nick-foley

When my last hotend failed, I took that opportunity to swap in a hotend I had wanted to try for a while - the Printrbot UBIS ceramic hotend. It has a lot of great features - light weight, small, no assembly, seemingly clog proof, extremely fast heating and very rapid temperature adjustments. The mount I designed is extremely compact, as well as simple to install and maintain. I think this is the best spent time/money I have ever put into the machine, as it has increased the print quality and drastically reduced time spent debugging the machine. Going along with that upgrade is a change from a printhead-mounted fan to a frame-mounted crossflow fan, as suggested by foehnsturm. (I purchased mine from Amazon here: http://www.amazon.com/gp/product/B001NPEBJM/ref=oh_details_o01_s00_i00?ie=UTF8&psc=1 ) Unfortunately, stock of this hotend seems to be low at Printrbot, but I believe this is the best setup I've ever tried on an Ultimaker. Files here: http://www.thingiverse.com/thing:208346

Here's the bracket I designed to simplify the process of converting to direct drive. It could be used in conjunction with the solution posted above to enable re-use of the original shafts. http://www.thingiverse.com/thing:208343

Flatout print speed is a mostly meaningless metric - it's all about how fast you can print while maintaining quality. It's also a very deceptive measurement because for many prints the max speed is never even reached due to acceleration limits. Also... printing the torture test with support mostly defeats the purpose of the test. I think the quality in the photos actually isn't that good - it's just ABS instead of PLA. ABS models always appear higher quality because ABS has lower gloss and thus shows the build lines much less clearly.

Filament was Ultimachine White PLA, though I'm out of that filament now, and will resume testing in Printbl White PLA. Which will also be a good time to test the possibility that filament variation is one culprit. Will probably do that tomorrow.

Good to know. With regards to PLA-optimized hotends though, I've been using the Printrbot UBIS ceramic hotend on an Ultimaker for several days now and I think it might be the best upgrade I've ever made. Will post more details soon, but compared to stock or the E3D, it's lighter, smaller, heats up (and corrects for temperature changes) wayyyy faster, and clogs seem to be a strange concept that it has never heard of...

I may have had more clogging than normal, but I still don't think there's much incentive to purchase this hotend if it PLA is your material of choice. It seems to work great for other materials - been printing heavily with XT for the past few days with zero problems. But if you're printing mostly in PLA, I think most of the benefits of this hotend are unnecessary.

STL file: https://www.dropbox.com/s/rf2ad815fwjg90u/Horizontal%20Banding%20Column%20v2.STL

Good point - don't think it's Z thread related as there is no 3mmi-ish periodic pattern to the banding.

Not really - 0.3 micron sandpaper is approx. equivalent to 50,000 grit sandpaper - and so it isn't really producing a burr. Another perk of doing this polishing (and the way I know that there is no burr in the nozzle) is that now, when I aggressively extrude hot filament through the nozzle by hand in thin air, instead of squirting out in a spiral or a curve, it shoots perfectly downwards.

On some of my prints a few weeks ago, something similar was happening - I think Cura was pre-emptively raising the nozzle temp, in anticipation of the fan being turned on full blast (and thus cooling the nozzle somewhat). Did you watch the nozzle temp when this happened? I bet the temp spiked and you started getting overextrusion.

For a while, I've been trying to track down the source of some minor horizontal banding which appears on my prints. Overall, I'm getting incredible print quality, but when you hold a finished piece up to your eye, it is clear that there is a minor amount of horizontal banding present in every print. I can imagine a few possible suspects, which have all been discussed fairly extensively here in the forums: variable material feed rates due to the extrusion system or material diameter, forces in the XY gantry which inhibit repeatability, or forces in the Z bed which affect layer height or XY position. Though it lags behind in other ways, I have noticed that prints from a Rep 2 do not have this problem. It does not seem that UM2's have this problem either. But all 3 of the UM1's I use exhibit it to some degree. (And again, to be clear, I'm talking about that last 5% of print perfection) Anyway, I would like to systematically identify and eliminate the cause of this problem. Tonight I tested two possible culprits - oscillations in the build platform/Z axis, and oscillations in the bowden tube causing deflection of the print head: To dampen bowden vibrations, I hung a light weight from the bowden tube - just enough to keep it from shaking while printing. To dampen the build platform, I hung a steel plate from the bed and stabilized it so that it wouldn't swing. It was hung outwards from the machine, so as to bias the build platform both downwards and to one side. The prints are roughly 40mm tall, 15mm-ish diameter irregular columns. Layer height is 0.15mm. Speed 40mm/s All 3 prints look nearly perfect in real life, but the extreme conditions of this photo (a macro shot with an LED flashlight directly overhead) show something fairly clearly - that dampening vibrations in the bowden tube can absolutely reduce some component of horizontal banding. The dampening of the Z stage is harder to judge - it doesn't seem to have made much of a difference in real life, or in the photo. It may be slightly worse. Regardless, a significant component of the horizontal banding remains.

Assigning nozzle temps to a material setting switch maybe?

After some of my bad firmware decisionmaking caused the printer to repeatedly hulk-smash its build platform into its hotend while it was moving around, I got worried that some elements of my printhead were knocked out of alignment. While I was inspecting all of that (turns out everything was mostly ok) , I noticed that the tip on my nozzle wasn't as perfectly smooth as it could be... and it got me thinking about what I could to to change that. Conveniently enough, the XY gantry on a printer is pretty much identical to a precision surface grinding setup in miniature... so I taped some ultra-fine grit (5 micron and then 0.3 micron) sandpaper to the bed, moved the z platform up so that it was just barely touching the nozzle, and moved the printhead back and forth several dozen times. The end result was a mirror polished nozzle tip, which is (in theory) perfectly parallel to the build platform. This has made a noticeable difference in my extrusion quality while printing. Previously, I could see a slight asymmetry to the lines being put down by the printer when doing infill, as the uneven finish of the nozzle dragged more in some areas than others. Now the lines go down perfectly uniform, with a clean, smooth surface. I don't know if this will have a noticeable impact on overall print quality, but I'm sure it can't hurt. This is the paper that I used: http://www.toolsforworkingwood.com/store/dept/THS/item/ST-MAF.XX And the whole operation took less than a minute. This may be particularly useful for those using old or off-brand nozzles with less than perfect surface finishes. This is probably not something to attempt if you do not have sandpaper adequately fine enough to complete the polish.

