Anders Olsson

Hi! First, make sure you have updated your machine with the latest firmware: http://software.ultimaker.com A lot of issues regarding z-calibration and filament feeding has been resolved past weeks, so it is really worth updating. Make sure your filament has a diameter smaller than 3mm, that it does not have flat parts from grinding and that there is no contaminations in the bowden tube. Your problems sounds a bit like when my teflon insulator did not reach the nozzle and half molten filament ended up in the space there causing excessive friction: http://umforum.ultimaker.com/index.php?/topic/5953-idea-on-underextrusion/&do=findComment&comment=55227 I bought my machine almost half a year ago and I had to do some minor adjustments for it to work good, but it was not scratched or damaged in any way.

Nice!! I have been thinking of designing something very similar to that for my UM2, but I did not find the time yet. Is there a spring somewhere pushing one of the feeding rollers towards the other one? How does the filament look after passing through the feeder?

Excellent! Your findings in the code seems consistent with what I have been measuring with my digital caliper the past weeks.

I mostly measure the thickness of the prime frame that prints before the print starts with a digital caliper. To get better priming of the head I have been printing a double frame lately. It is obvious too when I look at the first layer that it prints way to close to the platform before power cycling. It takes several layers of printing before the excess plastic from the first layer is hidden within the printed object. The outer dimensions of the first layers of the print will also increase due to the excess plastic. Last few days I tried calibrating the nozzle slightly close to the platform to get the plastic to stick better. I usually get like 0.3-0.35 mm after power cycling and now it is like 0.25 mm. The distance stays consistent at 0.25 mm after power cycling, so on my machine it seems everything works like it is supposed to after it is power cycled. If no one is experiencing this problem, maybe it is just something with my machine. Don't waste too much of your time with this in that case. I can live with it as it is now :smile:

Yes, this issue seems not to be directly related to the Z-homing issue. I think my machine always has been like this, but before the 14.06.1 firmware it was not as consistent, so it was difficult to figure out exactly what was wrong back then. It is not a major issue for me now that I know I have to power cycle after leveling the build plate. However, I might not be the only one with this issue so I thought it might be good to tell people here. To duplicate it, just level the build plate with a 0.1 mm paper and print something which has 0.3 mm initial layer thickness. Power cycle and print the same thing again. On my machine first print will have about 0.1 mm initial layer thickness and after power cycling it will be 0.3 mm (as it should be). My machine was delivered in mid January and has serial UM2A05-MES1337xx

Hi! Sorry, I accidentally wrote 13.06.1 in the previous post, that is wrong, here is the info: Jun 5 2014 09:13:03 Version 14.06.1

Yes, it seems so on my machine.

I have tested the 14.06.1 for some time now and have come to the following conclusion about the Z calibration on my machine: If I run the buildplate leveling procedure using a 0.1 mm thick paper for the Z-calibration then the first layer will print 0.1 mm thick until I power cycle the printer. It does not matter what I do to the printer, like pushing the platform down when in lower position, running lower buildplate and so on. The first layer will still be 0.1 mm after leveling until i power cycle it. After power cycling the first layer is suddenly 0.3 mm and stays like that no matter how many times I power cycle it. I have already reset the default settings on this firmware a few times so suppose it should be properly initiated. The models I have tested so far are supposed to print 0.3 mm initial layer. I have not tested 0.2 mm initial layer yet, so I don't know what will happen then.

If the diameter never went above 3.00 mm and dust did not exist, it would work fine. The feeder on the UM2 is quite hard on the filament though. It produces plastic particles which every now and then ends up in the bowden. It also tends to grind the filament from time to time, which smears the filament out and increases the diameter at that position. Filament change will often leave some plastic threads in the feeder/bowden too. Altogether this means you want at least a 0.1 mm clearance which makes 3 mm the upper limit if the bowden is 3.1 mm. I have both the 2.85mm Ultimaker filament and some 2.95-3.00 mm filament from another manufacturer and the extra clearance of the Ultimaker filament makes it much more reliable. When you reach 3.00 mm even slight disturbances tends to cause underextrusion or failed prints.

