TPU printing still kicking me around

[WobblyHand]

Hmm, another day, another print failure.  One with a long time into it, about 12+ hours.  In full disclosure, it was a tesla valve moderator with 45mm OD and 180mm long.  I designed it and realized it was a large OD, but I'm not that good in CAD and wasn't quite sure how to scale it cleanly and still have it work.  Going to have to figure out how to do that at some point.  But on to the failure.

This one was a little weird, the TPU filament decided to jam in the extruder, and the printer merrily kept on printing, although no filament was feeding.  The screws holding the extruder tube are still tight.  I read about possible causes, but I'm not really sure what to do just yet.  I read that for the Prusa MK4 you should tighten the tension on the idler assembly, as opposed to loosening on the MK3S+.  Read further that someone made a mod on the idler arm to apply more tension - so I'm not sure if I need to do this or not.  Also read about volumetric flow adjustments for the filament.  So I added a little more tension and reduced the volumetric flow for the material.  Of course, this makes the print even slower.

Since this doesn't happen often, it's tough to debug.  I have made a smaller cylindrical model that has similar features in it (30mm OD, 20mm high) and am printing it now, without input shaping.  Then maybe with input shaping.  I'm not all that confident that either model will actually be useful for forcing errors.

Material is Sainsmart TPU because it is one of the few with full characterization by Prusa.  It's not the high speed stuff, I haven't yet gotten some of that material.  And I'm still getting wispies, but that's not my primary concern at the moment.  As an aside, it might be relevant (or not), the filament is in a dryer box which is sitting on a table top, same surface as the printer.  The filament path goes up from the box, over a pulley on the printer and down to the extruder.  It's not super low friction, but I always see the spool turn (in the heated dryer box) as the filament is used and it doesn't seem to hang up, that I can tell.

For all I know it could be the filament winding on the spool, (which didn't look that great) or slicer parameters that need tweaking.  When I cut the filament to yank it from the extruder, I measured it and it seemed a little large, about 1.78-1.79mm.  I then measured it closer to the spool, and maybe it is 1.76-1.78mm, but it doesn't seem round.

I can post my Prusaslicer model of the test piece, if anyone wants to take a look.  The full model is far to large to attach here.  The test piece model file is 45KB when zipped.

Here's a picture of the test piece in mid print.  Temp is 220C.  Stringing was worse at higher temperatures.

[TorqueMaster]

I can't offer much advice, other than start where the failure happened -- "... the TPU filament decided to jam in the extruder..."  and leave everything else alone.  It sounds like it is feeding to the extruder ok, just something there is unhappy.  On my ender, they recommend slightly less pressure squeezing gears that drive the filament as compared to pla or petg.  Enough to squeeze, get a grip, but minimally flatten was the idea I took away.  Adjust flow percent if needed since TPU may slip differently than harder materials.  But your print looks good.  Good luck, not worthy on toast that it happens after HOURS, so troubleshooting will be painful.  Makes lots of smaller things until you "trust" it perhaps.

[WobblyHand]

Found some further advice on the Prusa forum.  The XL and MK4 extruders need a little more pressure for TPU.  I also got a suggestion of reducing the retraction speed from 60mm/sec to 20mm/sec.  Also reducing the temperature to 218C reduced stringing for this Sainsoft TPU.  I've printed a couple smaller samples, and they are ok, more or less.  I have found the faster print speeds just don't work well for this material - which is to be expected.  They print, but there's too much stringing.

Right now I'm running a slightly larger test print, at the new settings and have changed the perimeter generator back to Classic from the baseline Arachne. It seems the Arachne generated seams really are butt ugly.  There's either a wide seam that is greater than the diameter (if aligned), or big zits sticking out.  The random seam with classic perimeter generator is much less pronounced.  At least for these parts, I will be using the classic from now on.

So far really don't have a reason for the print failure, that I can see besides a jam.  Hopefully changing the retraction will help.  That or making the part print even slower.

[Airnut]

I had a similar issue and it turned out to be loose electrical/ mechanical connections at the hot end.

[WobblyHand]

I had a similar issue and it turned out to be loose electrical/ mechanical connections at the hot end.

I will definitely take a look for that.  Maybe it is that simple.  I found that my strange inclusions (in TPU) were partially caused by a loose mechanical connection.

[YEMX]

Is your printer in an enclosure?  Give that a try as well, when you're able.

[WobblyHand]

Is your printer in an enclosure?  Give that a try as well, when you're able.

No, it isn't.  I'd like to make an enclosure, but haven't printed up the parts.  Might make sense to start that up again.
From what I have heard, TPU doesn't need an enclosure.

I'd like to filter the air, specifically the stuff that is released during a print.  Seems to bother my eyes after a while.  Yet another rabbit hole to fall into...

[WobblyHand]

Jeepers, 0 for 3 on this moderator.  Think this spool of Sainsmart TPU is prone to jamming.  1st one extruder jam, 2nd one had adhesion failure, 3rd one I added a brim, but heard the click, click of the extruder and saw nothing was coming out.  Opened up the extruder, and sure enough, the TPU material managed to jam again.

