nick-foley

Agreed. The only real reason I see to make a direct extruder would be printing in highly compressible filaments like TPE, which I haven't tried, but seem like they would bunch up in the bowden tube and have a very long lag time between desired extrusion rate changes. Nylon does this to some extent because the filament is very soft. The lower speed of a heavy print head wouldn't be a huge problem because I'm getting my best results printing nylon at 30-40mm/s anyway...

I'm getting pretty good results using Taulman 645 Nylon, mostly stock hotend: 245C 30mm/s speed 0.2 layer height 2.6 specified filament diameter (3.X actual filament diameter) No fan during print Using a piece of 1/16 Birch Ply on top of the normal build platform for adhesion.

Printed for several hours today using the thin walled teflon tube I mentioned above. It definitely deforms a little more than the normal tube, and I had to print a thicker bowden clamp (blue piece) in order to prevent slipping (because the teeth of the bowden gripper don't dig in as much).... but once that problem was worked out it seems to be going great. Maybe retraction will be a little bit less responsive, maybe rapid changes in extrusion rate will be a little less responsive - but we'll see. Did several prints in the ~3.0 Taulman 645 Nylon without issue, which was previously clogging due to thickness variationin my stock machine and also on one with the E3D hotend. (I should also mention that I reamed out my entire hotend with a 3.4mm drillbit as well.)

To be clear - does the Z stepper stop turning? (meaning an electrical problem) or does the stepper keep trying to turn but the bed no longer goes down? (meaning a physical problem)

Another great execution. Those 3D printed outlets are exactly like what I had in mind.

If you're in the US, makergear has a lot of useful replacement parts. Not sure about the thermocouple. Switching to a thermistor is relatively easy to do though, and should only take you an hour or so. Anyway, what were you doing that cause such rapid destruction of your hotend? Particularly the thermocouple?

This looks awesome. Could you share some specs on the blower you're using? You should 3D print the terminating feature of the silicone tube - you could probably create some very useful airflow with the right nozzle shape.

I have some of the thin-wall teflon tube from Mcmaster that I'm going to be testing out this week in an attempt to solve this same problem. http://www.mcmaster.com/#5239k12/=nu16sx It seems promising. It's obviously easier to crush, but the overall strength seems similar. Haven't tried the nylon though.

The guys @ E3D have a good instruction for their hotend - make the final tightening of the nozzle while the hotend is above it's operating temp - probably 240deg. That way you can account for the thermal expansion. Another useful way of assembling is to rotate your entire hotend so that in order to loosen, the heater block has to spin past the M3 thumbscrew - essentially using the thumbscrew as a hard stop to prevent any loosening.

YES. Good point. I spent several days trying to debug this problem when I built my first machine. I thought I had everything tight, but once the machine is assembled, you just don't have enough clearance to get those pulleys as tight as they need to be. Really, when it comes down to it, this is another case of the wrong part being spec'd with the stock Ultimaker. A more effective solution is this: http://www.mcmaster.com/#catalog/119/3052/=nto4lp Which is a slightly longer, cone-point set screw of the same size. The extra length (not really extra, but correct, as it just fills the tapped hole of the pulley) allows you to properly tighten things without worrying about stripping threads, and the cone point is much more effective at setting into the hard steel of the axis rods.

Ditch the stupid M3 machine screws used for installing the limit switches and use #4 x 0.5" Pan head, torx drive plastite screws instead: http://www.mcmaster.com/#99512a219/=nt3cr5 They are faster, easier, and much stronger + more reliable long term. The current screws are a very dumb use of a machine screw (and they are tapping the switches even, good lord that's a lot of unnecessary work, for such a terrible result!) and it is inevitable that you will strip some of the switches either during assembly or later when you're adjusting them. It is not fun to have a machine whose limit switches you can not trust.

Cool. It doesn't seem to be happening now; whatever combo of voodoo and filament I'm currently using is working for the moment. Someone needs to design a hotend which is just one clean aluminum piece, with no internal transitions. The nozzle, heatsink, and heater block should then mount on external threads just like the Makergear hotend does. Having internal connections is just too failure prone.

I'm trying to debug this right now, and I think it is exactly the second issue described - I think the tolerances on this hot end are so tight that any minor filament grinding/deformation from an area that has a lot of retraction cause a change in the size of the filament enough that when that section makes it up to the hot end, it causes a plug. Printing Printbl PLA at temps between 205-220.

Interesting to note. That matches up with the quotes I've been getting back from US Sensor. Anyway... ...We seem to be having some clogging/plugging/underextrusion problems with this hot end. We haven't exactly figured out what's causing the failure, but we've had several prints over the past few days that slowly transition from solid part to underextruded fluff as the print progresses. It seems to be related to small/thin-walled prints, and maybe somewhat related to retraction, but not entirely, because we're seeing it even with retraction turned down to 1mm. (Edit - it is not being caused by retraction, as I now have a print that, 15 minutes in, is starting to fail due to underextrusion, and it has retraction turned off.) The current print, which failed 3 times on the machine with this hot end, is now printing without error on our machine next to it which has a (more or less) stock hot end Any ideas? Have you seen this before?