KISS. Just print a proven direct drive extruder from thingiverse, swap out the stepper for a GM3 from solarbotics, wrap the heater block in some rock wool, print an enclosure for the thing that holds a fan, a trigger, a switch, and a battery. Build it around an arduino if you're trying to show kids it's all easy to do.

What are people using for the build platform with XT? It sticks decently to blue tape, but I'm getting some curling when printing flat bottomed stuff. Would like to find something that it sticks to a little better.

The ooze shield looks way cooler than the vase! Go with it!

You can print reallllly hot things. I've printed at 290° with no problems. You don't need that for ABS, but I guess it's useful for PC? Haven't tried PC. (High temp example: The thermistor fell out today while I was reassembling it, I didn't notice, and I heated it for 15ish minutes without any temperature control... it probably hit 400 degrees, with no ill effects.) It's lighter than the stock hotend by some small to medium amount. It's maybe, possibly, slightly easier to assemble than the stock hotend, though not by much. Particularly the thermistor is a tedious PITA (see above...). Anyway, I would definitely not recommend it for printing in PLA - it just clogs too often when retraction is enabled, because of the incredible stickiness of hot PLA. It is great for ABS, and seems to be great for XT.

Just an FYI for anyone else using the E3D - after resenting it for months due to its ultra-cloggy performance with PLA, it may be finally redeeming itself with printing in Colorfabb XT. Retraction doesn't seem to be a problem, and unlike ABS, XT is pleasant to print without a heated bed.

Yeah the availability of the ceramic UBIS remains an open question. They didn't have it in stock for a long time, but it seems to be back now. I also have the cartridge version and it is indeed much clunkier and probably doesn't work as well in this application. Regarding thermal creep, I can say that so far, it seems to be totally under control. I'm currently 13 hours into a print with a decent amount of retraction and the clamp zone of the printhead still feels to be at room temperature. I also printed the star ornament last night (basically 3 hours of constant retraction) and there was no problem with clogging. Will post more thorough review when I get more prints in. edit: Just completed a 25 hour print with this setup; the print head was as cool when it finished as when it started.

Yeah, this is a problem I've been trying to solve for months. It is hard to eradicate. I've tried several different pairs of linear bearings and none of them have been any less sloppy than the others.

I get much higher quality prints when printing all objects at the same time, and group prints whenever possible, particularly when things are small. From a cooling/curling/deforming from heat perspective, printing two of the same object at once is equivalent to cutting your extrusion temperature in half! Also, there's something awesome about setting up a print, letting it go overnight, and waking up to a machine packed full of beautiful stuff.

Actually, maybe my suggestion of switching to Cura 13.12 for better retraction was wrong; having some problems with that at the moment.

Yeah, I'm printing the star ornament as a test since I just switched my hotend up. Lots of retraction and tiny printed areas. Anyway, I can confirm that 13.06.4 isn't moving in this illogical fashion - it is printing an island, retracting, and then moving to the next island while simultaneously un-retracting. Makes a huge improvement in print quality. I will play around with the minimum extrusion before retraction setting and see if it makes any difference in 13.12. It was previously at the default 0.1mm

What's going on here? If I'm try to describe the behavior, it is: Printing an island, Stopping in the middle of that island and retracting, Moving a tiny amount (<1mm) within that same island, Un-retracting while still on that same island, Then jumping over to the next island. Is this the way it is supposed to be? Is this the way it always was? It seems like it should retract, move to the next island, and then un-retract. I'm getting some pretty crappy results with stringing and blobbing. Will try moving back a few versions in Cura, but anyone else seeing this?

Well, despite my previously stated enthusiasm to test out a more minimal printhead enabled by the crossflow cooling fan, I had resolved to be content with my stock(-ish) printhead and crossflow fan setup as it was, because everything was just working so well and I have +1 other projects I need to be working on instead. Then, last night, during a somewhat challenging print (which I probably shouldn't have tried unattended), the print came loose from the build platform, with printer proceeding to knock it around the build chamber for 5 or 6 hours, until it "completed". At some point during that beating, it must have connected hard enough that the brass tube of the stock hotend actually cracked - because this is what I woke up to: Yikes! Anyway, I luckily had an entirely new printhead+hotend assembly waiting to be installed - one that uses the Printrbot UBIS hotend. After doing a bit of searching, it seems to be one of the smartest, lightest hotends on the market, and looked very well suited to an ultra-minimal printhead design: After setting aside some of my time today in order to make the install happen, I am excited to say that it works! And it seems to work well. I was worried that there would be thermal creep through the PEEK insulator into the PLA printhead, but after a few hours of printing at 230°, it seems to be stable. I need to do a lot more printing before giving it a more comprehensive judgement than that, but there seems to be potential here. This may be the lightest and smallest setup possible with commercially available hotends. The integrated ceramic heater enables a very small heat zone and blocks very little airflow from the crossflow fan. The pre-assembled seamless design and teflon liner should prevent any leaking or clogging. After more extensive testing, I will post more photos and all of the necessary files to replicate this setup in next few days.