I'm going to abandon this material for a bit, and try the Overture TPU again.  At least I've had a few ok prints with the Overture material.  All the Sainsmart test prints worked fine, but printing the moderator would fail after many hours.  Darn stuff was relatively expensive compared to the Overture.  The Sainsoft also looks like it has moisture and the dryer I have doesn't get hot enough to dry it out well enough.  (55C max)

Anyone know of a dryer that gets to 85C?  I found the PrintDry Pro 3, but it is pricey at $199.  Are there any other options?

[LDP]

Get a food dehydrator for drying filament. Its allot cheaper and can handle quite a few spools.

[WobblyHand]

Think it is the tension on the idler arm.  Someone on the Prusa forum checked their tension and said I had set it too tight.  Kind of strange since I had about 6 successful prints at the higher tension.  Backed off the screws a full millimeter.  (2 turns).  Going for try #5.  Will see how far I get.  I can tell I need to do a little more with the retraction length, but so far it seems to be a little better that it was.  What was really difficult about this, was simple prints would succeed, but this LDC wouldn't.

[Airnut]

When you lost extrusion and heard the extruder click was the hot end still up to temperature? You could have lost the hot end temp causing the lost extrusion.  This is exactly the same problem I had. Ran great on short prints but would have extruder issues on long prints. I found the hot end mounting screws loose along with the thermistor wire.
As for your drier box you just need to bake it longer 12 to 14 hrs should do.

[WobblyHand]

When you lost extrusion and heard the extruder click was the hot end still up to temperature? You could have lost the hot end temp causing the lost extrusion.  This is exactly the same problem I had. Ran great on short prints but would have extruder issues on long prints. I found the hot end mounting screws loose along with the thermistor wire.
As for your drier box you just need to bake it longer 12 to 14 hrs should do.

Yes, I believe the hot end was still hot.  The mounting screws seem to be tight.  The tension screws only have a few threads on them at this point, hard to make it any looser.  The print is failing in the region where there's a lot of internal fill.  It's just a mm or so of thickness, but at the large area it is, the layer seems to be just about at the limit of the hot end and the particular filament type.  I read about this issue, hitting maximum volumetric flow, and it seems I have reached it for this model.  The supposed cure seems to be to go even slower, but I'm not going to run a print longer than 23 hours, since my filament dryer will turn off after that time.  Sure there's a work around, but I have drawn the line at 23 hours.

I'm on try #8, if it dies, then I'm changing the design.  I'm using the Sainsmart TPU which has a higher MVS. 

Actually, I haven't waited for the print to fail.  I designed something today that is different with a lot faster print time.  Sort of an insert approach.  Print time of about 5 hours, rather than 20 hours.  My original monolithic approach doesn't seem to be feasible for this material.  It seems others have issues with big models in TPU as well, including those with a lot more experience than I.  I can print smaller models ok, but the big ones seem to stress the capabilities of the filament on the platform.  Live and learn I guess.

[WobblyHand]

Well the 9th time was the charm.  Success, or rather it completed without failing.  Kind of fuzzy and furry inside, a lot of stringing.  I'll try a little hot air to try to clean up some of the stranding.  Hmm, not doing that large a piece in TPU again.  That was a bear to print.

What worked:
Used Sainsoft TPU 95A
Changed MVS from original 5 to 3.6
Totally loosened the extruder tension to the point the screws were barely in the nuts.  Extruder springs are loose without filament.
Set retraction speed to 10mm/s (from 60mm/s)
Set deretraction speed to 10mm/s (from 20mm/s)
Used 10mm brim.

Took 19 hours to make this tower of fuzz.  Have to chase the coarse threads and print an LDC adapter in PETG.  Wasn't any point to making the adapter if I couldn't print this.  If this LDC somewhat works, I'll machine an adapter out of metal to replace the PETG one. 

[WobblyHand]

After printing an LDC adapter in PETG, I can tell this LDC design isn't too effective, at least at this power level.  Oh well.  I know what I'd change, but I don't think I'm going to print it again, at least not right away.

For the first time I tried some silicon oil as a tap lube to chase the threads into TPU.  That worked relatively well, reducing the tendency for the TPU to grab the tap.  Based on that, I used some silicon oil on the coarse male threads and it didn't take too much effort to screw in the adapter.  So at least I learned something out of all this.

[WobblyHand]

I found some Sainsmart GT-3 High Speed TPU, which seems to print pretty well.  Printing some spool holder inserts out of TPU and I'm finding the print quality is rather good, and that's at MVS=5 mm3/s rather than 3.6.  Now this might just be due to the spool holder design doesn't get near MVS, unlike my LDC design did. 

However, there's very little stringing at this point and the print looks quite clean at 1/2 way through.  Printing TPU is slow, but at least this new print is coming out way cleaner than the "normal" TPU which was quite stringy.  After the last TPU disaster, this is a rather welcome relief.

[rsterne]

I have very limited experience with TPU, but there was a HUGE difference in quality when I dried it before printing.... Some of the prints before drying were true Frankensteins....

Bob

[WobblyHand]

Two different TPU's, both Sainsmart.  Light blue GT-3 pretty good finish, very light stringing.  Darker blue had to print slower and had coarse strings.  Although both were rated at 95A the poorer printing dark blue one was a little bit softer.  Both were dried at 65C for 24 hours prior to printing.