So, as the title says, our Ultimaker occasionally freezes mid print. It stays at temperature, and stays turned on, and nothing in the Ulticontroller changes - but the print head stops moving. We are printing from a Raspberry Pi, and so my initial assumption was that the client on the Pi had crashed, and I'd been trying to debug that. However! Today we discovered, inadvertently, that if we use the Ulticontroller to "Disable stepper motors" while the printer is frozen, the printer starts working again. It then finishes the print without issue. This makes me think it is related to the stepper motors or stepper boards overheating. Is there any indication given by the Ulticontroller when the machine shuts itself down due to an overheated stepper board?

Not sure, but how'd it happen?

How big is the STL? Anything under 60 megs usually loads relatively quickly in Cura.

Interesting idea. I'd love to try it, as I'm currently building another Ultimaker that I don't have an E3D hotend for, and I would love to use something other than the stock hotend. I wasn't familiar with the J-Head nozzle. Are there examples of it being used successfully with a bowden tube? I'd definitely purchase one and redesign the print head to accept it, but only if it's something I can actually test on an Ultimaker.

I just built up a machine using GT2 belts and pulleys (sourced from SDP-SI and adafruit) and am seeing some big improvements too. The print quality is definitely higher, though some of that is likely due to the improved hotend and not just the pulleys. Regardless, moving the head by hand is much smoother, and the entire machine is much quieter in operation. I don't think this is too much a factor of MXL vs. GT2 - I think it's mostly just that the GT2 pulleys are higher quality and not crazy eccentric like the standard UM pulleys. I do like the firmware elegance of GT2 pulleys requiring exactly 80 steps per MM though, vs 79.XXXXXXXXX or whatever it was for MXL. With no adjustments, parts were coming off within .1mm of accuracy.

Awesome. I just finished up an improved xy block design that I'll post shortly. Seems like a big usability improvement.

Sanjay - The only major problem I have with this hotend so far is the thermistor - it is crazy ghetto. Shoving a tiny fragile element into a tube, packing aluminum foil behind it, and wrapping the whole thing in kapton tape is some macgyver garbage that doesn't fit with the rest of this otherwise elegant design. You should change it to be something like this: http://www.ussensor.com/h3457-fast-response-micro-probe or this: http://www.ussensor.com/h3134-s/s-set-screw-8/32 Don't those look sexy? God, they would be sweet. They would probably perform much better as well. I'm going to order one and modify the existing thermistor hole to accept it... ...actually, they may be ridiculously expensive. Will try to find something reasonably priced and more elegant.

Also (probably) a nozzle, teflon tube, peek, brass tube all made for 1.75mm filament. The simplest/most effective way to switch would probably be to purchase an E3D hotend that's made for 1.75 filament, and have two print heads that you switch between when you want to print in different filament diameter - assembly of the stock UM printhead is a pain in the ass, because the tightness of the thumbscrews affects the nozzle location so much, and the wire routing is challenging. I suspect in this setup, you could probably keep the 3.0mm bowden tube.

We use them in many critical plastic pieces on the bike-share systems we design, and there has yet to be an instance of one working itself loose. They see a lot of vibration. They hold extremely well in plastic because they have a triangular profile, which allows the plastic to creep around the screw threads after the screw is inserted, effectively locking the screw in place. Apart from being way easier to drive (Torx and aggressive threads FTW), they should be much more resistant to vibration and loosening over time. The also allow for much simpler assemblies, because you don't have to worry about being able to access the nut in order to tighten them. The only downside is that you probably need an extra mm or two of thickness in your pilot hole bosses, because you don't have the compressive force of the nut on the other side - you're using the walls of the boss itself. Even with the extra plastic though, there is still a weight savings vs. an M3 + nut. The other downside is that getting your hole size correct is a bit more critical, but on my machine a 2.2mm pilot hole and a 3.5mm clearance hole seem to work with a variety of print settings when using #2 plastite screws. Does anyone have any good info (anecdotal or otherwise) about the long-term stability of printed PLA parts, btw? Like creep, fatigue, UV, and humidity resistance?

Haven't tried the colorfabb but I can confirm that the Printbl Grape is an awesome material. I have two machines printing in front of me, one in Printbl Lime Green and one printing in Grape. The Lime Green is ever so slightly cleaner than the Grape, but that may just be due to machine calibration. Of the Printbl colors I've used a roll of, I'd rank them Lime Green > Grape > Banana Yellow > Ruby in terms of appearance of finished prints. Regardless, they are all wayyyyyy better than the Ultimaker PLA. Definitely worth the money.

Hah, absolutely, didn't think anyone would want them. Uploading to thingiverse now... ... here they are: http://www.thingiverse.com/thing:121293