Concerning the filament grinding, Would it be possible to lower the motor current at the fast forwarding part of the priming cycle? Or even better, when the filament is fast forwarded in general? That way you could make sure the filament has reached the head with no risk of grinding it down. You could then run the final part of the priming at a low feed rate so that the filament has time to melt and thereby avoid grinding. My Z-axis calibration is still a bit strange with 14.06.1 by the way. The gap between the head and the platform still increases with about 0,1mm at some point after the calibration was done, possibly at first power cycling. But it then stays at that distance it seems like, so I can accept it as it is now.

You might be onto something here. I took apart my nozzle assembly to change nozzle and with the new nozzle I suddenly had some extrusion problems which I have never seen before. Since I used the same components (apart from the nozzle) and the same filament as before I could narrow it down quite quickly. The symptoms were that the filament loaded okay but when I started to print it did not extrude well. Pushing the filament through the nozzle by hand was also harder than it used to be. It turns out that the teflon piece on my machine has such a tight fit that the spring was not able to push it all the way down to where it meets the upper part of the nozzle. When there is a gap between the teflon piece and the nozzle part, half-molten plastic will collect there and increase friction. It gets particularly bad after retraction. The molten plastic also seems to catch on to the filament when retracting, possibly pulling the teflon piece even further up. I took everything apart again, cleaned the components and grinded a few hundreds of a millimeter of the outside of the teflon piece to make the fit less tight. Now it prints just like when it was brand new again :smile: I think it is really crucial that the fit of the teflon piece is not more tight that that the spring always is strong enough to push it firmly against the upper part of the nozzle. I imagine that if the teflon piece change diameter due to the heat near the nozzle it can possibly grab on to the filament on retraction and then not get back to it's original position due to molten plastic entering the space.

Since a lot of people asked about this I made this video:

I had to take apart my nozzle assembly yesterday so I did some measuring of the components. The outside diameter of the thick part of the teflon piece is about 11.1 mm when the nozzle has been at 260C for some time compared to 11.0-11.05 mm when the nozzle assembly is taken apart. The temperature of the teflon piece is about 50 C when the nozzle is at 260 C. I can not see any major gradient in the temperature in the teflon piece with my thermal camera, but the resolution is a bit low for this kind of measurements. The force of the spring is about 1.5kg when compressed as when assembled, measured using kitchen scale. The lower part of the teflon has a tight fit in on my machine, so if it has to expand it can not increase in diameter.

I have noticed a few other bugs in the firmware lately, which are not related to this topic. What is the preferred way of reporting these? Should we have an official bug report thread for each upcoming release of the firmware?

Interesting. I have been trying to figure out what is going on with my UM2 and the bed leveling for some time, but I did not investigate it this far yet. I never had the nozzle hitting the glass yet, just the opposite problem (printing in air). On my machine the leveling stays fine after calibration as long as the power is on. After power cycling the gap increases 0.1-0.2 mm or so, enough for the walls of the first layer not to stick to the platform, ruining the print. I was almost about to ask for a tune-option for fine tuning the Z-calibration on the fly when the print is started. It would be more convenient than having to abort the print and go through the leveling cycle every time the printer has been powered off. If there is a clear cause for this problem and it can be fixed it would of course be even better though :-)

I have tried cleaning my UM2 nozzle in acetone in ultrasonic bath and the ultrasonic did not really seem to help that much. Based on my experience, ultrasonic cleaners are best at removing hard materials from hard materials. Soft materials, like plastic, tends to stick to the surface and just absorb the ultrasonic waves. However, taking apart the UM2 for cleaning the nozzle is completely unnecessary, here is how i unclog it now days: 1. Heat the nozzle to 250C and push a wire form a wire brush through it from below to push the dirt inside the nozzle. 2. Remove bowden and gently push some ABS filament through the nozzle by hand. If it clogs again, repeat step one. 3. Change to nylon filament and gently push filament by hand until pure nylon comes out. 4. Lower the temperature to 130C, wait a few minutes and then gently pull/twist the nylon filament to get it out. (It is possible that you can skip the ABS and go directly on the nylon, however I had other reasons to start with ABS so I did not try that yet) The problem with all other cleaning methods is that there is a seat inside the nozzle where the clogging particle tends to hide. I have not managed to clean this space properly with any other method than using the nylon filament. The nylon filament comes out like a nice casting of the internal space of the nozzle and often there is particles inside the filament where the arrow points: Maybe I should also mention that I did not have any clogging with normal filament yet, all clogs was due to my home made experimental filament. When running normal filament you should not experience clogging very often.

