gr5

Illuminarti is correct. However, if I was determined to print this and I had created the CAD model, then I would print out each leaf seperately and flat and then peel them off the bed and attach each one to the trunk separately. With glue. This is a difficult print. Shapeways could print it if the leaves are thick enough. They have a FAQ that states the minimum thickness. I think it is somewhere around 1mm so as long as the leaves are 1mm thick shapeways should be able to print this just fine. Probably cost you about $50. If you upload the CAD file they tell you right away the cost.

I think it's a little of two things: belt a little too loose (backlash/play) and a little bit underextruded. I would also recommend increasing flow to 110 or reducing filament diameter by 10 or 20% area (multiply by square root of .9 or .8). Alternatively reducing speed or increasing temp should increase flow a little. But I would stick to fast printing and instead increase flow.

Well not "at the same time" but you do get it on the same model. Because of the lag or delay, as you slow down for a sharp vertex (corner) you get some overextrusion as the extruder is slowing down but there is still pressure in the head. Getting overextrusion before the corner means you must get underextrusion somewhere else. The somewhere else is as you start the next corner - again there is a delay so the pressure isn't building back up fast enough in the nozzle as it starts the next line. As long as you print at typical speeds the delay isn't a problem. The slower you print the less under/over extrusion you get. If your part overall is "underextruded" it will be more noticeable *after* a sharp corner. If your part overall is "overextruded", well this is more rare, but it will most likely happen just *before* or *at* a sharp corner.

Do you get a lot of stringing on this print or is it not a problem?

The link to the pictures works great!

1) Marlin has "software limit switches" set to 205, 205 (in x,y). If you hit one of these limits, the print will do this kind of shift so this is possible. 2) Which axis slipped? X or Y? This is very common if your pulleys slip. I would tighten all 6 pulleys for the slipping axis and mark all 6 pulleys and rods with a permanent marker so that if it slips again you can see if it was a pulley and which one. Make sure you get all 6! There are 4 pulleys on the long belts and 2 pulleys on the short belt (for each axis). These setscrews for the pulleys can get loose after many months with no trouble. Even though there is no "noise" it happens very fast and would be impossible to spot a pulley slipping. You could hear it if you went back over the video (sound of stepper moving but with no movement of print head briefly).

Great question. The extruder will automatically adjust. All moves are performed by the Marlin firmware which is pretty good stuff. It is pretty smart. All of the moves in gcode are X,Y,Z,E (E for extruder) moves - they are 4 dimensional or 4 axis moves (actually typically Z doesn't move for a large percent of the moves). These moves are required to be linear moves. This means you can't move the X at the same velocity as Y because the distance to cover may be different. So to move in a "straight line" you have to make sure the X,Y,E velocities are all at the same ratio as the total distance to cover. This means as the X or Y movement decelerates, the E (extruder) also decelerates. Unfortunately, because of the long bowden tube, and because of the somewhat high pressure of plastic in the nozzle, there is a lag (time delay) in extrusion. So what happens is you get a tiny amount of overextrusion as it slows down and underextrusion as it speeds back up. This is mitigated with slower print speeds. It's rarely a problem though. But when I look at a print of a cube I can tell if the print head moved clockwise or counter clockwise because of this "delay" or "lag".

Holy crap. I never thought of that. This is why I read these posts every day! Something so simple and obvious... Now to see if my webcam comes with software to do stopmotion. Do you recommend 1 frame per second maybe?

This looks familiar! It looks like bed issues. Most likely height or Z issues but it could be so many things. Usually Z screw issues repeat every 3mm (one rotation) but not always. Did you put the green grease on it? I bet it looks better with .2mm layers and I bet .1mm layers looks much worse. But I could be wrong on this. So some ideas: 1) Play with moving the bed up and down - a lot! do it by hand (turn off steppers) and spin the Z axis up and down many times. See if you see the bed get stuck and then slip suddenly. See if things get tight near the bottom of travel (bottom 2 inches) which can be caused by bed not assembled quite right. Z nut (that huge nut in the bed) should be able to slide slightly from side to side, but maybe it's better to glue it in place. 2) Look at where the Z screw is connected into the coupler at the bottom. It should be all the way. Is it possible that it's loose or not all the way into the coupler? You shouldn't be able to see where the z screw gets thinner at the bottom. 3) is the bed hitting anything? Wires? steppers? Is there anything that might make the bed tilt, or get stuck slightly, or reduce it's speed. 4) Are the 4 screws agreeing? If you push on the plexiglass corners, does it wobble like a chair with one short leg? Is one screw not being used? 5) Move the z axis with the stepper. With Cura or Pronterface (not with ulticontroller!). Move 100mm at a time. Anything? Could your acceleration be set too high? Did you mess with it? The default with cura is 100mm/sec^2 you can check with the ulticontroller or if you don't have one use pronterface to query the Z acceleration. Mine is set to 200mm/sec^2. You can go quite a bit higher but 1000 or 2000 mm/sec^2 acceleration can cause this pattern due to skipped steps. Put a ruler or something taped or clipped to the side of the UM near the bed. Put a pencil or flat ruler or something on the bed so it just touches the ruler. Move by 10mm or 1mm and look carefully. Very carefully at how much it moves exactly.

