We have lift off :( Print failure due to loss of adhesion to the sheet.

[WobblyHand]

While fooling around printing some LDC's, I have had two print failures out of 4.  LDC's are relative tall and thin.  I am printing on a Prusa satin sheet sheet.  In the past I had issues with PETG sticking too well.  Now it seems the parts aren't sticking well enough!  What cured my problem for sticking too hard was spraying with AquaNet.  Before that I had super adhesion - pulling up the coating and ruining the sheet.  So I don't want to go back to that. 

Examining the last failure, it seems there was minimal contact on the bottom layer before it broke loose, like only about 20%.  The rest had warped upwards off the plate.  Another way of saying this is the part warped upwards and away from the print sheet.  This was for a 40mm diameter print.  I had neglected to use a brim for this print, which in retrospect was a bad idea.  The part fell off the plate after printing 153mm out of 166mm.  Almost, but no cigar.  This print failure happened overnight, and I was greeted with a whole lot of extruded spaghetti on my printer.

Anyone have some helpful suggestions on getting the right amount of adhesion and still be able to get the part off the sheet without damage?  I'm using the Prusa slicer and PETG.

[hoteltwofive]

Could try a number of things - playing with bed temperature, filament temperature- slowing down print speed. 
I would say if your print is pulling up like that check to make sure your bed is super clean, re-level and try some test files.  My guess would be temperature related. 
Maybe try to also print on a different spot on the bed.
Good luck- nothing worse than losing a print so close to the end and having a pile of spaghetti.

[aluminumfetish]

  Make sure "part cooling" is off for the first few layers (5 LAYERS). A breeze, a fan, open window can all cool the plastic fast enough to contract and curl and lift. I used a glue stick on smooth glass for years with great success. Once I switched to PETG, I have never printed with PLA since. PETG uses NO part cooling at all and I never get liftoff or adhesion issues.
  Perhaps also try cleaning the bed with rubbing alcohol. I had good luck printing large diameter monocores with no brims. These printers are amazing but a bit finicky. Fun Hobby !!!!



Good luck,
Mitch

[WobblyHand]

Gave it a good cleaning with isopropyl alcohol.  Then some AquaNet, most of which I wiped off.  Added a wide brim to the part.  Hoping this will help.  Giving it another go.

If this fails, on to more severe measures.

[WobblyHand]

Found the problem, pretty much by coincidence.  The issue is occasionally a small blob is emitted by the nozzle.  The blob hardens and is above the z setting of the nozzle.  So when the nozzle comes back to that location, it hits the hardened blob.  When the z is low, it doesn't do much.  When z is high, like 6", the bump jams with the nozzle and causes the print to tilt and then jam with the extruder.  I saw one of these blobs before things came unglued.  And I saw the head get jammed.  For the third time.  Darn.  Another wasted day.

I am trying some retraction settings, and reducing my extrusion rate to 99% from 100%.  Hope that might help reduce the booger production. 

Oh the brim really stuck well.  Had fun removing it from the sheet.  But if the print head slams into the print at 6" above the sheet, that print is going to become undone.

[WobblyHand]

It's funny, the settings I made to improve things, made it significantly worse.  These were settings I found in the Prusa forum, specifically to address PETG blobbing.  So much for that.  Reverted the settings.

Did a first layer test.  The zigzag was non uniform, especially at first - it was thin, but it got better as it progressed.  The pattern was ok, save for the first line, where there was a break in the line.  The pad that was formed looked ok, decent amount of squish. 

Just completed a 3x3 set of squares.  Doesn't look too bad, except one square, which has little blobs.  The other 8 squares look ok.

I'm beginning to think it is the filament.  I will replace it with a different brand and see what happens.

[Airnut]

Whenever I have print issues the first thing I do is dry the filament in my dryer for about 5 hrs on 50 degrees Celsius. More often than not it does the trick.

[avator]

I have zero experience with 3d printing but I'm just wondering... would it be advantageous to use conditioned power supply?

[WobblyHand]

Whenever I have print issues the first thing I do is dry the filament in my dryer for about 5 hrs on 50 degrees Celsius. More often than not it does the trick.