Here is how the nozzle assembly looks at the UM2: The yellow part is threaded onto the nozzel part. The two horizontal holes in the nozzle part is for the heater and temperature sensor.

I am also very interested in this! I have been printing some custom made materials for research projects lately which has attracted some attention. While single extrusion of these materials is already interesting, the dual extruder kit would open up completely new possibilities. If there would be a possibility to get a pre-production version of the UM2 dual extrusion kit, it would be very interesting for me.

I found a "window scraper" in the hardware store and it quickly became my general tool for removing printed parts from the platform, removing excess glue from the platform and for cutting/scraping excess plastic off the printed parts: I mostly use the blades without the holder and they are very nice for making precision cuts in plastic. You have to watch your fingers though :smile: The "Tesa Easy Stick" might not be considered as a tool, but I included it because it works much better with ABS for me than the "Staples glue stick" provided with the Ultimaker 2 did. The main advantage with the Tesa glue is that it is harder and therefore it is much easier to apply a very thin, even, film of glue to the platform. The provided "Staples" glue tended to come off in big chunks, which made the bottom side of the printed part uneven. The Tesa glue also dissolves instantly in hot water, which makes cleaning the printed parts and the platform very easy. I also found that a vacuum cleaner is really handy to have next to the printer. I don't know if the vacuum cleaner might be clogged by glue in the long run, but it is very convenient for keeping the printer clean.

If I designed something, it would probably be inspired by forest machine harvester heads: The filament feeder is somehow a similar engineering challenge, just on a slightly different scale :grin: I made a quick draft with standard gears and bearings from RS Components, just cutting a trace for the filament: I would personally like this kind of design, where both feeder rollers/gears are powered and there is a large surface area for the filament to rest on and at the same time sharp edges for traction. Now, I did not think this through very much, so considerer it as a wild idea, not anything you should go and build before thinking it through extensively.

Just out of curiosity, I ran the extrusion test with ABS at 200% speed to find the flow limit for ABS. (I am sorry for torturing you, my dear Ultimaker :mellow: ) Interestingly, it was okay up to 12mm3/s (6*200%). Then is shows very similar behavior as PLA did: When checking the feeder, it is obvious that the knurled gear is starting to slip at high speeds. It is not skipping steps, but the marks from the knurled gear on the filament is smearing out more and more the higher the speed is. The nozzle temperature stays stable at 260C throughout the print according to the control panel though. Therefore, I guess that the maximum flow rate depends on the flow characteristics or thermal conductivity of the plastic rather than the feeding mechanism or the nozzle heater. Now, I promised the professor to print a number of things which has to be out of white ABS coming week, so I really have stop using it for tests before I run out of it. :grin:

I have done some tests now. The ABS came out perfect on first try. When trying PLA, I suddenly realized why I never switched back to printing things in PLA after I tried ABS.. PLA certainly has a more sticky surface. I added an extra pulley before my extruder to guide the filament into the feeder, just to avoid potential problems with the spool unwinding: This thing has not been tested for normal printing though. I don't know how it will handle the filament retracting. With ABS I never had problems with this part of the printer. The version in ABS came out perfectly fine on standard settings (260C) as you can see here: When running the PLA at 210C the feeder started skipping and underextruding at 6mm3/s. I quickly increased the temperature to 230C, which I read people here had been using for printing this thing in PLA. It worked until 10mm3/s, when it underextruded again and I increased to 240C. I then tried another time at 240C, but the filament unwinding jammed a bit in the end of the print. Next try at 240C worked well except when going from 9-10, where I had some underextrusion. I then tried at 250C, which seemed a bit better, but still there was some slight underextrusion the first layers at 10mm3/s. My conclusion is that maybe we should have a look at the extrusion properties of PLA, trying to determine if it is really suitable for flows rates this high. Many plastics has kind of a non linear viscosity. (I am not a chemist, there are probably better words for this). This property of the plastic will make the extrusion force increase more in an exponential way than in a linear way, when pushing the plastic through a small nozzle at an increasing flow. This is my current thought about why the force needed to extrude PLA seems to increase rapidly as flow is increased. As I said, ABS came out perfectly fine on standard settings, but ABS is a different plastic too which most likely has different extrusion flow properties. Particularly interesting is how underextrusion occurred for both 240C and 250C the first few turns when going from 9-10mm3/s. I don't see exactly how this would happened if it was a mechanical problem. Looks more to me like there is something going on with the melting and the viscosity of PLA at these flow rates. So maybe in the end you are just trying to defeat the extrusion-properties of PLA when running it at this high flow? For me this is a bit of a non problem in the end, since I am not going to print PLA at this rates anytime soon anyway. ..time to get back to my experimental HDPE extruder..

