gr5

Are you going to post the STL files?

Awesome!

Strange. Well it speeds up on the second layer so if your second layer height is similar to the bottom layer (typically bottom layer is .27 in cura and second layer is maybe .1mm but if second layer is more like .2 or .3...) then maybe you are just putting too high a volume of filament through the nozzle per second and it is underextruding (which is kind of what this looks like). Or it could be the fan. The fan comes on starting the second layer. Maybe it's blowing too much on the print head and cooling it down. Or maybe the fan shroud is touching the heater block (very common problem on UM2s) and when the fan comes on the nozzle temp drops like a rock. Or it could be the Z screw needs cleaning - although that would not affect the second layer - 3rd layer would be the first possible layer with an issue I think.

It might be easier to just straighten the rods you have now versus sourcing new ones. There's a guy who posted a video of how he rotates the rods and can measure easily where the bow is the worst - then he markes the rod then bends it the other way and then retests it. I think he fixed all his rods in an hour or so. I am not a fan of E3DV6 all metal. This hotend you are thinking of - is it all metal or does it have some PTFE? I strongly recommend a head with PTFE. I believe their 3mm head is all metal but they have a 1.75mm head that is quite excellent and has some ptfe in it (teflon). The problem with all metal is you can print anything but PLA and PLA is the best material out there (overall properties and ease of printing). But if you are willing to switch to 1.75 or if you are never ever going to print PLA again (which is fine - PET/nGen/CPE is a fantastic material also) then go for it. They may tell you it prints PLA but it doesn't. It gets stuck. molten PLA sticks very well to all metals.

Is your printer still using 19V? Or is it running 24V? Hopefully 19V because if you switch to 24V then that changes the wattage of your heaters and the voltage regulator will get quite hot as well. So you should be able to control PID from the menus. If not you can certainly change them using gcodes. There's a gcode to autotune the PID. There's one to set the PID, one to read back the PID and one to save into eeprom (permanent memory) all settings including latest PID values. But don't worry about PID unless your heater temp is oscillating (typically it oscillates with roughly 20 to 100 second period) badly. Such as more than 4C error for hours.

It's okay to turn the 3 screws an equal amount (CCW from below) about 1/4 turn at a time to get it to stick better *while* printing the bottom layer.

Oh! This is good. This is the problem. Even though it doesn't audibly sizzle at 225 it's still a problem. Is your dry box just a box with desiccant? You have to recharge the desiccant often. Put the PVA on a heated bed with a towel over it at 55C (it gets soft at 65C) for 2 hours. That should get the outer portion at least so it should work well for many hours. If it gets bad again try drying overnight the next time. For PVA I bought large 25g boxes of desiccant that changes color when it needs to be recharged.

What kind of printer is it? I think the pva extruding nozzle is too far from the bed. You need to calibrate the difference between nozzle height better. For example if this is an S5 then try heating the nozzle and cleaning it. If there is a tiny bit of filament on the tip of the nozzle during auto leveling it will throw off the measurement.

If that black thing in the photos is your bed I'm sure it's fine. It looks like "build tak" or a similar material. Build tak is great. No need to switch to glass (which is also great).

I found 2 used ones.

Leveling has other purposes. The more you squish the bottom layer the better the part will stick. I don't know what you mean by "leveling high" and "leveling low". On all of my printers the bed moves and the nozzle does not but even so, many people think of it as the nozzle that moves. Instead it's better to say "close" and "far" or "squished" or "big gap versus little gap". high/low doesn't mean much. Anyway nominal is to have things such that when Z=0 the nozzle touches the bed. Nominal sucks. You want to squish more than that. For example if your bottom layer is sliced at 0.3mm cura will set Z=0.3mm and extrude enough for .3mm of filament. But you want the nozzle to be closer to .15mm to .2mm so it squishes extra hard. The worst thing is for a part to come loose 5 days into a 6 day print. Better squish means the part sticks better.

So the issue is you have coasting turned on and the volume is set to 0.15. It might print just fine - I don't know. I've never used coasting. I notice you have a gap in your teal colored skirt also. This is also due to coasting.