#1 Keep track of every setting. Whenever you save a gcode file the settings that you used get saved along with it at the end. You can later choose "file" "load profile from gcode" to see what settings you used at the time. Write down the most important settings: speed, temperature, layer height. I recommend you stick with .2mm layers for 99% of what you print. But vary speed and temperature often depending on your needs. Everytime you start a print have a notebook where the student must write down the date/time which gcode file was printed, color of filament, and a few details like the 3 parameters I mentioned above. #2 I don't know if there is something that explains them. And they change when you start printing (tune menu). Don't mess with jerk. lol. It's fine as it is. In physics velocity is the change in position (derivative). Accel is change in velocity. Jerk is change in acceleration. But Everything is discreet in the software (not continuous) so jerk is always infinite at the vertex of every move (theoretically). So instead they created something called "jerk" but it's not jerk. It's a max delta velocity vector at the vertex. So if X velocity is 100mm/sec and y velocity is 0 and you hit a 90 degree corner at that speed, jerk would be considered to be sqrt(100^2+100^2) or 70.7mm/sec. Since max jerk is set to something like 10 or 20 mm/sec the X axis is instead slowed down to the speed =(max jerk*1.404). This is for 90 degree corners. for only slight angle change the UM doesn't slow down as much because the velocity change at the vertex is smaller.

oh! So that probably means the old ultimaker board was breaking that pin - actually probably the X servo driver was destroying that arduino pin! You could possibly reroute using another signal. You would have to make your own custom drivers that use a different arduino pin.

There shouldn't be any stringing so in cases like this definitely go with hot and fast. There shouldn't be accuracy issues. Going fast has more to do with stringing and underextrusion. The UM can do 300mm/sec if the damn nozzle could keep up (which it can't). How do you know if you were in bed when it popped off? That is most likely the problem why it popped off but if you had used ISA it might have hung in there (and brim).

Also I am not patient so I would print this at 240C and at 120mm/sec.

It looks like the part moved off the blue tape. Is that true? There are tricks to keep a part stuck to the tape. The most important trick is to use isopropyl alcohol. Blue painters tape has a kind of waxy non-sticky stuff on the non stick side of the blue tape to keep it from sticking from the sticky side of the blue tape. You need to remove this stuff with isopropyl alcohol aka rubbing alcohol found in any store that sells bandages such as a pharmacy or drugstore or supermarket or convenience store. The next helpful thing is to use the brim feature. Do 6 to 10 passes of brim. This helps quite a bit. The next thing is to make sure your first layer is getting squished into the tape. The very first layer - the top should be very very flat. Like a pancake. Each thread, each trace should be clearly wider than it is tall. It should be pressed into the blue tape. There are more tricks but if you follow these 3 it should be good. Having said that - the way the fins flare outward is a bit tricky - due to pla shrinking they will be pulled *upwards* and the head or the fan shroud might hit those fins. Once you get to the part where the fins start going vertical your print can be left unattended. One solution to that upward pulling due to shrinkage is to add support. Another solution is to print in a heated chamber but the support will help much more than a heated bed or heated chamber. You should be able to print this without any heat. Another solution is to have a putty knife in your hand and every 4 layers press down on those fins. Again once they go vertical you can walk away.

mic6 aluminum is... well... wonderful. It's very very flat. But make sure you always put kapton tape on it and then blue tape on top of that potentially. This is because you don't want to ever scratch it (aluminum is softer and easier to scratch than steel or glass). If you get an aluminum plate do not get any other type of aluminum. Get mic6. Look it up on wikipedia. However you might also have warping/bowing in the hardware that moves the head around! An alternative would be to characterize your bed and adjust all the Z values depending on XY position. This would be a great plugin - it would add a Z coordinate to *every* move that has only X and Y. And adjust Z by up to .2 or .3mm depending on the XY location. Wouldn't that be amazing? The plugin itself would be quite easy to write. The hard part would be characterizing your bed.