Definitely one of the things to try.  Doesn't seem to be doing the trick for me in this case.  When I print, I print directly from the heated dryer.  I also have about 100 gm of desiccant in the dryer.

[WobblyHand]

I have zero experience with 3d printing but I'm just wondering... would it be advantageous to use conditioned power supply?

Bill, the whole unit runs off of 24VDC.  So there is effectively a conditioned supply.  Doesn't mean that it's really a good supply, but it is at least slightly isolated from the line.

I think there's a bunch of little stuff that has come to bite me in the behind.  The bed isn't particularly flat, and it's a pain to straighten it out because of the design.  I did discover that my first layer was off, so I need to fix that.

It might be the filament itself.  This batch seems to have a few issues, it was quite full of moisture as received.  It originally steamed, hissed and popped due to the moisture.  Drying helped, but I suspect quality issues - as it never should have been shipped that full of water.  The filament is blobbing a little, which means hard little nubbins are above the print head height.  If the print head hits it (and it should as it is building up the height) it knocks the LDC a little.  Enough times, and the LDC just falls over.  If the filament is full of moisture, why should I think it has really uniform properties?  If there's a quality issue in one area, there may be one in another.

[Airnut]

Whenever I have print issues the first thing I do is dry the filament in my dryer for about 5 hrs on 50 degrees Celsius. More often than not it does the trick.

Definitely one of the things to try.  Doesn't seem to be doing the trick for me in this case.  When I print, I print directly from the heated dryer.  I also have about 100 gm of desiccant in the dryer.

Bruce
How long did you bake the filiment before printing?
I have had some filaments that were so bad I had to bake them about 12 hrs prior to printing.
I would also heat the nozzle and bed then level the bed and recheck the zoffset. You can also try setting your retraction distance to 6-8.

[WobblyHand]

Whenever I have print issues the first thing I do is dry the filament in my dryer for about 5 hrs on 50 degrees Celsius. More often than not it does the trick.

Definitely one of the things to try.  Doesn't seem to be doing the trick for me in this case.  When I print, I print directly from the heated dryer.  I also have about 100 gm of desiccant in the dryer.

Bruce
How long did you bake the filiment before printing?
I have had some filaments that were so bad I had to bake them about 12 hrs prior to printing.
I would also heat the nozzle and bed then level the bed and recheck the zoffset. You can also try setting your retraction distance to 6-8.

Originally, maybe 6 hours.  Then the warm reel was put in a vacuum bag with a 1lb bag of desiccant for a couple weeks.  It helped.  Prints after that were from a heated filament holder.  Total amount of baking has been over 24 hours I'd guess.  After printing, I remove the reel and put it back in the vacuum bag full of desiccant and pump out the air.

One issue is the bed isn't flat.  So leveling is tough.  The bed correction algorithm in the firmware assumes a flat bed, ie a plane, that may be tilted.  The nine point calibration pattern I printed shows thicknesses that are non planar.  One corner is high, and doesn't fit any plane.  Most of the other points could be described as a plane if you only consider a precision of 10u.  But I have over 50u difference between corners and you can see it in the print quality as well as the measurements.  My z height is still too high on average.

Think I will bite the boolet and see if I can find the nylon locknuts I bought and redo the bed mounts.  That is, once I figure out how to access them when hot.  At the moment, I don't know how to adjust them without moving the printer.  As I understand it, moving the printer may cause changes.  In my opinion, the bed adjustment design on this printer is woefully inadequate.

[WobblyHand]

I think I now understand the nylock nut trick.  I'm going to do that.  At least there's some chance to adjust things.  On the printer, the heater PCB is held to the frame by a flat head screw.  Under the PCB is a "precision" spacer.  Then the bed holder.  Precision is relative, and things are not quite flat enough.  The Nylock nut mod, is to remove the spacer, and add a nylock nut that is snug to to PCB, but allows the screw to turn some.  This whole assembly is then screwed onto the frame.  What is nice is there is some adjustment.  Turning the screw some fractional turns will raise or lower that point.  But it is still a 9 point system, which isn't that easy to planarize.  The print pattern is a 3 x 3, in the same pattern as the screws, so there is some guidance.  I'm sure it will take a few iterations. 

