gr5

That wiring sounds correct. Especially after looking at this picture: To turn on the fans, select PRINT and choose anything to print then immediately go to TUNE menu and then adjust the fans to 100%. The printer will not print until you exit the TUNE menu. When done testing just kill the power. If the fans still don't work I would look at the wiring underneath the printer - you only have to remove 2 screws to remove the (larger) botton cover.

Look at the part in cura "layer view". Is it filling in the holes on purpose? If so uncheck "fix horrible type a' in cura. Please show a photo of your result. For example: May mean different things to you and me. For me completely means you can't even tell where it was supposed to be because the slicer put diagonal lines over it. Alternatively if it is just a bit small then make the hole larger in CAD by say .3 to .5mm to compensate for your result.

Understanding the I value is important here. I is for integral (or sum) of the temp error. To calculate the I portion you find the error (actual temp minus goal temp) and you average that forever. You multiply that by the "I" constant to get how much heat to add. The I term basically tells you how much energy you need to achieve a steady state. It may take a while (a whole minute) for the I term to be accurate. Eventually it is guaranteed to be the exact right amount of energy to keep the nozzle at the goal temperature. Once you get there the P and D terms evaluate to nearly zero. Once you turn on the fan, everything changes. The I term which has finally settled to the perfect amount is no longer valid. The nozzle cools, the I term slowly builds and eventually you settle on the new integrated error with the higher energy. The latest Cura, by default, turns on the fan slowly for this reason. So that the I term has time to adjust to the new, colder, environment. One possible firmware fix would be to print for a while with the fan off and then print for a while with the fan on and somehow read back the integrated error for both of those states. Find the difference in integrated error (think of it as the "base" power level needed to maintain temp with fan on and fan off). Then everytime the fan is switched on or off, you cheat and instantly bump the error by that amount. That would require some changes to the code. I don't think this is critical as I just turn the fan on slowly. The existing PID values are properly tuned and work quite well (but not for E3D nozzle) so I wouldn't mess with them other than maybe expand that 10 degree range of where they are used.

It's a 2 year old machine. It's out of warranty. They are very busy. Eventually they should reply to you. Also you will get better response from them *after* you "probe" your machine as you will have more information for them.

Daid the final object to be printed should be allowed to be as tall as you want. The gantry won't hit the last object. Only the objects before.

Really? That's enough to get the filament off the top of the bowden (at the top of the arc) and resting on the bottom of the bowden (at the top of the arc)? I need much more. Closer to 4mm.

Leave combing on then. Also change 42mm to 40mm/sec. If you still have the older UM firmware it could make a HUGE difference. If you have the latest firmware it will make no difference (latest caps the speed at 25mm/sec - older firmware does bizzarre things).

The stresses are where the bowden and wires meet the head and so that's almost always where the wire breaks. Push on the wiring near the top of the head and push the head around. If the temp changes suddenly (more than 5 degrees in 1 second) then it's the sensor path and can't possibly be the heater. If it doesn't change suddenly then it could be lots of things. Once you get it to mess up push many times in different spots to get it to repeat and eventually you can often find the spot that is bad and replace that wire. There's a similar but I think different common problem when the fan wires are too close to the other temp wires - before the probe is amplified. The signal is very sensitive to noise from the thermocouple but once it goes through the tiny amplifier board it is not sensitive so keeping the fan wires and the thermocouple wires at least 10mm apart can help a lot. Do you know what kind of temp probe it is? Is it a thermocouple? Or some other design (thermistor, etc)? The UM Original has a thermocouple. And if thermocouple, where is the amplification board? On the print head like the UM Original?

Oh - and while you are in there messing around you might want to increase the max temp so you can go over 275C.

