gr5

Yes - once air gets under there it's pretty hard to recreate that vacuum seal.

I recommend tinker firmware version 16.01 if you are having problems with any newer versions. I don't understand it (and I'm using the latest) but a lot of people have less crashes and issues with 16.01. It's on the releases page with all the other versions - just scroll down - a lot!

agreed!

There is a "fuzzy skin" feature in cura you might try some time.

Oh. If retractions are off then definitely turning off combing will speed things up further. and you will get more stringing for that "grimy vine" look, lol.

retraction also takes time. So sometimes combing speeds up a print because it doesn't have to retract. Sometimes it slows down a print because it's not going direct from point A to B but sometimes going the long way around (so it can stay within the print). I don't know why combing and retraction are inextricably linked - I wish this feature had 4 modes instead of just 2. But lots of people turn off combing for the bottom layer to avoid ugly lines on the bottom layer. Turning off combing means you get more retractions and fewer extra lines.

If it doesn't crash then it's up to the task. The problem is almost certainly that there are errors in the model. STL format is just a bunch of unordered triangles. Likely you have wall inside of solids or you have holes in your walls such that cura can't figure out what is inside and what is outside. If you look at the model in xray view you may see some red areas - if so then those are problem areas. In addition each triangle has a "normal" which indicates which side of the triangle is "inside" and which is "outside" aka touching air. I would try this - first I would decimate the model a bit so it's easier to work with - this step is optional though: http://www.shapeways.com/tutorials/polygon_reduction_with_meshlab Then I would repair the model: netfabb free repair service is here (but you have to create an account): https://service.netfabb.com/login.php Do those two steps and it will probably slice fine. Or at least the second step. If you created the model yourself and you did it in sketchup then I can give you a guide to repair sketchup models which is a better route (if you plan to make changes to the model later) than using netfabb.

I think this. I think you are printing too hot - the filament needs to cool very fast as it touches the layer below so it stays in place. When PLA comes out of the nozzle it cools very rapidly (milliseconds) and shrinks and comes out like snot - like mucus. It sticks to itself very well (not sure if it's surface tension or just likes to stick to itself like snot). So it's like a liquid rubber band. when you go around those corners it gets pulled inside the part sometimes. to get it to stay in place like mortar on a brick wall you need to either slow down and/or lower the temp. With a volcano it heats the PLA extra well and you can print much cooler. What temp are you printing at? I would lower 10C or 20C down to maybe 190C or even 180C. You might also have to print slower so the filament has a few milliseconds to cool before the nozzle moves away and starts tugging on the filament.

Your Y axis slipped some steps. This is a common problem and easily fixed. 95% of the time the problem is that one of the pulleys needs to have it's set screw tightened. Almost always one of the more difficult pulleys to get to. There are 6 pulleys on the Y axis - 4 on the long belts and 2 on the short belt (I forget but in some cases two of the pulleys are merged into one so maybe 5 pulleys). The ones on the short belt are usually the problem - the one on the motor most likely. You can push the head around until you can see the set screw and get a long hex screwdriver into there. Or worst case you have to remove the cover (only two screws need removing to remove the rear corner cover). Tighten the hell out of these! So tight that the tool is twisting. So tight that if you use and hold an L shaped allen wrench by the short end your fingers hurt. A lot. If you don't believe me mark the motor shaft and pulley with a sharpie. Same with the other pulley on the short belt. See if it moves the next time you get lost Y steps.

Did you use sketchup? sketchup can ruin a model. Try a few things - try looking at the part in xray mode and if you see any red in the ears then your model is not manifold. If you see no red and you used sketchup then click on all the faces around the ears and make sure they are all white. If any are gray or light blue then right click and select "reverse faces". More info about sketchup used for modeling here https://i.materialise.com/blog/3d-printing-with-sketchup/

Well you can make the bottom layer thicker - there's a "bottom layer thickness". I've never had this problem with petg. My problem with petg is it sticks too well. So well that it lifts pieces of the glass off and I get glass shards in the bottom of my petg prints. I use a very very thin layer of PVA (I like to mix wood glue and water but there are other ways to do this). Also stick your petg in hot water and see where it softens and make sure the bed is above this temp. This will allow the bottom layer to warp just the tiniest bit (< 0.1mm) and spread the load (relieve the strongest stresses). Another thing is you want to make sure you are squishing that bottom layer. The more the squish, the better it sticks. But seriously, If I were to print the UM robot with petg, you can pick up the whole printer with the tiny robot and spin the printer around. It won't come off. The glass will shatter first.

There are many possibilities but it sounds like this is happening on the first layer. If that is not true let us know asap. If that is true then your printer is probably smaller than the print. What are the exact limits of your printer? What are the limits you entered into Cura? is it a delta printer (they have 0,0 in the center) or non-delta (0,0 in the corner)? What did cura say the dimensions were of your part (in addition to the dimensions you sometimes have things like skirts and brim that go beyond that).

I can't imagine why the temp would go *down* with temperature. However, there is a common problem with electrical interference. The signal on the thermocouple is very sensitive. make sure those wires are at least 2cm away from other wires to the heater and/or fan wherever possible. Once they get past the little circuitboard the signals are not sensitive to electrical noise. You didn't answer the above question as the answer could explain everything.

Well this guy might be buying a UM2 so it might not matter.

You should be using a 2mm hex key but sounds good. If it works, it works.