I did notice that some of the screws had backed off, ie, they were loose.  The PCB is FR-4 fiberglass, so the material may have yielded over temperature cycling.

I know what I will be doing this afternoon.

[WobblyHand]

Installing the lock nuts wasn't that bad.  Now I will use Octoprint to visualize the bed and figure out roughly how to turn the screws.  Preheating the bed so whatever growth due to thermal expansion will be present.  Goofed up and flipped the sheet on the wrong side.  I had forgotten to clear of the wipe area.  The bed visualizer alerted me to a z sensing fault.  The corner of the sheet was a lot higher than it should have been due to the material "crud" being there.

Level attempt #2 ongoing.  Nothing will be very good, since not much care was taken to figure things out.  I'm off by a lot on the right hand side.  No worries.  I now have a table of how much to turn the screws.  Given enough tries, this should get me closer.  Waiting for the cool down.

[WobblyHand]

A few more iterations later, the bed is leveled to within 10u over the whole surface.  -0.002 to 0.007 mm.  Good enough for me.

Now do I do the Z height setting?  So I print a 3x3 pattern and measure the thickness and then adjust Z height according to the relative thickness of the print?  My file is claiming a 0.2mm layer height.  So if I measure 0.22mm, then I adjust Z height by 0.02mm, correct?

[Airnut]

Bruce
I am not to familiar with your machine and don't know the procedure to adjust z offset. Utube should have a video of the process for your machine.

[WobblyHand]

Bruce
I am not to familiar with your machine and don't know the procedure to adjust z offset. Utube should have a video of the process for your machine.

Frank, the adjustment is buried in the first layer calibration procedure.  What I have found is the unit isn't that dimensionally stable, or the repeatability isn't that good.  After getting everything dialed in, I redid the bed visualization and it's not flat anymore.  It's either high or low in the center as compared to all the other zones.  Kind of frustrating.  But from the 3x3 pads I printed, I found a region that had good first layer characteristics.  Fortunately, the area is good enough to print an LDC on it.

My last prints were knocked over roughly around the height when I switched to variable thickness printing.  So I realized using a constant layer thickness.  I figured a very long but successful print was better than 3 faster failures.  We'll see if this strategy works.  I'm a couple hours from completing this print.  It was a 12 hour print, so I'm crossing my fingers!

[WobblyHand]

Well, I'll be.  The print completed.  Won't be doing that again though, as it took 12hr, 52minutes. 
That's about twice as long as I'd like it to be.  But it did work.

I think the issue might be that using a coarse print, then transitioning to variable layers, needs adjustment in the variable (finer) layer region.  I was getting some boogers in that region, maybe because I needed more retraction (or something).  Or it could be as simple as filament uniformity.  I really don't know, but I need a better way to reduce print time, and get detail where I need it.  That may require gcode editing.

While this was printing - I only noticed one sort of charred looking small booger.  Fortunately it was near an edge and it was no longer in the print path.  It's inside the part at this point.

My level setting is much better.  There's no evidence of the part being knocked hard enough to see any disturbance in the brim.  No cracks.  It appears that in this region of the sheet I have a good first layer.  Still need to work on that to make it better over the whole sheet.  That's a different story, but I see a way to get there.

[Airnut]

Glad you are on the right track!

[rsterne]

I share your pain.... I am currently trying to print tensile test coupons (in the Z Axis) for 3DXTech PA6-CF-G3 (Nylon Carbon Fibre), and it is consistently problematic.... I have tried Garolite, PEI and the original black coated Creality beds, with temps all the way to 110*C, with and without PVA glue stick, and like Bruce, the problem is mostly a blob getting stuck to the print when 80% finished and then the print getting knocked over by the nozzle hitting the hardened blob.... REALLY frustrating.... I dried my filament per the 3DXTech instructions for it, and it made no difference, except a bit better surface finish.... The lower the print temperature the worse the blobs are (higher viscosity and harden quicker)....

If anyone uses Garolite (FR4) do you sand the surface to remove the gloss, if so, what grit did you use, and did it improve the adhesion?....

Bob