I had severe under-extrusion on long PLA prints on my UM2 upon delivery. After reading through the 24 pages quickly I did not see anyone mentioning the nut that sits behind the stepper motor for the feeding mechanism (or did I miss that post?): On my UM2, the hole in the chassis plate was not deep enough to fit the nut. This made the stepper motor sit with a slight angle. The resulting angle on the motor shaft moved the knurled gear away from the ball bearing. This resulted in: 1. Less pressure on the filament, making friction in the spool unwinding mechanism more of an issue. 2. The filament not centered, making it grind towards surfaces inside the housing. 3. The ball bearing positioned too far out, making it grind on the plastic housing. 4. The mechanism more sensitive to filament thickness, since there is less play in the bearing spring available to compensate for thinner filament. 5. The chance of the filament getting stuck when slightly grinded much higher. The solution in my case was to simply remove the screw and nut. Since then I have not had any under extrusion issues. Although I have mostly been printing ABS since the repair and under extrusion happened with PLA, so I should try a long print with PLA some day to confirm that it works with PLA too. I also experienced that the noise of the feeding motor decreased significantly after removing the nut. Here is a video of my feeder mechanism, loading ABS, after the nut was removed: As you can see, the ball bearing is pushed in significantly when the filament is inserted. This was not the case before I removed the nut, the bearing was barely pushed in at all then. Regarding the filament spool holder, there was an extra "knob" delivered with my UM2. The manual said nothing about it, but using it as a guide for the filament improved feeding (before i found the nut behind the stepper motor): (the knob in the lower left corner)

telophar My UM2 came with a plastic knob which there was no information about in the manual, but I just assumed it was to be used as a guide for the filament. The feeding, at least for ABS, appears to work much better for me when the filament runs around this knob. I have also noticed a few things regarding the feeding mechanism. - When changing material, the stepper motor apparently is running at reduced current, since it skips steps when the filament hits the nozzle. - On my printer there appears to be a firmware bug, because I have to power cycle it after changing material for the "move material" option to work. - Using the move material option, the stepper motor is much more powerful than when changing material, which means it easily grinds down the filament if one is trying to feed to fast. I am curious to why the stepper motor apparently has much less torque when changing material (at the low speed in the change material cycle). When the filament hits the nozzle, it still extrudes with a much higher speed than for normal printing. Is there a setting somewhere in the firmware where the torque of the stepper motor could be reduced? I mean, when the stepper starts grinding the filament down, it is not feeding anyway, so limiting the torque a bit should not have a negative impact on the feeding i think?

What is your initial layer thickness? I had no luck printing 0,1mm initial layer on my UM2. Using (at least) 0,2mm initial layer makes leveling less critical. When the first layer becomes transparent, the nozzle is too close to the build plate. Running my UM2 that close at large surfaces will cause problems with the feeder. The plastic can not escape the nozzle and the knurled gear starts slipping, grinding down the filament. This makes the feed rate irregular on other parts of the print later on. I have to be super careful when leveling the build plate to get good results. A standard laser printer A4 paper should move with a very slight resistance between the nozzle and the bed for my UM2 to print well. Every now and then I have to redo the leveling process since my UM2 seems to loose calibration with from time to time. One step too high/low when fine tuning during calibration is enough to ruin print quality and reliability on my machine.