There are a few settings to deal with this. First of all I would cut P,I and D in half since you want less power. Especially P. That should help a little. Maybe cut it in half again if it still overshoots. You can play with these values with the ulticontroller but it's so much easier with pronterface: printrun/pronterface/prontrface download: http://koti.kapsi.fi/~kliment/printrun/ Then you can also limit the max power setting for the heater if you want. Unfortunately you might be forced to do that. #define PID_MAX 255 // limits current to nozzle; 255=full current #ifdef PIDTEMP //#define PID_DEBUG // Sends debug data to the serial port. //#define PID_OPENLOOP 1 // Puts PID in open loop. M104/M140 sets the output power from 0 to PID_MAX #define PID_FUNCTIONAL_RANGE 10 // If the temperature difference between the target temperature and the actual temperature // is more then PID_FUNCTIONAL_RANGE then the PID will be shut off and the heater will be set to min/max. #define PID_INTEGRAL_DRIVE_MAX 255 //limit for the integral term Above is a section from Configuration.h. You can limit PID_MAX to say 128 instead of 255 so it will put out half the power - this will make it heat up half as fast. You can also open up the range of 10 degrees to something more reasonable - by the time your head is 10 degrees below the goal temp it's too late for the PID controller so make that 20 degrees. That alone should be a huge help. Unfortunately doing these Configuration.h changes requires a recompile of Marlin: First get the source code files here: https://github.com/ErikZalm/Marlin Then edit Configuration.h – this is by far the hardest step and it's not bad. I recommend you go here: http://marlinbuilder.robotfuzz.com/ and use that website only to get the Configuration.h file. Then run winmerge or some other diff program to compare the latest ErikZalm version of Configuration.h to the one from robotfuzz which is usually a few months behind. Edit the ErikZalm version to match the robotfuzz generated version. Make any other edits as necessary. It sound complicated but it is extremely clear and well commented. Sometimes with paragraphs of explanation. configuration.h file detailed explanation: http://airtripper.com/1145/marlin-firmware-v1-basic-configuration-set-up-guide/ Then you need to build Marlin. There are instructions that come with the erik zalm download in the "README.md" text file. Basically you download and install arduino ide: http://arduino.cc/en/main/software Then copy the sanguino software as explained in README file. Open Marlin.ino file in Arduino IDE by double clicking it (not pde file as stated in README - I think that's old). Select board as "Mega 2560" as explained in README file. Go to "file" "preferences" and select "verbose output" so you can find your hex file. Then build it by clicking the check box in the upper left corner. At the bottom you will see it compiling Marlin. At the end of this it says where the hex file is. If you are currently connected to your UM through USB you can just click "file" "upload" and you are done! But you should locate that hex file and save it somewhere along with the Configuration.h file used to create it so you can recreate the same version with maybe one change. Also you can upload the hex file using Cura in expert menu. Alternatively you can build Marlin with somewhat more detailed step by step instructions the command line way (which I don't prefer): http://www.extrudable.me/2013/05/03/building-marlin-from-scratch/

Sounds like a Marlin issue. First try reducing your print speed by 2X so that Marlin doesn't have to send so many steps per second. I believe Marlin can only handle one axis going at 300mm/sec or something like that and any more than that causes the cpu to have trouble keeping up. So if 2 axes are going over 75mm/sec and you have double stepping you might also go over this limit. Not sure. Just a hypothesis here. That's all I've got. Printing too slow over USB should just make it stutter and print slow - not a jagged path. I would think. Whatever the issue is, I think it's firmware related.

gantry is wrong. You can print parts MUCH taller than 15mm without hitting the X/Y gantry bars at the top when you home again. Please be aware that you do not need to print to see if it is printing one-at-a-time - simply go to layer view and you will see only 1 part printing when you show layer 1. Second part won't start until first part finishes. Right click (edit - actually I think you hover!) on these settings to see the default settings. If any part is taller than the gantry height it will not print in one-at-a-time mode.

printbl also has high impact resistance PLA: printbl.com

Excellent!

