donmilne

Way too fast I think. Also the temperature is too high for PLA, I think you are risking deformation of the teflon spacer which could lead to even more underextrusion than I see in the pictures. I would suggest establishing a baseline using say: T=220°C, speed=50, retraction enabled, 0.1mm layers, 0.8 or 1.2mm shell thickness, ditto for bottom/top layer thickness.

From what others have previously reported, I'd suspect a buildup of stuff in the hotend - not necessarily confined to the nozzle. At the very least you should try the "atomic cleaning" trick.

Another vote for OpenSCAD here. I tried Designspark Mechnical and it just didn't... er, spark. SketchUp on the other hand I have found to be a joy to use for woodworking plans - but not for 3D printing as it doesn't make true solid models. I'm not working with others, so the ability to import STEP etc is not a major consideration for me. I would love to give Solidworks a go, but the price is insane.

If I'm understanding this correctly (a picture would be nice) I believe this can be caused by not levelling the bed properly, i.e. having the bed a little too far from the nozzle, so the filament doesn't get squished enough.

Yes, a traditional putty knife should work quite well, especially if sharpened.

You would lose a lot of torque on larger wheels.

Re the mention of geared steppers: bear in mind that increasing the motor torque, while necessary, isn't enough by itself. All that happens is that we reach the stress failure limit of the filament itself, i.e. we get grinding. If we want to return to a stronger motor then we need to distribute the force. I assume the 360° suggestion means 3 or 4 drive wheels around the filament? That might be ideal in the sense of distributing the load, but wouldn't it complicate filament loading? Also, we'd be losing the spring loaded idler bearing, so maybe it wouldn't cope as well with varying filament diameters.

I found that if the part was short with a lot of contact with the glass then it often didn't need glue. However if the part is quite tall, with relatively small contact area with the bed then you want to use every trick in the book: include PVA glue and "enable brim". Think of it as a torque problem: the torque applied by the print head is minimal, but if you have a very tall part and minimal contact with the glass... The choice of filament also matters. Unlike UM PLA, I've found that ColorFabb doesn't adhere to a glass bed as well as it adheres to itself, and really needs glue for all prints. Otherwise you just end up with the filament being peeled off the bed and dragged around by the nozzle.

That was my feeling too, more or less. I.e. that two wheels are better than a belt, because there would be no belt to stretch, slip or wear out. However I wouldn't go with rubber wheels. I do understand the intuition that led to that suggestion, but IMHO rubber would itself wear and slip. My choice would be two knurled wheels, ideally tuned so that they both run in each other's indentations - we don't want the filament to be chewed up more than it needs to be, e.g. during retraction-heavy prints. My understanding of the weakness of the existing design is that if you go for a powerful motor (UM1), it grinds the filament and stalls the print. If you go for a weaker motor (UM2), it can still grind the filament, but if it doesn't do that then it jumps back, causing underextrusion. It seems to me that if you had two points of contact on the filament then you could go back to a stronger motor, but have a reduced risk of grinding. I.e. each wheel would act as a backup for the other, and the load on the filament would be distributed.

Have you actually tried a wood chisel, or is all this coming from your gut? I tried a variety of thinner blades before arriving on the chisel idea, and the chisel works the best. A 25-30 degree bevel is perfect for cutting while maintaining the edge, which is why chisels are given that bevel to begin with. There is absolutely no danger of damaging or scratching toughened glass with a carbon steel tool unless you deliberately stab it into the glass, which I have no immediate plans to do. If you use a razor then it is not "the weakest link in the process"... that description better applies to your hand - an uncovered razor blade is about the most unsafe option I can think of. It's all very well to offer advice on things not to do with a razor blade, IMHO the best advice is not to use one at all, as it's totally unnecessary.

Try the wood chisel, nothing beats it. A razor blade is a terrible idea, you don't want something that bends - and if it breaks you'll very much regret it. Believe me, the perfect tool is a robust hand tool purposely designed to cut or separate things under a bit of pressure while keeping you safe, and includes a convenient handle to grip with. That describes a wood chisel, not a razor blade. Yes, you do need to get off your lazy ass, get down to the hardware store and buy one, but it's worth it. It has to be a broad chisel to distribute the force, otherwise you could mark the part. I find 40-50mm just about perfect. Don't use a kitchen knife either. You want a narrower cutting edge that you push away from yourself like a snow shovel.

I have no issues with using a metal tool; it isn't like I'm intending to stab it into the glass! (which in any case I assume is toughened). My favorite implement is a 40mm wood chisel, very sharp. Rap it against the print bevel side up, and the print pings off. You want something sharp enough to get in under the print. I think a more blunt tool could damage the print.