Okay well I downloaded your stl file and it sliced in cura just fine. No gaps.

You have to decide what you care about the most. Speed? Beauty? Accuracy? There are tradeoffs. You should pick one of the 3 and ignore the other 2 as much as possible. Which of these 3 things do you want to learn to do first?

Well there are many possibilities but the most likely two: 1) The model is not manifold. Try viewing the model in xray view. If you see any red or brown then the model is really bad. You can probably fix it with the free netfabb repair service. You have to create an account but the repair service is then free. 2) Maybe the wall is just too thin there. Try setting the line widths to 0.2mm just for an experiment to see if the problem goes away. If so then this wall is too thin. There are lots of things you can do but playing with line width might be the easiest. There are thin wall options as well. A 0.4mm nozzle can print reasonably well down to 0.3mm. Line width smaller than that for a 0.4mm nozzle will start to get bad quality. 3) Normals are wrong. Probably the problem is #1 or #2 above.

So this looks like basic underextrusion - when underextrusion is on the bottom layer it's usually a leveling issue. Notice the gaps between linfill lines as well. There is not enough filament coming out to fill in the gaps on the bottom layer. If this is happening on every layer then you are printing too cold or too fast for your extruder/nozzle combination. But most likely your leveling needs to have the bed closer to the nozzle when printing that bottom layer and this problem goes away by the 3rd layer.

90mm/sec doesn't mean much without knowing layer height, line width and temperature. And nozzle size. I doubt you are comparing apples to apples here. Also some types of infill only print one angle on one layer and the other angle on the next such that every other layer is "empty". Basically it's purposefully printed with 50% infill. Which type of infill pattern are you using? The UM2+ and UM3 and S5 can more consistently print twice as fast as the below speeds but I don't like to push the limits. Here are top recommended speeds for .2mm layers (twice as fast for .1mm layers) and .4mm nozzle: 20mm/sec at 200C 30mm/sec at 210C 40mm/sec at 225C 50mm/sec at 240C The printer can do double these speeds but with huge difficulty and usually with a loss in part quality due to underextrusion. So if you are printing 90mm/sec at 210C with .13mm layers, and .35mm line width, yes the printer can do this but with not much margin for error (like if there is a mild tangle or your bowden is getting old or many other minor annoyances that can push the printer into underextrusion territory.

I absolutely love the UM2go. I have 3 of them. I use them more than my other printers (UM2/UM3/S5/UMO). See if you can find a used um2go. In my opinion it's the best quality of all their printers and it's just a fantastic printer. I modified all of my um2go's. I added a heated bed to all 3, I added an olsson block and the iroberti feeder with the medusa 2x feeder upgrade.

Damn. Didn't mean to post the above - meant to start a new paragraph. Anyway on the UM3 I just use the default settings in cura usually for temperature (but turn off that stupid retract on layer change thing that made it's way inexplicably into the AA 0.8 profile). I change other things - I set all the speeds the same. I like to print slow (30-35mm/sec) and thick (0.3mm) with line widths at 0.8. But I haven't had to mess with temperature on the AA 0.8 profile. Also on the UM3 if I use a 3dsolex core I have to lower the temp on the 0.8 nozzle versus an Ultimaker AA 0.8 by 10C or it will string a lot. the 3dsolex core heats the filament more thoroughly because of the dual passage that the filament travels through in the nozzles. So you don't need to raise the temp at all with the 3dsolex core.

I agree that sounds like a lot. I usually print with the same temperature. If you are printing near the limits of capability you should raise the temp. On the UM2go I print with the same temp and print about 3X more volume than with the 0.4mm nozzle.

