80.00 Y80.00 Z2020.00 E93.50
M92 X80.00 Y80.00 Z2020.00 E93.50
READ: echo:Maximum feedrates (mm/s):
Maximum feedrates (mm/s):
READ: echo: M203 X100.00 Y100.00 Z5.00 E14.00

2020*5=10,100 which is <11,000 so I guess that's not the issue right?


As for observing the live output of the X,Y,Z coordinates. What's good way to do that? Should I start an actual print and try to watch the numbers during those first few layers?

[RetireeJay]

[rsloan76]

Ok. I started a print and watched the printer (the computer is about 5 fit across room from printer. I have to roll my chair to get from each device haha). As soon as the nozzel started scraping, I killed the power.
I went over the the "Communications" tab and copy pasted everything.

In the text doc, this is every instance of "G1 Z"
SENT: G1 Z0.200 F1000
SENT: G1 Z0.300 F1000
SENT: G1 Z0.200 F1000
SENT: G1 Z0.300 F1000
SENT: G1 Z0.200 F1000
SENT: G1 Z0.300 F1000
SENT: G1 Z0.200 F1000
SENT: G1 Z0.300 F1000

Not sure what to make of that.


But I think I may have found the issue. When reassembling the extruder cart, I noticed that no matter how tight I tightened the screw that holds the extruder to the motor, the hotend would not tighten up in its hole. It had about 1mm of play up and down with the screw in place (through the notch at the top of the hotend).
Looking into this issue last week, I found this thread:

So I decided to take the part, put it into a vise and slowly clamp it a little bit more closed. After that, the hotend felt a bit tighter but was still slipping. Then I noticed if I rotated the hotend within the hole it would tightened up. That's weird right? That would imply either the hotend or the hole is not perfectly circular right? Perhaps carelessly, I took that as a win and tightened everything up. I found that the hotend was not slipping anymore. Yay? (This was all before I made this forum post)

So today I started a print (and recorded that data above), then I started another print after that without any filament to really stare at the hotend and also try to keep and eye on that z-rod to see if it moved at all. The print got about 5-6 layers in and I shut it off. Everything was fine and nothing moved. I put filament back in and started another print, again, watching the hotend and z-rod intently. I wondered if some harmonic resonance or other vibration might be causing the hotend to slide down over time? A few layers in, sure enough, the hotend started going down (as observed at the nozzle end). I shut it off, and noticed that the hotend appeared to be just the slightest bit lower ~0.5mm. I was looking at the spot where the top of the hotend lays flush with the aluminum clamp-holder part. I started another print without touching anything. Sure enough the nozzle crashed, but as expected, much sooner in the print (1st layer)

I took it all apart and test fit my original ceramic hotend in the hole. Of course it was tighter than I remember because of the vise stuff I did to it, but it rotated within the hole smoothly. So that tells me my Ubis13S is not perfectly circular. All that said, what might I try to make sure the hotend stays put? As per that thread I linked, I could try shimming it I guess. I really hate to do that. It feels so "hacky." I like the idea you put forth in that thread about drilling and tapping a hole for a set screw. I would easily be able to do that, but I'm a little apprehensive since it is a "destructive" solution. Anyway sorry for the long post. That's all the info I have right now. I think I'm going to try to shim it with PTFE tape as suggested by Mooselake in that thread.

[RetireeJay]

[MartyS]

If I had that issue I would take out the screw, put a single layer of aluminum foil around the area of the hot end that gets clamped, pry the clamp open to get the hot end in the hole, then put the screw back in and clamp it down. The small size screw and screwing into the stepper motor limits the force you can put on that fairly thick aluminum clamp, so best to have it tight from the start.

It's kind of amazing how much sideways force the nozzle puts on the clamped area while it's spitting out melted plastic, add to it the fact that it's constantly changing direction and it really needs very good clamping action to keep it in place.

[rsloan76]

Ok. I think we're good now <fingers crossed>
I tried wrapping the top of the extruder in teflon tape then running a print. The cube printed fine. Next I started a print of the new fan shroud I modeled for the Ubis13S. It printed mostly fine but there were some things I wanted to fix on the 3d model so I stopped it about 20min in. I started the print again and the nozzle crashed! I looked closely, and sure enough, the extruder had moved down just a tiny bit!

I decided to try another shimming method mentioned in that thread. I cut out a little strip of aluminum from a soda can and stuck it in there. I successfully did the 2hour print of the fan shroud and continued some 25-50min prints throughout the day. So far so good.

Thanks a lot for all the help! So glad this forum is still active enough to get some help with this thing.

[MartyS]

Don't know why anyone would suggest Teflon tape for this type of application. Even o-rings made from more dense Teflon require constant re-tightening as the stuff just oozes out with pressure.

Metal shim was the way to go.