Interesting. You might want to consider removing the fan wires such that they are farther away from the temperature wires. I have gotten maxtemp with fan power off. So for me it wasn't the fan. I can duplicate it by jiggling the temp sensor wires. I found it happens more in the corners of the UM where it stresses the wires the most. Anyway I improved the stress relief and switched to the other cable and we will see what happens.

I'm jealous!

The black part in this silver pulley is the setscrew: http://foehnsturm.com/3dp/ulitmaker-pulley-vs-lathe-pulley.jpg There are 6 pulleys on each axis X,Y - 12 total. Each has 1 setscrew. You need to tighten all 6 X axis screws and maybe all 6 Y axis screws. Tighten them very hard.

I also want to know what version of Cura you have. Assuming you are using an Ultimaker then you probably have a .4mm nozzle so you need to change the wall to .8 (not .9). Using .9 with nozzle set to .4 will cause problems (like the loops or "rings"). Wall should always be a multiple of actual nozzle size. The stringing is mostly between support. There may be a bug regarding support in Cura where it relates to retraction - not sure. There is a new secret beta version of Cura that has better support features. You can try it here: http://software.ultimaker.com/Cura_closed_beta/ I'd like to see a better picture of the "rings under ears".

Illuminarti's 3 suggestions are good but the most important is this: Wipe your blue tape with isopropyl alcohol. You can find this product wherever people sell bandages.

It sounds like the PCB is fine. So probably the Arduino is bad. Did you say you tried 2 different Arduinos and they both had the same problem? Or was there only one Arduino? Or did the first Arduino have a different problem? I think you should buy another Arduino board and also just to be safe, I would remove the X stepper driver. I would test the voltage on the DIR pin on the X stepper and only if it works, *then* put the X stepper driver back in - there is a small (very small!) chance that the stepper driver damaged the Arduino. This is pretty unlikely as the Arduino is designed to not get damaged easily.

I haven't printed with ABS but there are lots of ways to get it to stick better. Google "how to make ABS stick to print bed" or something similar and see what you get. Certainly I recommend "brim" more than "raft". When plastic lifts up off your print bed it starts at the corners due to pulling/cooling of upper layers. If your print has corners the problem is much worse. If the print has rounded corners it isn't as bad but if you have the brim feature turned on it helps even more and keeps the corner from lifting in the first place. Once the corner starts to lift it's easier for more of your print to start lifting. Read about "glue sticks" and "abs glue" and other techniques.

Maybe your arduino board is loose or bent pin? I think you should create a ticket for Ultimaker Support.

Nice job debugging! Can you send the board back? Is it under warranty? If it is new you can do a support ticket and they should send you a new board out right away as this is a clear problem. Also did you load the Marlin firmware that comes with Cura into your Arduino? There are many configurations for which pin of the Arduino controls which stepper pins and so on. I would also cut power and remove the arduino and check the signal path from the arduino connector to the stepper driver. Did you check this signal on the stepper driver or the arduino? It might be a bad connection between the arduino and the ultimaker PCB. And the stepper driver may be gently pulling that signal to 5V.

Assuming there is nothing wrong with your ultimaker (and there might be..), then you should also be aware that the faster you print, the higher the pressures get in the print nozzle so that when you want to stop printing, it keeps coming out. The way to fix this isn't necessarily to print colder because then the pressures get higher. On the other hand colder means "more like toothpaste" so colder can help. A better way is to print slower. Having said that, 210C at 40mm/sec with .2mm layers is reasonably slow enough that you tend not to get too much leakage but if you are unhappy with the quality then try 25mm/sec or 15mm/sec and just be patient and wait for that print. Also there is a feature that should be on by default i believe called "retraction". This tries to immediately remove all the pressure in the nozzle and is used when the head is jumping over a spot that should have no extrusion but is outside the print. When retraction isn't working quite right we call it "stringing".