The PID parameters can be modified in many ways including by storing them on the core itself. So you could have a "low temperature core" that has a different set of PID values than normal. You can play with PID values to find the best values using the command_util.py. Have you used ssh to connect into your printer yet? Here are a few of the steps: First put your UM3 or S5 into developer mode - it's in the menus on the printer. Your machine needs to be on your network (wifi or ethernet) if it isn't already. Once it's on the network it will show the IP address at the top of the main screen. Next you need ssh which is built into linux and Macintosh terminal but not windows. For windows I recommend putty: https://www.chiark.greenend.org.uk/~sgtatham/putty/latest.html In putty you don't type the "ssh" part but just put the ultimaker@111.222.333.444 part and click "open". ssh ultimaker@1.2.3.4 (don't enter 1.2.3.4 - enter the ip address listed on your UM3) username/password: ultimaker/ultimaker (much easier than root/ultimaker as it takes you straight into the utility to do sendgcode) At this point you can do: select printer printer/head/0/slot/0 list And you will see all the parameters for the left core. To change one do for example: set pid_Kp 5 Type list again to make sure it worked. Then heat up the core manually on the front. You can change parameters live. If you remove and insert the core you will lose these values. To save these values you can edit py files or json files on this linux computer inside the printer. Which is a little dangerous - make sure you have an olimex cable first (I sell them if you are in usa) or you can modify them on the core itself which is safer. Email me for more details or more questions. Email me directly as I'm hesitant to explain how to modify the core on this forum. I'm at gr _at_ gr5.org. After emailing me check your spam folder as you have to go through a very simple approval process to get to me (prove you aren't a spamming robot). Or post your email here and I'll email you directly but I might not check back into the forums until tomorrow so emailing me is much faster.

A picture would help. But if the bottom layer (and only the bottom layer) is underextruded as you describe then maybe your bed needs to move closer to the nozzle. You want to have a good squish on the bottom layer or the part will come off or at least corners will lift off. Also the bottom layer tends to print much faster volume (0.3mm typically versus 0.1mm for the rest of the layers typically) so you may just be printing too cold or too hot. But more likely it's a leveling issue. The gap can also be partly caused by backlash/play. Especially if it's not symmetrical. Here's more about backlash and gaps between wall and infill:

The cracking is caused by insufficient layer bonding. The layer being laid down is not hot enough to melt the layer below. Or it melts it but only along a 0.1mm line and you need much better bonding/adhesion. 1) You need to cover the front and the top of your UM3 to get the air temp up to 35C. Just use a gallon ziplock or saran wrap and for the top use a box - the type for copy paper is perfect (no cutting needed, no taping needed - it just fits on top perfectly. 2) You need to lower the fan speed. Play with fan speed on UM3. For my UM3 (yours is likely different) 10% fan speed is identical to 100% so cura's suggested 50% fan speed is rather stupid. I use 3% fan speed with ABS on my UM3. 30% on my UM2s. 3) Bed should be at 105C minimum. This is how you solve the other issue (the lifting corners) combined with consider using pva juice or ABS juice (google abs juice) and use the brim feature. Brim makes a large difference. I recommend 110C bed temp. This is impossible to reach without covering the front and top. ABS gets soft at 99C so if the glass isn't above this temp it will warp off the bed instead of relieving the shrinkage forces. Much more about all this on my video. 4) Raise nozzle temp. I don't recommend messing with this as the UM suggestion works well. ABS is very difficult to pick the nozzle temp. To cold and you get bad layer adhesion and "cracking". Too hot and you get clogs. ABS at even just normal printing temps but left for a few minutes at this temp in the nozzle will turn it into a horrible gummy clog. 5) Layer heights. Some people say thinner layer is better as the nozzle is closer to the layer below to heat it. Some say thicker is better as the mass of heat is enough to melt the layer below. I don't know what the truth is as I haven't played with this much but try sticking with 0.1mm to 0.2mm if you don't want to experiment yourself. For adhesion to the glass, in addition to the stuf in #3 above there are a lot more things you can do. I mention them all in great detail in this long video but it's packed with information. You might not need to watch the video but if brim, abs juice and 110C bed aren't enough then watch this video:

No need to change any code. Go into the power budget settings and set your heated bed power to 1 watt and the software will realize it can keep the bed on full power even when the nozzle is on full power. Or alternatively set the power budget to 255. Again the software will realize there is no need to limit power.

@smartavionics have any ideas?