Acetone won't really help - it isn't a solvent for PLA. What printing speed are you using, and what temperature? A slower print speed and/or a slightly higher temperature may reduce the back pressure from the printhead and reduce the risk of grinding. Also, it may be necessary to open the feeder to clean out all the grinding debris. Particularly make sure the grooves on the motor drive isn't filled with gunk. Also, when you can get the printer working again, I highly recommend printing out a copy of IRobertL's replacement feeder. It is so much simpler to tune and to keep clean of debris. It is also easier to see what's going on when something like this happens. A final thought that occurs is that, once the filament started grinding, the printhead may have overheated. The PLA cooks and causes obstructions, and the PTFE spacer can distort or be plugged. Take heart: there's a heater at one end, a motor at the other end, and a tube in between. Whatever the problem is, it must be simple.

I'm not getting any of those problems with it. Could you have a faulty temp sensor?

White PLA from Ultimaker? Personally I've found it to be my most reliable (least fussy) filament of late. I'm printing at the default temp - 210C I think.

I have my doubts that the problem is really solved. It may just be covered up. I know from experience that a thicker top layer might appear to produce an ugly but solid top even when it's underextruding badly, but the part as a whole will have a strength similar to pummice (i.e. crumbly and not very strong at all). Underextrusion is my guess. The walls look just as bad to me as the top. Perhaps this is due to going for that wider nozzle and not making all the consequent tweaks that should be needed? We seem to have a lot of missing info here - I don't even know what printer it is. I'm guessing UM1?

Fairy noof. I've ordered a 500g tin of automotive copper grease.

Does the anti-sieze compound have an affect on function? E.g. act as an insulator?

No, I'm even sure how that would be done. I'm not on YouMagine or any other filesharing site, and this site doesn't allow zip attachments. I suppose I can try pasting the OpenSCAD code in a code block below, people can then do whatever they like with it. /* OpenSCAD Bowden Clip */ inner_D = 7.95; lip_thick = 1.4; outer_D = 13.65; H1 = 1; H2 = 2.6; /*..............................................................*/ module flat_sided_cylinder(diam,thick,height) { difference() { cylinder(r=diam/2,h=height); union() { translate([thick/2,-diam/2,-1]) cube([diam,diam,height+2]); translate([-diam-thick/2,-diam/2,-1]) cube([diam,diam,height+2]); } } } /*..............................................................*/ module holes() { translate([0,0,-1]) cylinder(r=inner_D/2,h=H2+2); translate([0,0,H1]) cylinder(r=outer_D/2-lip_thick,h=H2); translate([-outer_D,-(outer_D/2+3.11),1]) cube([outer_D*2,outer_D,H2*2]); translate([0,0,-1]) rotate([0,0,-135]) cube([outer_D*2,outer_D*2,H2+2]); translate([outer_D/2-0.75,-outer_D,-1]) cube([outer_D*2,outer_D*2,H2+2]); translate([-outer_D*2.5+0.75,-outer_D,-1]) cube([outer_D*2,outer_D*2,H2+2]); translate([0,0,H1]) flat_sided_cylinder(outer_D-2,(outer_D-1)/2,H2); } /*..............................................................*/ module main() { difference() { cylinder(r=outer_D/2,h=H2); holes(); } } /*..............................................................*/ $fs=0.4; $fa=2; main();

After noticing that the bowden clip on my feeder had vanished, I just took measurements off the one on the extruder end. It's quite a simple shape and easy to make a good facsimile in OpenSCAD. I printed off a dozen or so in Colorfabb PLA/PHA. They seem to work just as well as the original. Yes, I found the Ultimaker drawing - which gave dimensions that didn't seem to match the part I had. Intentionally wrong drawings to foil the Chinese? I also found alternatives on Thingiverse which did the job but weren't the exact match I wanted.

I wonder how one would go about putting a teflon coating on a ceramic substrate?

Anyone have any tips on the easiest possible way to get that spacer out of the UM2 printhead and back again?

That ridge doesn't look like wear to me, or even heat distortion (though heat may play a part). It looks to me like the plastic has been formed around a former of some kind... e.g. around the top thread of the brass nozzle, perhaps while the nozzle was being fitted?

I've been rereading this thread with a view to turning my own spacer at the weekend. What was the verdict re PEEK vs PTFE? Cleven talked about PEEK to begin with and then I think switched to PTFE? Daid asserted that one PEEK test jammed within an hour, but didn't suggest why that should be the norm. I have both PEEK and PTFE rod here so I'm somewhat neutral, though PEEK seems harder and easier to machine. Comments?

Hmm. I wonder if the glass filled teflon is more susceptible to abrasion, especially with lots of retractions.