Cool. thanks for the posts. Those trees look great - what don't you like about them? Rather than use meshmixer it would be better to use meshlab which can also decimate: http://www.shapeways.com/tutorials/polygon_reduction_with_meshlab Even better - learn the sketchup tricks such that you don't need to pass your model through an extra step: https://i.materialise.com/blog/3d-printing-with-sketchup/ That way you can make tiny edits to your model and not have to pass them through meshlab or meshmixer.

Actually I think it's worse than it sounds - I think this is absolute abort/crash limit and the dial that controls the bed temp only lets you go something like 15C cooler than that. However I bet @amedee would add this feature to his builder if you ask nicely.

He added heat sinks and fans to his steppers. He has reached 80C air temp. Don't let it get hotter than that without moving steppers outside. I believe you can use the existing 4 holes for the X,Y stepper mounting screws and drill a large hole for the pulley and I think it will reach far enough in through the wall of the printer to work without buying any parts. But I'm not 100% sure. If the shaft on the stepper is 1mm too short you would have to buy a whole new stepper.

Wow - that builder got better recently. And it has the crucial features needed for a high temp printer including swapping direction of steppers, setting max temp of nozzle to 400C. However it seems not to allow you to set bed max temp which needs to go to 160C and the default max temp for some marlins (not sure about this version of marlin) is 150C.

Do you have 1) the older UMO with the thermocouple and the circuit board on the head that converts the thermocouple voltage to a voltage readable by the arduino or 2) the newer UMO+ with the white circuit board and a PT100 temp sensor and no circuit board on the print head that converts the voltage What is the reading of the temperature when the "thermistor" when it is at room temperature. There are thermistors, thermocouples, and PT100 (which is technically a thermistor but different enough that it deserves it's own category. Thermocouples are 2 pieces of different metal touching and this creates a battery like voltage that changes with temp. Thermistors change resistance with temperature change. Most can't handle temps above 250C. Actually most can't handle temps above 100C. Thermistors vary widely. every manufacturer of thermistors has a different table that translates resistance to temperature. PT100s are consistent across all manufacturers because they are simply made of platinum. also pt100 can handle much higher temps.

You should get zero stringing between prints. What is your: 1) print speed (faster speeds means more pressure in the head which means more stringing) 2) temperature (higher temps means less viscous filament) 3) filament type (some filaments print better than others - PLA is the best) 4) filament color (some colors string more than others - white is particularly bad and I usually print it 10C cooler than other colors if I'm getting stringing) 5) printer/nozzle/core (a 0.8mm nozzle will drip more than a 0.4) A 3dsolex nozzle heats the filament more efficiently so you need to print cooler with those nozzles. Increasing retraction is most likely only going to make things worse. You want to retract only enough to release pressure - if you retract too much then air gets into the nozzle and that causes all kinds of over/underextrusions and messes everything up (too much filament in some places, too little in other places). Maybe you have retractions disabled or aren't always occuring: in cura layer view check the blue checkbox (travels) and uncheck all the rest. The lighter color blue is for travels with retraction. Maybe some of them are dark blue? If so you might need to tweak your retraction settings. If all the travels between parts is light blue then I wouldn't mess with retraction settings.

more info here: http://ultimaker.com/ER02 This is almost certainly a problem at the bed and not the board. When things heat up they expand. A contact along the temperature sensor path is a bad connection and when things heat up and expand it loses connection. Slowly. As it loses connection the resistance suddenly rapidly increases and that looks to the logic just like increased temperature. But it's not increased temperature - it's a slow motion disconnection. The problem is most likely right where the connector is. Also it could be a trace on the board or it could be the sensor. I'm an engineer - if you or someone you know is good with a multimeter you can fix it yourself pretty easily. This is like a 20 minute fix for me. I would take it all apart so you can get access to the tiny little PT100 chip (it's the only chip on the board - you can follow the traces visually that the smaller two wires go to over to that chip. I would first just reflow the solder at the PT100 chip and at the connector. It's almost certainly the connector. There is mechanical stress there when the board moves and several people have had problems on the forum with that connector not soldered well enough. At least on the UM2. Basically the UM3 is identical in this regard and I don't remember if anyone has had a problem with the UM3 connector but if they get what you describe it's probably the same spot - the connector. 99% chance this fixes it. If not I would connect it back up and just hanging (not screwed in) and connect the larger two cables to a 24V 200Watt supply (not easy to find) and heat it up to 100C and while that is happening measure the resistance of the PT100. It will slowly increase from 109 ohms at room temp to around 200 ohms when hot but then suddenly sky rocket to infinite ohms. At that point you can probe each point along the path and find out where the failure is exactly. It might be that you need to run two tiny wires from the connector over to the pt100. In fact I'd just run those 2 wires if the reflow trick doesn't work and not even bother heating it. OR you can get a new bed heater from your reseller.

I wouldn't worry about "voiding warranties". Ultimaker is very nice about warranties. For example if you damaged the a20 with static you would indeed have to buy another one yourself but if your power supply died and that has nothing to do with your modifications then the supply would be covered. Besides, every part on the UM3 can be purchased from Ultimaker. So if you break anything you can get a new part.

You could certainly hire ahoeben - he's a competent programmer so he's not cheap. He used to work for Ultimaker. He'd need to rent a UM3. You should first explain what the goal is. As neotko says, the interface might not be able to handle the required quantity of gcodes per second.