Did you check "retraction"? There are 6 retraction settings... 1) Make sure retraction is checked on basic/quality 2) In expert settings set minimum travel to 0mm or at least something small like 1mm. 3) In expert settings set "minimal extrusion..." to 0 or something well under .1mm. There is a 160X factor on this for .1mm layers so if you extrude .02mm filament you will get 3.2mm of straight line extruding. So the default is to not retract if you print for less than 3.2mm of linear printing (for .1mm layers). 4) In expert settings you can set "combing" on or off - it shouldn't make a huge difference for this part. Lately turning this "off" seems better in the current version of Cura but in the next version it might be best to turn it back on. 5,6) For your UM Original, you can also control the amount and speed of retraction. I recommend 40mm/sec (which works well for old and new Marlin firmware) and if you printed this part then 4.5mm retraction should be about perfect: http://www.thingiverse.com/thing:46157 Without the above part you need at least 5.5mm retraction distance. 5,6) On UM2: On the printer, you can control retraction there also Leave the retraction speed alone but change the retraction amount from 4.5mm to 5.5mm if the bowden moves up and down at the head (try pulling it a bit).

show off!

First be aware that only one axis has to have backlash in order to get gaps between infill and border. This post gives you some ideas of why (see post #7 and #8): http://umforum.ultimaker.com/index.php?/topic/1872-some-calibration-photographs/?p=14474 Backlash can be caused by too much friction or too much play. Too much friction because there is still quite a bit of pulling in the belts when the head grinds to a halt and changes direction without getting to the goal. Too much play causes it because the motors are getting to the right spot but not the nozzle. Turn off power and test for too much friction. Push the head around with your fingers. It could be your long belts are too tight or more likely your end caps that hold the rods are pushing too hard against those rods and they are having trouble spinning. Consider loosening a few if friction seems high. You should also oil all the rods (6 rods in the top area and 2 vertical Z rods) with a light petroleum oil hopefully with very few additives (WD40 will work but I recommend you find something without powerful additives). Just a light oil. Even baby oil is probably better than WD40. Check for play. Hold your coupler still if possible and push on the head and also push on the nozzle itself. Sometimes the nozzle is loose in the head. Sometimes one of the 2 linear bearings in the head doesn't fit right and there is a bit of play. The amount of play you are looking for should be visible and you should be able to feel it. It's the same amount of gap you are seeing on your infill-to-wall-gap. It could be something as simple as the head being too loosely put together and need some tightening down. It could be loose long belts. It could be many many things but it shouldn't be too hard to find.

What speed did you print at? Keep in mind you did 60 from horizontal but you can get almost horizontal. You can print about 10 degrees from horizontal. The surface will be very rough but you can do it.

Hey good luck! You could just try printing and posting a picture if you have issues... The assembly instructions are here: http://wiki.ultimaker.com/Mechanics_build_guide Download and run Cura if you don't have it yet: http://software.ultimaker.com/

Do you have an Ultimaker Original? This is very common on the UM1. It often starts happening after the printer is many months old. Usually the problem is in the wire on the print head. You can heat the nozzle to say 100C and then push the head around to the 4 corners while watching the temperature and duplicate the problem. Also you can duplicate it by pushing on the wires on the top of the print head. The easier fix is to use the extra cable already installed - the cable hanging loose that isn't used - that is meant for a second nozzle. Unplug the current cable and plug that in at the head. Do the same thing underneath the ultimaker at the PCB. But this time make sure the wiring goes through the black strain relief shaped like a F so it won't break again.

Just move the glass a little closer to the nozzle. Or alternatively you can increase flow on the first layer - maybe 150%. If you increase too much the pressure will build up too much and the stepper will skip. Same problem if you move the glass to close to the nozzle. Also don't do .1mm bottom layer thickness as that is too difficult to get perfect. .2mm bottom layer is a great compromise.

This is probably a UM Original. On the UM Original there is a tiny circuit board on the print head. The voltage coming from that circuit board should be 0V for 0C and 5V for 500C and linear in between. So at 20C it should be at 200mv. Now you can determine if the problem is at the head or somewhere else.

You are looking for the scattered dots to disappear. Daid wrote the program so obviously he knows it better than me but I thought if infill was 0% you might not get those "scattered dots".

I see what you mean. That would be very simple for this part - just make a 3.3mm high cube. But it would a bit difficult to remove without damaging the part. I think you are still better off using meshmixer and it's AMAZING support algorithm and you can place supports manually with meshmixer and it's so fast. Click click click and you have 3 supports. And it's free.

Except that now you have to re-level every few hours for about 10 months as it takes about that long for the arms to stop drooping. But I agree - it sounds like a great modification.